NIH Director's Pioneer Award: Antigenic Cartography

点击翻译按钮获取中文摘要

Influenza viruses are classic examples of antigenically variable pathogens, and have a seemingly endless capacity to evade the immune response. For example, since the influenza A(H3N2) subtype entered the human population circa 1968 the vaccine against it has had to be updated 24 times to track the evolution of the viral quasispecies and to remain effective. Most of the bioinformatics methods for analyzing viral evolution are based on genetic analyses; however, it is the phenotypic (antigenic) properties of the virus that determine its success at escaping prior immunity and causing infection. Indeed, the antigenic data are the primary criteria for selecting the virus strain used in the influenza vaccine, and which are important for much basic and applied research on influenza. However, there is no reliable method to determine quantitatively antigenic differences. Antigenic differences are typically determined using some form of binding assay (for influenza virus, the hemagglutination inhibition assay, for other pathogens a neutralization assay, ELISA, etc.). In such assays, typically a panel of antisera is titrated against a series of antigens and the data are organized in tabular form and analyzed by eye. This has been the procedure for over 50 years. These data are difficult to interpret quantitatively, and sometimes even for experts give an inconsistent picture. The primary reason for this difficulty is that the data contain irregularities, or paradoxes. On such irregularity is that one antiserum might detect a difference between two antigens, while another will not. Another irregularity is that heterologous titers are sometimes higher than homologous titers. Furthermore, it is often difficult to compare data from different laboratories. These, along with other irregularities result in binding assay data only being considered reliable enough to judge large antigenic differences?in the case of influenza virus differences of sufficient magnitude that they necessitate an update of the vaccine strain. By only being able to judge gross differences among the thousands of influenza strains characterized each year, one misses opportunities to optimize the vaccine strain choice, to detect signals in the evolution of the viral quasispecies that could give advance warning of the necessity to update the vaccine, to understand the epidemiology of influenza, to judge how vaccination affects the viral evolution, and to devise novel intervention strategies which have the potential to fundamentally change our options to control epidemic and pandemic influenza. Influenza virus is one example of an antigencially variable pathogen; others include human immunodifficiency virus and hepatitis C virus. The degree of antigenic diversity will increase as interventions increase selection pressure to generate escape mutants, and the characterization of these phenotype differences will thus only increase in importance.