Abstract

The adaptive immune system relies on randomness, in generating a huge and diverse set of lymphocyte receptors through random DNA rearrangements. Randomness occurs also at the level of gene expression, which was shown to be a stochastic process at the molecular level. We study the interplay between randomness at the molecular level and reliable function of the immune system at the cellular and multi-cellular level, focusing on two cases: the T cell receptor repertoire, and differentiation of naïve CD4 T cells. Although the T cell receptor (TCR) repertoire is produced by a random process, studies during the years identified biases in the repertoire, for example the unequal usage of V and J gene segments. However, the mechanisms that govern such biases remain poorly understood. Using a quantitative high-throughput TCR sequencing approach applied to murine T cells, we observe a very high level of similarity in various properties of the repertoire between individual, genetically identical mice. This similarity suggests that common underlying mechanisms determine repertoire structure. We revealed one such mechanism, showing that chromatin conformation at the DJβ genomic locus explains most of the biases observed in Jβ usage. Remarkably, chromatin conformation also explains Jβ usage biases measured previously in human T cells. We demonstrate that as a consequence of these structural and other biases, the TCR repertoire, despite its random and highly diverse nature, contains a surprisingly large number of public sequences that are shared among individuals. We derive a necessary mathematical condition for this surprising finding, which indicates that the TCR repertoire contains a “core” set of receptor sequences that are abundant among individuals. We will present also implications of stochasticity in gene expression for Th1-Th2 differentiation, following experiments in which we mapped cell decisions under mixed input conditions.