THE TRANSMISSION-VIRULENCE TRADE-OFF AND SUPERINFECTION: COMMENTS TO SMITH

The virulence-transmission relationship in an obligate killer holds under diverse epidemiological and ecological conditions, but where is the tradeoff?

Parasite virulence is a leading theme in evolutionary biology. Modeling the course of virulence evolution holds the promise of providing practical insights into the management of infectious diseases and the implementation of vaccination strategies. A key element of virulence modeling is a tradeoff between parasite transmission rate and host lifespan. This assumption is crucial for predicting the level of optimal virulence. Here, I test this assumption using the water flea Daphnia magna and its castrating and obligate-killing bacterium Pasteuria ramosa. I found that the virulence-transmission relationship holds under diverse epidemiological and ecological conditions. In particular, parasite genotype, absolute and relative parasite dose, and within-host competition in multiple infections did not significantly affect the observed trend. Interestingly, the relationship between virulence and parasite transmission in this system is best explained by a model that includes a cubic term. Under this relationship, parasite transmission initially peaks and saturates at an intermediate level of virulence, but then it further increases as virulence decreases, surpassing the previous peak. My findings also highlight the problem of using parasite-induced host mortality as a "one-size-fits-all" measure of virulence for horizontally transmitted parasites, without considering the onset and duration of parasite transmission as well as other equally virulent effects of parasites (e.g., host castration). Therefore, mathematical models may be required to predict whether these particular characteristics of horizontally transmitted parasites can direct virulence evolution into directions not envisaged by existing models.

Superinfection and the coevolution of parasite virulence and host recovery

Parasite strategies of host exploitation may be affected by host defence strategies and multiple infections. In particular, within-host competition between multiple parasite strains has been shown to select for higher virulence. However, little is known on how multiple infections could affect the coevolution between host recovery and parasite virulence. Here, we extend a coevolutionary model introduced by van Baalen (Proc. R. Soc. B, 265, 1998, 317) to account for superinfection. When the susceptibility to superinfection is low, we recover van Baalen's results and show that there are two potential evolutionary endpoints: one with avirulent parasites and poorly defended hosts, and another one with high virulence and high recovery. However, when the susceptibility to superinfection is above a threshold, the only possible evolutionary outcome is one with high virulence and high investment into defence. We also show that within-host competition may select for lower host recovery, as a consequence of selection for more virulent strains. We discuss how different parasite and host strategies (superinfection facilitation, competitive exclusion) as well as demographic and environmental parameters, such as host fecundity or various costs of defence, may affect the interplay between multiple infections and host-parasite coevolution. Our model shows the interplay between coevolutionary dynamics and multiple infections may be affected by crucial mechanistic or ecological details.

Co-infection and super-infection models in evolutionary epidemiology

Multiple infections are intensively studied because of their consequences for the health of the host but also because they can radically alter the selective pressures acting on parasites. I discuss how multiple infections have been modelled in evolutionary epidemiology. First, I briefly mention within-host models, which are at the root of these epidemiological models. Then, I present the super-infection framework, with an original focus on how the definition of the super-infection function can lead to evolutionary branching. There are several co-infection models and, for each of them, I briefly go through the underlying mathematics (especially the invasion fitness of a mutant strain) and I discuss the biological assumptions they make and the questions they consequently may ask. In particular, I show that a widely used co-infection model should not be invoked for invasion analyses because it confers a frequency-dependent advantage to rare neutral mutants. Finally, I present more recent frameworks, such as the Price equation framework in epidemiology, that can account for increased parasite diversity. To conclude, I discuss some perspectives for the study of multiple infections in evolutionary epidemiology.

Multiple infections and the evolution of virulence

Infections that consist of multiple parasite strains or species are common in the wild and are a major public health concern. Theory suggests that these infections have a key influence on the evolution of infectious diseases and, more specifically, on virulence evolution. However, we still lack an overall vision of the empirical support for these predictions. We argue that within-host interactions between parasites largely determine how virulence evolves and that experimental data support model predictions. Then, we explore the main limitation of the experimental study of such 'mixed infections', which is that it draws conclusions on evolutionary outcomes from studies conducted at the individual level. We also discuss differences between coinfections caused by different strains of the same species or by different species. Overall, we argue that it is possible to make sense out of the complexity inherent to multiple infections and that experimental evolution settings may provide the best opportunity to further our understanding of virulence evolution.