T cell differentiation and repertoire selection depend critically on several distinct thymic epithelial cell types, whose lineage relationships are unclear. We have investigated these relationships via functional analysis of the epithelial populations within the thymic primordium. Here, we show that mAbs MTS20 and MTS24 identify a population of cells that, when purified and grafted ectopically, can differentiate into all known thymic epithelial cell types, attract lymphoid progenitors, and support CD4(+) and CD8(+) T cell development in nude mice. In contrast, other epithelial populations in the thymic primordium can fulfill none of these functions. These data establish that the MTS20(+)24(+) population is sufficient to generate a functional thymus in vivo and thus argue strongly that all thymic epithelial cell types derive from a common progenitor cell.

The thymus is the primary organ responsible for generating functional T cells in vertebrates. Although T cell differentiation within the thymus has been an area of intense investigation, the study of thymus organogenesis has made slower progress. The past decade, however, has seen a renewed interest in thymus organogenesis, with the aim of understanding how the thymus develops to form a microenvironment that supports T cell maturation and regeneration. This has prompted modern revisits to classical experiments and has driven additional genetic approaches in mice. These studies are making significant progress in identifying the molecular and cellular mechanisms that control specification, early organogenesis and morphogenesis of the thymus.

The thymus and parathyroids originate from a common primordium that develops from the third pharyngeal pouch in mice and humans. The molecular mechanism that specifies this primordium into distinct organ domains is not known. The Gcm2 and Foxn1 transcription factors are required for development of the parathyroid and thymus respectively, and are attractive candidates for this role. However, their embryonic expression patterns during pharyngeal pouch development and early thymus and parathyroid organogenesis have not been described. Here we report that Gcm2 is expressed specifically in the developing second and third pharyngeal pouches at E9.5, and is further confined to a small domain of the third pouch endoderm by E10.5. In contrast, Foxn1 is not expressed until after the common primordium is formed, beginning at E11.25. Our results show that Gcm2 and Foxn1 expression mark two complementary domains that prefigure parathyroid and thymus regions within the common primordium before morphological distinctions are present.

T cell development depends critically on several distinct thymic epithelial cell types that are organized into two main compartments: cortex and medulla. The prevailing hypothesis suggests that these derive from ectoderm and endoderm, respectively. Here we show that lineage analysis provides no evidence for an ectodermal contribution to the thymic rudiment. We further demonstrate, via ectopic transplantation, that isolated pharyngeal endoderm can generate a functional thymus containing organized cortical and medullary regions and that this capacity is not potentiated by the presence of pharyngeal ectoderm. These data establish that the cortical and medullary thymic epithelial compartments derive from a single germ layer, the endoderm, thus refuting the 'dual-origin' model of thymic epithelial ontogeny.