DBIO Dissertation Award: Low-dimensional models of immune sensing and signaling

T cells have the remarkable ability to sense fine differences between antigens presented in mixtures. Quantitative models of T cell activation are crucial to harness T cell specificity for immune therapies. However, the biochemical details of TCR activation quickly become intractable. To address this challenge, in my doctoral thesis, I worked in close collaboration with experimentalists to develop low-dimensional theories of T cell sensing and signaling. We first analyzed the production of cytokines – extracellular messenger proteins – by activated T cells. We found a two-dimensional representation of high-dimensional cytokine dynamics, parameterized with simple ballistic equations. We used information theory to quantify antigen encoding in this latent space, revealing a continuum of T cell responses. Building on the insight that early antigen sensing determines these responses, we revised previous kinetic proofreading models of TCR activation to explain paradoxical cross-receptor interactions in chimeric antigen receptor (CAR) T cell immunotherapy. Our model predictions quantitatively matched in vitro experimental results. The model finally led us to design CAR T cells in which inhibition by weak TCR antigens protects healthy tissues from treatment toxicity.