Gene regulatory logic underlying the differentiation of mixed lineages in human thymic epithelial cells

Abstract

The thymus enacts central tolerance through thymic (TECs), which present self-antigens to developing T cells self-reactive cells, preventing autoimmunity. Human TECs include populations that express lineage-committed programs corresponding to distinct peripheral tissues, yet how they acquire these programs and whether these TEC subpopulations are genetically related and conserved in mice remain unknown. Using matched single-cell gene-expression and chromatin-accessibility profiling of 126,687 human TECs, we identified 17 TEC subsets defined by their distinct expression profiles of 1,843 cell type restricted-antigens (CRAs) from 48 peripheral cell types. With a new phylogenetic reconstruction method, we show that lineage-committed TECs share a common ancestor, retain a thymus-specific mixed identity through co-expression of multiple CRA signatures, and are specified by coordinated activator and repressor regulons. We also identified a basement-membrane-collagen-rich state (BMC-TECs) with progenitor-like properties. Finally, mouse TECs shared the regulatory strategy but conserved few master transcription-factor drivers. This study mechanistically characterizes human TEC subpopulations and reveals the gene regulatory logic that leads to their distinct diversification and mixed identities in humans and mice.