Immune avoidance strategies in RNA viruses: fitness continuums arising from trade-offs between immunogenicity and antigenic variability.

Highly exposed and protruding amino acid sites on the surface of viral capsids are subject to fewer residue interactions and packing constraints than those buried within protein interiors. Consequently they often experience higher rates of non-synonymous substitution and exhibit greater genetic variability than buried interior sites. However such protrusive surface structures often induce host immune responses and are likely to constitute B cell epitopes. Genetic variation at these surface sites is therefore likely to correspond to antigenic variation. This may be of adaptive value to the virus for two quite different reasons. The first is that antigenic variation arising over the course of a viraemia may result in greater net viral replication, and increased opportunities for viral transmission. The second is that antigenic variation generated rapidly over a single infection or incrementally over several sequential infections may give rise to variants that are sufficiently immunologically distinct that they can reinfect host individuals with previous infection experience of related virus. This would lead to an extension of the susceptible host pool with consequent increase in transmission opportunities. The surface architecture of viral capsid proteins is therefore conceivably subject to two opposing selection pressures: one to minimize the surface area accessible to interaction with elements of the immune system, the other to increase the potential access to antigenic variation by adoption of exposed and unconstrained protein conformations. Therefore, there exists a possible trade-off between the fitness benefits deriving from potential ability to generate antigenic variation, and the increased immunogenicity with which such potential may be associated. We propose that the existence of this trade-off would lead to a continuum of different strategies by which a virus might combat an immune response. We explore this strategy space with simple mathematical models, and show that peak loads of infectious virus particles are proportional to levels of antigenic diversity, and inversely proportional to immunogenicity, thereby creating the potential for a trade-off by which fitness might be maintained with a continuum of strategies. This may remain possible even if the antigenic variants are not transmissible between hosts, so long as immune resources are sufficiently dispersed between antigenic variants. The diversity of possible strategies is discussed with reference to the Picornavirus family.