Levin and Parker (2014) have developed a very clever method to track the movement of an avian blood parasite vector in the Galapagos Islands using a system consisting of great frigatebird hosts, hippoboscid fly vectors and the blood parasite Haemoproteus iwa. In their study, they used the most variable microsatellites in the frigatebird population. Then, they compared the microsatellites from the blood of a bird to the bloodmeal found within a fly collected on that same bird. They inferred from mismatches between paired microsatellite signatures from the bird and from a bloodmeal within the fly, whether or not the fly had recently moved between hosts. Additionally, by keeping track of the sex of the bird from which a fly was collected, they found that flies that had switched hosts were collected more often from female birds. To see what role blood parasite infection might play in fly host switching, they determined the Haemoproteus infection status of birds and flies in their study. Haemoproteus parasites are found in the vertebrate blood cells, and so it is straightforward to determine the infection status of a frigatebird blood sample by using PCR. However, determining the infection status of a fly requires first ruling out detection of Haemoproteus DNA from an undigested bloodmeal; because fly infection can take a few days, parasites taken up in a meal do not necessarily indicate that a fly is infective. Haemoproteus iwa parasites mature and migrate to the fly's thorax when they reach the infective stage, whereas the bloodmeal only passes through the thorax on its way to be digested in the abdomen. To determine the infection status of a fly, the authors first isolated the thorax by dissection, and if it contained Haemoproteus DNA but no bird DNA, they could infer that (i) the fly had finished processing the bloodmeal through the thorax and (ii) the fly was infected with Haemoproteus. Using a predictive model that included the infection status of both host and vector, they quantified the importance of host sex and infection status on vector movement and host choice (Fig. 1).
The use of molecular tools to track vector movement is an exciting advance for the study of arthropod-transmitted pathogens. Previous studies of vector movement have involved tedious mark–recapture methods. In a tremendous effort more than 50 years ago, 560 flies were individually marked with unique colour combinations of three dots of cellulose paint and were released one at a time with individually marked birds (of several species) on a small island in Shetland (Corbet 1956). The birds were later re-trapped and the movement of the flies and any preferences for certain species of birds were determined by the data collected from this massive re-trapping effort. Corbet (1956) found more flies on larger birds, on hole-nesting birds and on juvenile birds (compared to adults). About 25% of the flies had switched hosts. Of the recovered flies, more than 50% of male flies had changed hosts, but less than 20% of female flies had switched hosts, suggesting that behavioural differences between sexes in the vector population may play an important role in disease dynamics for systems where both sexes take blood and transmit pathogens.
Levin and Parker's (2014) examination of vector movement in the frigatebird–hippoboscid fly–Haemoproteus system is eloquent. By using molecular methods to estimate fly movement, they avoid laborious trapping and re-trapping of birds to try to physically follow the path of each individual fly. This approach simultaneously reduces the stress placed on the birds, and increases sample sizes, because only one trapping ‘time point’ needs to be collected to tell the story of a fly's travels. Their results are intriguing and open up new questions to be investigated. Interestingly, they found that uninfected flies were more likely than infected flies to have a bird genotype in their bloodmeal that was different from that of their current host (i.e. more likely to have switched hosts). Why do uninfected flies move more (or infected flies move less)? It could be that the flies are ‘sick’; the parasite could be either causing direct harm to the fly or indirect harm by removing resources that may be used for locomotion or other activities. Flies infected with Haemoproteus parasites have shorter survival and produce fewer offspring than those that are uninfected (Waite et al. 2012a). Alternatively, it could be that infected birds preen less, as was found in a similar avian malaria system in Hawaii (Yorinks & Atkinson 2000). Hippoboscid flies are vulnerable to preening (Waite et al. 2012b), and so flies that have found a host with lowered defences may stay on this host longer, and subsequently become infected. It remains to be investigated whether flies would be more likely move again as hosts recover from infection, because defence behaviour would likely increase with recovery. Work on vector preferences has shown that in another avian haemosporidian system, mosquito vectors prefer chronically infected birds over uninfected birds, regardless of the infection status of the mosquito (Cornet et al. 2013). If a similar phenomenon occurs in the frigatebird–hippoboscid fly system, uninfected flies might be more likely to move to a more attractive (infected) host, becoming infected as they stay and feed.
This research provides a powerful new approach for understanding and predicting how vectors move between hosts, and could be modified for other systems to address the many questions about vector movement that remain unanswered. By finding microsatellites to uniquely identify individuals in a small population, and then using these microsatellites in a comparative approach to estimate vector host-switching frequency, we might be able to determine what factors influence vector movement in many other systems, potentially opening up new avenues for disease control. The approach of Levin and Parker (2014) was so successful in part because of several system-specific details, and such elements can also be found in other vector systems. The hippoboscid fly vector has a close association with the frigatebird host and is not often found in the environment without a host. Additionally, the frigatebird populations are localized to the Galapagos Islands, a considerable distance from other landmasses, and so they may be more traceable than, for example, human populations across a larger land mass. Employing their techniques could be possible at a local scale for even a seemingly dispersed disease system like the human–mosquito–malaria one. For example, it seems feasible to sample within a village, using an anthropophilic mosquito like Anopheles gambiae. It could be informative to compare the microsatellite signatures of mosquito bloodmeals to microsatellites of people in the same house as where the mosquito was found, and to people in nearby households and the rest of the village. Amassing real data on vector movement and preference in nature would be invaluable to informing control efforts, making more predictive models, and for evaluating our current efforts to protect ourselves, livestock, and crops from biting arthropods and the pathogens they carry.