Each organ develops at its own time — usually in the embryo. The discovery of progenitor cells that give rise to two structures in the thymus hints that this immune organ can continue to develop after birth.
The thymus is the source of the mature T cells that fight many types of infection. In this issue, two papers1,2 identify the potential progenitors that enable its development. Rossi et al. (page 988)1 demonstrate that single embryonic epithelial cells contribute to two major structural components of the mature thymus (the medullary and cortical epithelia). An accompanying paper by Bleul et al. (page 992)2 shows that thymus epithelial progenitors continue to be active after birth, and probably generate new thymus epithelium constantly. Remarkably, genetic activation of thymus epithelial progenitors in ‘nude’ mice that do not make a functional thymus, results in the de novo formation of units of functional thymus.
The immature progenitor cells that give rise to T cells are constantly supplied to the thymus from the bone marrow. On entering the thymus, they divide rapidly before undergoing a series of well-defined developmental steps, requiring tight contact with the thymus structure. Specialized thymus epithelial cells (TECs)3,4 nurture the immature T cells by providing growth factors and cell–cell contacts. Moreover, contact of immature T cells with TECs is vital for the selection procedures that ensure that the mature T cells do not react against the tissues of the body itself (which would cause autoimmune disease). TECs come in two major types — medullary TECs (mTECs) and cortical TECs (cTECs) — that differ in their anatomical location in the thymus. During their development, T cells migrate through these distinct regions, and the different TECs have separate functions in T-cell growth, selection and export5,6.
During embryonic development, the thymus arises from the third pharyngeal pouch, which is part of the primitive gut7. Classical histology suggested that the thymus is derived from two separate embryological tissues: the endoderm and the ectoderm. It was originally thought that the typical thymus architecture, with medullary and cortical compartments, results from folding an ectodermal layer around an endodermal layer8. However, Gordon et al.9 suggested that endoderm alone may be sufficient for thymus development.
So, what exactly are the thymus-forming cells? Do thymus stem or progenitor cells exist; and if so, how and when do they contribute to the formation of the thymus? In the classical embryological view, thymus-specific stem cells are not necessarily required. Nevertheless, there has been speculation that epithelial stem cells exist that might give rise to both thymic cortex and thymic medulla cells10,11. When transplanted into adult mice, populations of embryonic or fetal thymus epithelial cells harbour all the components required to generate a functional thymus environment in vivo12,13,14. The presence of both cortex and medulla in such reassembled organs13,14 (and the common origin of cortex and medulla in certain ‘chimaeric’ mice15) has been taken as evidence for thymic epithelial progenitors with the dual potential to make both tissues. With the exception of cell clusters of thymic medulla reported previously as being derived from a single cell12, there have been no assays to analyse the progenitor–product relationship and the developmental potential of single cells.
Taking complementary cellular and genetic routes, Rossi et al.1 and Bleul et al.2 searched for further evidence of thymus stem or progenitor cells during embryonic and adult life. Rossi et al. used single epithelial cells tagged with yellow fluorescent protein (YFP) isolated from embryonic thymus (at day 12) and followed the cells' development in a non-fluorescent ‘foster thymus’ over several weeks in vivo (Fig. 1a). In every case where fluorescent progeny cells were detected, the single embryonic epithelial cell had produced both mTECs and cTECs in the adult thymus. This is consistent with epithelial cells in the day-12 thymus containing a large proportion of progenitor cells with this dual potential. One cell can only come from one embryonic tissue. So, these results (supported by those from Bleul et al. discussed below) show that both cortex and medulla are derived from a single layer — probably the endoderm9.
Bleul et al.2 used complex genetics to examine the role of epithelial progenitors in post-natal thymus development. The fate of the cells was again followed using a YFP tag, but in this case expression of the fluorescent protein was switched on only after birth. The switch occurred randomly and only in very rare epithelial progenitors, akin to a ‘random generator’. The authors' data provide independent evidence that a single epithelial cell can give rise to both types of TEC in vivo, but they also show that this dual potential is maintained in the thymus after birth. Some of the thymi showed fluorescent clusters of only medullary or only cortical epithelial cells. This suggests that, in addition to progenitors with dual potential, there are distinct progenitor cells that are committed to either an mTEC or a cTEC fate. The distance of cortical from medullary epithelial cell clusters indicates that these distinct mTEC or cTEC progenitors migrate away from each other after branching off from the common TEC progenitor1,2,12,15.
In a second set of experiments, Bleul and colleagues attempted the more challenging task of determining whether a single epithelial cell, or very few such cells, can generate a functional thymus. They used the well-known ‘nude’ mouse16 which carries a non-functional copy (FoxN1nu) of the thymus- and skin-specific gene FoxN1. Nude mice have an epithelial cell defect that renders TECs unable to initiate thymus development16,17,18. As a result, the primitive embryonic thymus develops into cysts composed of a wall of immature TECs and lacking any developing T cells.
Bleul et al. constructed a novel nude mouse strain in which the FoxN1 gene was switched on randomly in single postnatal epithelial cell precursors (Fig. 1b). Remarkably, thymus tissue developed in these mice. This implies that single epithelial progenitors are not only able to contribute to some medullary and cortical epithelial cells but to a fully functional cohort of such cells. The authors have gone a long way to show that these ‘neo thymi’ can support the development of T cells bearing a diverse T-cell repertoire, and sufficient in numbers to mount a T-cell-dependent immune response.
Several points are worth considering in the light of these papers. The anatomical position of the non-functional thymus cysts in the chest is far away from the original location of the primitive thymus in the pharyngeal pouch, implying that thymus epithelial progenitors can initiate thymus formation outside their normal embryological context. This is remarkable because many reports have stressed requirements for other cell types for proper thymus development3. It is possible that adult thymus development is independent of such cellular interactions, or that cells surrounding the nude thymus cysts can act in lieu of the embryological counterparts.
Because it was sufficient merely to re-express FoxN1 in adult life to make a thymus, it would seem that FoxN1 could be the master regulatory factor in thymus development. The data suggest that, in the absence of FoxN1 expression, thymus epithelial stem cells become dormant without permanently damaging their viability or developmental potential. Finally, TEC progenitors could continually replace TEC populations in the adult thymus, which might allow the TECs to update the antigens that they use in T-cell selection.
The nature of thymus-forming cells will continue to challenge immunologists because it is likely that thymus epithelial stem cells are not only the cells underlying the formation of the well-known thymus, which is found next to the heart, but are also the origin of the recently noted thymus in the neck19,20. Rossi et al. identified most, if not all, embryonic day-12 thymus epithelial cells as progenitors, and Bleul et al. demonstrated stem or progenitor activity in the adult thymus. However, isolation and better characterization of these cells is required, and it would be interesting to learn whether self-renewing epithelial stem cells or TEC-committed progenitor cells are responsible for the observed thymus development.