Theoretical systems biology holds the promise of ultimately building a predictive causal theory of biology that has the potential of revolutionizing medicine. In this talk I will offer an example of how computational models can provide insights into T cell mediated autoimmunity.
T cells, the frontline cells of our adaptive immune system, are characterized by the immense diversity of the antigen binding receptors (T cell receptors or TCRs) they bear. The repertoire of TCRs is forged in the thymus by somatic recombination followed by positive and negative selection. The latter process involves testing the response of the TCR on immature T cells (thymocytes) against self-peptides. Thymocytes that show very weak responses die through neglect (positive selection); those that show very strong responses die through apoptosis (negative selection). Experiments indicate that both selection processes use the same signaling mechanism that filters weak signals down one pathway characterized by a graded response of some key downstream molecules, but sends stronger signals down another pathway that is characterized by digital-like kinetics of the same downstream signaling molecules. We present a new molecular explanation of this phenomenon via a computational signal transduction model, which we also test by experiments. We show that Ras activation in thymocytes is characterized by the presence of a molecular switch due to a positive feedback loop in a Ras-activating enzyme. However we show that an important adaptor protein, LAT, acts as a filter, and allows weaker signals to activate Ras by another pathway present in thymocytes. Our model also yields a new mechanism for digital signaling of the Erk protein in mammalian cells, and has important implications for autoimmunity.