Understanding T cell fate selection in Gene Regulatory Network terms
By comprehensive genome-wide screening building on a foundation of years of gene-by-gene measurements, we have characterized the whole set of transcription factors that change expression during the fate selection events of early T cell development. Stage-specific CRISPR/Cas9 acute knockouts have helped us to show that many of these factors beyond those known previously actually modulate the success, speed or direction of T cell development. The time windows for the activities of these factors overlap throughout the early stages of T-cell development. Notably, many have different impacts on gene expression depending on the stage at which they are removed or overexpressed.
Coupling stage-specific perturbation with whole-genome transcriptome and accessible chromatin analyses, we have been able to model the gene regulatory network that guides the cells into the T cell developmental pathway under Notch signaling and to explain the mechanisms involved in shutting down alternative pathway access. Through deeper DNA binding analyses of key transcription factors and the relationships between their genome-wide binding and transcriptional activity, we can learn rules for the activity of factors at the nodes of this regulatory network.
Several notable features distinguish this network from other developmental network models in development. First, some of the most important transcription factors for T cell development can play roles in both T and non-T programs; thus, their expression is necessary but not sufficient to determine T-lineage entry. Second, repression in this network is often "soft", delaying or disfavoring gene expression without silencing it completely. This allows extended developmental plasticity for the cells before they make final choices. Finally, activity of the network is acutely sensitive to small changes in the levels of key transcription factors, with T cell development inhibited both by too little and by too much of a given essential factor. At least one transcription factor we have studied changes the default choices of T cells between different developmental outcomes if its level is reduced by less than twofold during a key stage of T cell development.
How transcription factors interact with the chromatin state epigenome
The unusually high resolution of T cell developmental stages in which we can manipulate transcription factor activities has revealed underlying principles about transcription factor rules of engagement and interaction with chromatin states. Most transcription factors we have studied prefer to bind sites in open chromatin, but this preference is far from absolute. Transcription factor interactions with other transcription factors determine not only the impact of their binding to target sites, but even the choice of genomic sites where the transcription factors bind. Different partners draw the same factors to different sites competitively. Altered doses of a given factor may both change their titration by different partners and change their abilities to gain access to closed chromatin sites. We are exploiting the T cell development system now to learn the rules by which particular factor combinations may or may not remodel chromatin at important regulatory sites to establish new long-term baselines that translate into a new cellular identity.
Major Issues currently under investigation
Causality in the "developmental ratchet": what makes the T cell fate decision irreversible? How do certain batteries of T cell genes get marked to stay on permanently, even when the factors that activated them are turned off? During the developmental process, what determines when chromatin states radically change, and when gene expression changes while keeping chromatin states conserved?
Predictiveness: what determines the speed with which individual cells progress through the T-cell commitment process? Are separate regulators required to make fetal T cells develop much faster than adult T cells? How is the order determined in which multilineage precursors lose access to non-T cell options?
Is the cellular path to T cell identity history-dependent?