Engineering cross-reactivity in the antibody response to HIV and influenza

Abstract

Antibodies are a key component of the human immune response to infectious disease. The best anti-viral antibody response is one that is potent and broad, covering a large number of the many diverse viral variants. Unfortunately, natural human antibodies are rarely perfect in that they don’t cover the entire spectrum of possible viral variants. In this thesis, I describe my work using computational design with the ROSETTA software to re-engineer human antibodies to improve their coverage of large viral panels, also known as their breadth. I developed two new methods for increasing the scale of computational design against many viral proteins, a technique known as multistate design, and applied one of these methods to an anti-influenza system to define the molecular limits of breadth and affinity. In addition, I use computational modeling to identify and optimize antibodies that cross-react between different viral proteins. Finally, I use computational modeling to engineer peptides capable of recapitulating the activity of a broadly neutralizing antibody with increased breadth.