Phenotypic differences in viral immune escape explained by linking within-host dynamics to host-population immunity

Abstract

Viruses that do not cause life-long immunity persist by evolving rapidly in response to prevailing host immunity. The immune-escape mutants emerge frequently, displacing or co-circulating with native strains even though mutations conferring immune evasion are often detrimental to viral replication. The epidemiological dynamics of immune-escape in acute-infection viruses with high transmissibility have been interpreted mainly through immunity dynamics at the host population level, despite the fact that immune-escape evolution involves dynamical processes that feedback across the within- and between-host scales. To address this gap, we use a nested model of within- and between-host infection dynamics to examine how the interaction of viral replication rate and cross-immunity imprint host population immunity, which in turn determines viral immune escape. Our explicit consideration of direct and immune-mediated competitive interactions between strains within-hosts revealed three insights pertaining to risk and control of viral immune-escape: (1) replication rate and immune-stimulation deficiencies (i.e., original antigenic sin) act synergistically to increase immune escape, (2) immune-escape mutants with replication deficiencies relative to their wildtype progenitor are most successful under moderate cross-immunity and frequent re-infections, and (3) the immunity profile along short host-transmission chains (local host-network structure) is a key determinant of immune escape.

Introduction

Selection by host immunity, whether natural or artificially induced, is a major driver of viral evolution. Host immunity imposes selection on viruses at two linked scales: replication within hosts and transmission/establishment in subsequent hosts. Immune responses within hosts impact transmissibility through affecting the time course of viral abundance; moreover, levels of immunity in the host population modify patterns of transmission by determining establishment of infection and hence the length and structure of host-to-host transmission chains. A mechanistic understanding of the interdependence of immunity structure and viral evolution across scales is thus key for understanding disease-emergence risk and efficacy of viral disease-intervention methods.

In addition to host immunity, intrinsic genetic constraints in the viral genome govern immune escape. Phenotypic studies of immune-escape mutants show that these mutations are often pleiotropic, causing reduced replication rates (Berkhoff et al., 2006, Novella et al., 2005, Rimmelzwaan et al., 2005, Rudneva et al., 2005). This reduction in replication can impose a significant fitness trade-off for the virus, since experimental adaptation studies show that immune-escape strains acquire compensatory mutations when the specific immune pressure is removed (Rimmelzwaan et al., 2005, Rudneva et al., 2005), and wildtype strains often re-emerge several seasons after emergence of immune-escape strains (Berkhoff et al., 2007, Boon et al., 2002). However, immune-escape mutants with replication deficiencies can arise and predominate in a wildtype population extremely rapidly during seasonal epidemics of influenza, even when the margin of advantage from immune escape is thought to be small (Berkhoff et al., 2007, Voeten et al., 2000). What conditions explain this counter-intuitively high rate of immune-escape evolution?

Epidemiological models of immune escape in acute-infections (e.g., influenza viruses, noroviruses, rotaviruses) have identified how individual factors that govern the rate of escape operate; from viral life-history traits such as replication rates or infectious periods, to immune-mediated processes such as cross-immunity and recovery rates, to population processes such as epidemic size and herd immunity (Boni et al., 2006, Gog, 2008, Gog et al., 2003, Nuno et al., 2007, Park et al., 2009, Recker et al., 2007). However, these types of population models do not explicitly account for the effects of immunity on transmissibility throughout infection (i.e., time-varying transmissibility), thus separating the scales at which selection operates. Recently, nested models of host-parasite evolutionary dynamics, which link within- and between-host dynamics, have been developed to address this deficit (Alizon and van Baalen, 2008, Andre and Gandon, 2006, Coombs et al., 2007, Ganusov and Antia, 2006b, Ganusov et al., 2002, Gilchrist and Sasaki, 2002, Lange and Ferguson, 2009, Luciani and Alizon, 2009, Mideo et al., 2008, Read and Keeling, 2006). These models have derived fitness trade-offs that arise between within-host parameters and host responses, and shown that linkage of within- and between-host selection predicts different evolutionary dynamics than when these scales are considered separately. Thus, nested models provide an essential mechanistic foundation for examining the evolutionary epidemiology of viral immune escape.

Two recent studies emphasize the utility of nested models for understanding viral immune escape. Lange and Ferguson (2009) show how different host-contact rates and contact-network structures predict different between-host fitness optima that can be reached through selection on viral replication rate (i.e., a within-host viral parameter). Luciani and Alizon (2009), on the other hand, address how between-host fitness guides evolution of within-host viral replication, showing that optimal replication rates are affected by mutation rates and cross-immunity. Both studies reveal that viral replication rate evolves differently depending on between-host fitness optima. However, neither study investigates how the within-host dynamics change the profile of immunity in the host population, which alters immune-selection pressure. Furthermore, previous nested models have focused on invasion fitness rather than how fitness may change as the host-population immunity profile changes. In order to predict immune-escape evolution in acute viral infections, several processes remain to be examined mechanistically in more detail: (1) how within-host interactions such as direct competition between strains and cross-immunity impact the immunity profile in the host population, (2) how the profile of host-population immunity modifies host-to-host transmission chains, and (3) how previous infection dynamics determine selection of immune-escape phenotypes.

We used a nested model to examine how the reciprocal influences of viral phenotype and host immune response impact selection of an immune-escape mutant. Our two-strain within-host model specifies viral phenotypes according to two parameters: replication rate and cross-immunity. We begin by identifying how replication rates and cross-immunity affect the within-host viral and immune effector dynamics and transmissibility. We then investigate how viral-induced immunity structure in the host population impacts the evolutionary trajectory of an immune-escape phenotype by comparing host transmission chains with arbitrarily determined immunity to those with immune memory (i.e., from previous infection dynamics). By linking within-host dynamics, transmission and immune memory, we found that a mutant with a moderate replication deficiency can have a selective advantage in the host population. We show how host-population immunity, which is the selective force behind immune-escape evolution, critically depends on within-host dynamics and interactions between particular viral phenotypes.