T cell-mediated autoimmunity
Certain autoimmune diseases, such as multiple sclerosis and type I diabetes, result from damage caused by T cells to endogenous tissues in specific organs. A major focus of our group is to work on understanding the principles that underlie the misregulation of T cell activation in such organ-specific autoimmune disorders. T cell activation in response to molecular markers of “self” is the result of collective dynamic processes that span multiple length and time scales and involve many cellular and molecular components. Phenomena that occur on large scales (e.g., tissues) influence cooperative molecular events in single cells which, in turn, influence the tissue environment (Fig. 1). This inherent hierarchical cooperativity makes it difficult to intuit the context-dependence of competing interactions between various components and processes. Manipulating a variable in the same way can lead to contradictory consequences (e.g., disease inhibition versus enhancement) in different settings because the relevant mechanisms cannot be parsed in terms of additive or autonomous components. It is very important to understand how individual molecular components in the system work, but predictive capabilities will remain in their infancy until we understand how cooperative phenomena on different scales of time and space regulate molecular and cellular events.

Using the principles of statistical mechanics and harnessing today’s computational capabilities, the hierarchical cooperative processes shown in Fig. 1 can be simulated in a manner that makes such studies productive partners of experiments. Intracellular signaling in response to ligands presented by cells that make up tissue is a key element in the development of a T cell response (Fig. 1), and recent studies [] show that membrane-proximal and intracellular signaling in T cells can be strongly influenced by spatial organization and stochastic fluctuations. These spatially resolved stochastic dynamic events can be simulated on a computer []. In Fig. 2 (central panel), the various colored particles represent intracellular or membrane proteins that can bind other molecules and undergo various biochemical transformations (nucleotide exchange, phosphotransfer, etc.) according to known features of the signaling pathway or hypotheses that need to be tested. Then, the principles of non-equilibrium statistical mechanics (implemented via Monte-Carlo, Langevin, or Gillespie algorithms) allow elucidation of the dynamic consequences of these well-established rules or new hypotheses. Complexity, like cooperative behavior, emerges naturally from these computer simulations. Computational models for the behavior of a single cell interacting with its environment can be linked to simulations of the migration of cells in tissues such as the lymph node, the thymus, or the site of organ – specific autoimmunity (Fig. 2). The stochastic motion of the cells are influenced by biochemical cues in the environment (cytokines, chemokines, etc.) as well as intracellular signaling in response to interactions with molecules presented by tissue cells (e.g., molecular markers of “self” that are ligands for receptors on the T cell). Thus, simulations that probe different length and time scales can be coupled, and the results understood in terms of field-theoretic techniques to study the hierarchically arranged collective dynamic processes. Of course, as in our work on T cell activation, this research is carried out in close synergy with experiments in collaborating immunology laboratories.

Specific issues currently under investigation include: 1] Thymic selection and factors determining the probability of escape of self-reactive T cells. 2] The integration of co-stimulation, cytokine-mediated signaling, and TCR signaling that determines the balance between Th1 and Th2 type cells. 3] Signaling by auto-reactive TCR in the periphery.

T cell activation in response to pathogens