Predicting the Antigenic Evolution of Influenza Viruses with Application to Vaccination Strategy

Seasonal influenza viruses cause substantial worldwide mortality and morbidity every year. The evolution of these viruses comprise unique systems for studying natural evolutionary processes in real-time. Host immune systems recognise pathogens based on binding affinities between host antibodies and pathogen antigens. Pathogens with similar antibody binding are said to be antigenically similar. Seasonal influenza viruses evolve antigenically over a timescale observable by humans. Vertebrate immune responses also adapt rapidly, such that a single infection usually leaves a host protected for life against antigenically similar strains. In humans this imposes natural selection for antigenic novelty in wild virus populations. Influenza vaccines contain virus antigen which elicit the production of antibodies that protect against antigenically similar strains. After major antigenic evolution in wild viruses, vaccines must be updated to remain effective. I use a simple model to show that even in best case scenarios current influenza vaccination strategies cannot completely avoid antigenic mismatch. I then present the first study that quantifies the link between degree of antigenic mismatch and vaccine effectiveness. Knowledge of the molecular variation responsible for major antigenic change in natural influenza viruses has improved greatly in the last decade. I review this work, and present the application of a computational approach from the field of quantitative genetics to this problem. Subsequently, I show that patterns in biophysical features of substitutions responsible for major antigenic change are far from random, begging the question: how predictable are these amino acid substitutions? I answer this question by developing a computational framework to rank candidate substitutions by their biophysical similarity to substitutions responsible for major antigenic change and show that predictions can be made that are substantially better than chance selections. Finally, I discuss the application of these rankings to advanced influenza vaccination strategies based on the principle of immunity management. I expand on how this work should be the basis for further investigation into the mechanisms that govern different components of influenza virus fitness and ultimately the antigenic evolution of seasonal influenza viruses