An experimental approach to uncovering the consequences of parasite co-infection in Peromyscus
Infectious diseases are ubiquitous in nature and can be strong selective forces for the evolution of wild populations. Individual hosts are commonly infected with multiple parasites simultaneously, and this pattern has been documented in many systems. Recent evidence suggests parasites can partition host resources as direct competitors or indirectly through resource- or immune-mediated interactions, and these interactions can have significant implications for host fitness and the health of the population. Nevertheless, much of this work has focused on single parasite infections, though in nature and in humans, most hosts are infected with multiple parasites. The consequences of ‘co-infection’ can be important for the immune response and host health by changing the ecology of host population dynamics, negatively affecting host fitness in the form of disease, and disrupting the normal host response to infection; however little research has focused on this area until relatively recently. Understanding the complex dynamics of co-infection can be difficult; yet, using a community ecological approach can provide a framework to make sense of these complicated interactions.
Indeed, parasite infection was common in the Mountain Lake Biological Station populations of wild Peromyscus hosts, with over 90% of mouse captures over the course of 3 consecutive mouse breeding seasons (May – August: 2010-2012) infected with 3.4 different parasites on average. Using rigorous and controlled statistical models, I found that when co-infections were present, parasite associations were positive in nature and variable in strength from weak to strong. Most parasite associations were contemporary, but a few were time-lagged, and the combination of all associations created a complex within-host parasite community.
Using a perturbation experiment, I found there was a decrease in nematode burden with ivermectin treatment, and a concomitant increase in coccidian protozoa. Importantly, mice treated with a single oral dose of anthelmintic had significantly lower survival than control mice, suggesting that treatment may be detrimental to survival and fitness because of unintentional increases in co-infected parasites. Furthermore, these results suggest that there are strong negative interactions between nematodes and coccidia. Additionally, effects of ivermectin treatment were not isolated to treated individuals only, as there were knock-on effects to untreated individuals within the population of a similar type and magnitude as seen in mice that received ivermectin treatment.
Experimental perturbation of the within-host parasite community revealed that parasites interact in a reciprocal manner. To examine direct and indirect interactions between microparasites and macroparasites across differing host systems (i.e. blood and gastrointestinal tract), I applied a randomized four-level factorial treatment (control, ivermectin, fipronil, dual ivermectin and fipronil) to individuals in the host Peromyscus population designed to reduce either microparasites (via fipronil to reduce fleas, a vector of Bartonella), macroparasites (via ivermectin to reduce nematode infection), or both (via dual treatment). Again, ivermectin treatment reduced nematodes, and an interaction with coccidia resulted in reduced recapture. However, fipronil treatment had no effect on Bartonella infection, and there were unexpected effects of fipronil on coccidia. It has been suggested that fipronil also acts to reduce nematode infection, so the effect of fipronil on coccidia was also, possibly, an effect of nematode removal. These results suggest that microparasites may impose a heavier cost to their hosts than macroparasites, especially if infection intensities increase suddenly or rapidly.
Overall, these data provide support for conducting manipulative field experiments, rather than relying on observations of non-random associations to determine that parasites are interacting with each other. However, if manipulative field experiments are not possible because of logistical or ethical constraints, then the most reliable way to detect the presence of strong parasite associations is to conduct a combined cross-sectional and longitudinal analysis, controlling for as many confounding individual host traits and environmental factors as possible, to look for associations that remain consistent across both tests. In addition, the data from these studies present strong evidence that within-host communities provide an arena for direct and indirect host and parasite interactions that apply strong selective forces to populations at larger scales.