Drivers of individual variation in influenza vaccine response and protection from infection

点击翻译按钮获取中文摘要

PROJECT SUMMARY The induction of protective immune responses through vaccination is central to the management of many pathogens. For antigenically variable pathogens such as influenza, protective immune responses impose a major selective pressure on viral populations and indirectly influence vaccine strain selection and vaccine effectiveness. Our poor understanding of the generation and maintenance of protective immunity to influenza hinders vaccine development and the accuracy of evolutionary forecasts. Antibody titers to the hemagglutinin (HA) surface protein were established as a correlate of protection 50 years ago, and more recent evidence shows many anti-HA antibodies directly and indirectly contribute to viral neutralization. However, HA titers remain only moderately predictive of an individual’s risk of infection on exposure, and the contributions of other immune responses are less well understood. Understanding the causes in addition to correlates of protection could increase the accuracy of forecasts of viral fitness and provide reliable endpoints for vaccine development. Here, we propose complementary approaches to identify the correlates and drivers underlying protection from infection and heterogeneity in vaccine responses. We will integrate diverse variables, including infection and vaccination history, baseline antigen-specific and antigen-agnostic immune states, intrinsic characteristics including age, sex, and body mass to predict responses to influenza vaccination and extract mechanistic insight. In order to address our specific aims, we will leverage data from existing, longitudinal studies of immune parameters following influenza virus infections and vaccination in humans. First we will use computational and multimodal single-cell approaches to investigate how vaccination and infection impact host immune status. Emerging evidence, including our own data, suggests that vaccination and infection can establish new antigen-agnostic immune set points that affect future vaccine responses. Next we propose to integrate complementary computational approaches, spanning machine learning, causal mediation analysis, and mechanistic modeling to predict and develop causal mechanistic insight into vaccine responsiveness and protection from severe and mild infection. We will develop and distribute a suite of accompanying tools to make these novel approaches accessible to bench and computational biologists. Improved prediction of immune responses, especially protective immune responses, could lead to more effective vaccination strategies that mitigate vaccine failure in different subpopulations and improve the public health impact of influenza vaccination. The methods and tools that we develop can provide foundational frameworks to dissect responses to other vaccines and pathogens.