Predictive Biosignatures for Complicated Novel H1N1 Influenza

Clermont, Gilles; Ross, Ted

Abstract

This proposal will join for the first time advanced computational algorithms, multi-scale mathematical models, state-of-the-art imaging and biological measurements, a highly relevant animal model, and existing human data towards a translational problem of imminent relevance to human health: pandemic influenza A virus infection (IAV) and its potential to cause critical illness and death in a large number of individuals. Indeed, the 2009 H1N1 influenza reassortant virus caused an estimated 18,000 deaths so far, targeted younger individuals and pregnant women, caused acute lung injury in close to half of cases admitted to the intensive care unit, and was also associated with a secondary bacterial infection in up to 35% of the these cases. Not only is 2009 H1N1 of major significance in itself and expected to display at least a third wave in the fall of 2010, but it represents a prototypical emerging infectious disease of pandemic proportion. As such, it offers an exceptional opportunity to deepen our understanding of (1) mechanisms associated with influenza a virus pathogenicity, (2) prognostic biomarkers of severity and of complicating bacterial infections, and (3) of the potential contribution of tools leveraged from the physical sciences to enhance knowledge and preparedness for the next 'big one'. We propose to (1) model real-time observations of cellular, inter-cellular processes and organ function at multiple levels including in vivo imaging, (2) develop non-invasive, model-based predictors of severe complications and of outcome of high translational relevance, (3) develop robust methods for parameter identification and estimation on a variety of mathematical frameworks such as 3D and compartmental dynamical systems, (4) and use these models to map these experimental data to existing human data and generate predictions of very early biosignatures of complicated disease. To achieve these goals, this proposal will assemble a database of the most detailed multiscale, longitudinal observations ever collected that will be undoubtedly used by other groups of biological and physical scientists beyond the proposed effort. The underlying premise of this proposal is that model-based interpretation of complex biological data in a relevant animal model of IAV, combined with incomplete, yet relevant human data, will lead to new predictive biomarkers of complicated illness and new therapeutic approaches.

Public Health Relevance

Even if the 2009 H1N1 pandemic influenza a virus was dubbed a relatively mild infection, it nevertheless killed 18,000 to date, infected 60 million people in the US alone, and overwhelmed critical care resources in many areas. The 2009 experience was a clear demonstration of our ongoing vulnerability as a society and has not attenuated fears of a doomsday scenarios should a big one indeed strike. This proposal will accomplish the most in-debt study of influenza an infection to date and combine it with advanced mathematical and computational tools to improve our ability for early detection of cases of influenza A that will eventually become severe. A successful completion of this research will also provide tools that will integrate early detection and strategies to avert such a complicated course, and possibly death, in individual patients.