cortical epithelial cell(s)
medullary epithelial cell(s)
re-aggregate thymic organ culture
thymic epithelial stem cell(s)
Stem cell-based therapies hold much promise for the rejuvenation of aged or damaged tissues; however, before such cells can be used therapeutically, they must first be accurately identified. In this issue of the European Journal of Immunology it is reported that MTS24, a marker previously associated with progenitor cells of the thymic epithelium, fails to accurately identify epithelial cell populations with the ability to reconstitute a functional thymus. This finding demonstrates that much progress needs to be made before thymic epithelial stem cells can be harnessed for clinical benefit.
See accompanying article: http://dx.doi.org/10.1002/eji.200737275
Stem cells are defined by two key functional criteria – the ability to self-renew, and the ability to give rise to differentiated progeny of multiple lineages. Tissue-specific stem cells have now been described in a number of different tissues, although by far the best characterized of these is the hematopoietic stem cell (HSC), and the identification of these cells represents an elegant paradigm for stem cell discovery 1.
In 1951, the ability of bone marrow cells to reconstitute the hematopoietic system of an irradiated host suggested the presence of HSC 2, 3. Thereafter, it was quickly realized that bone marrow contained distinct subpopulations of cells capable of giving rise to multiple cell lineages, and the clonal nature of HSC was soon demonstrated 4, 5. Once the existence of an HSC population was proven, attempts to define HSC on the basis of size, density, and surface marker expression eventually lead to the discovery that HSC reside within the lineage-negative (Lin–), Sca-1+, c-kit+ fraction of the (mouse) bone marrow 6, and that single cells with this Lin–Sca-1+c-kit+ phenotype could give rise to all blood cell lineages in conditioned hosts 7, 8. This success in defining the HSC on the basis of cell surface marker expression and functional capacity has spawned interest in identifying other tissue-specific stem cells. In this issue of the European Journal of Immunology (EJI), Rossi et al.9 report that in the case of thymic epithelium, much progress remains to be made before thymic epithelial stem cells (TESC) can be identified on the basis of their surface phenotype.
The thymic epithelium is comprised of two major cell types – medullary and cortical epithelial cells (mTEC and cTEC, respectively) that have distinct functional properties. The idea of a common thymic TESC from which both cTEC and mTEC arise was proposed as early as 1988 10, and two groups have recently provided fresh evidence for the existence of such cells 11, 12. This evidence comes in two forms; firstly, individual thymic epithelial cells from E12 thymi have been demonstrated to give rise to both cTEC and mTEC lineages when injected into intact embryonic thymic lobes 11, and secondly, cell marking experiments have shown individual progenitor cells in the thymus can give rise to TEC colonies composed exclusively of either cTEC or mTEC, as well as colonies containing cells of both epithelial lineages 12. Collectively these results demonstrate that multipotent progenitor cells are present in the thymic epithelium, and may also suggest the existence of distinct cTEC and mTEC progenitors, a conclusion supported by the presence of clonal islets of mTEC in chimeric mice 13. Importantly however, although these studies are consistent with the presence of a TESC, they do not constitute proof that such a cell exists, as the ability of epithelial progenitor cells in the thymus to self-renew (another defining feature of a stem cell) has not yet been demonstrated. Therefore, currently these cells must be viewed as thymic epithelial progenitor cells (TEPC), rather than stem cells, although further investigation may indeed reveal them to be so.
Is it possible to isolate thymic epithelial progenitors on the basis of cell surface markers? In two earlier studies, a surface structure detected by the MTS24 antibody was associated with thymic epithelial progenitor status 14, 15. In this issue of the European Journal of Immunology, however, Rossi et al. 9 challenge this conclusion, by demonstrating that both MTS24+ and MTS24– populations from E12 thymi can generate a functional thymus upon transplant. What then is the reason for these contrasting results? Differences in the sorting strategies and methods of preparing sorted cells for transplant likely explain the conflicting findings. In the study published by Bennet et al.14 E12 thymus cells were sorted into MTS20+MTS24+ and MTS20–MTS24– populations and compared for their ability to generate a thymus on transplant. The problem with this comparison arises from the identity of the MTS20–MTS24– population – Rossi et al. 9 demonstrate here that MTS24 is expressed on most E12 epithelial cells, and therefore suggest that the MTS20–MTS24– population at this time point is composed primarily of other stromal cell types such as fibroblasts, which would not be expected to reconstitute a functional thymus. When a pan-epithelial marker is used in combination with MTS24 to sort MTS24+ and MTS24– epithelial cells from E12 thymi, it is clearly apparent that cells with progenitor capacity are contained in both epithelial cell subsets 9.
Gill et al. 15 also investigated the ability of MTS24 staining to enrich for progenitor cells in E15 embryonic thymi. In this study the ability of MHC2+MTS24+ and MHC2+MTS24– cells to give rise to a functional thymus was compared, and progenitor activity was found to reside exclusively in the MTS24+ population. By gating on MHC2+ cells Gill et al. 15 ensured that both of the transplanted cell populations used in their experiments were of epithelial origin (as at this time point no other cells in the thymus express MHC2), and in light of the findings of Rossi et al.9 it is, therefore, unclear why MHC2+MTS24– cells failed to generate a thymus in this study. One explanation is the total number of cells transplanted. Only 2500 sorted cells were transplanted in the study of Gill et al. 15, while Rossi et al.9 transplanted 100,000 cells. Regardless of the reasons for the differences between the three reports, the finding that MTS24– cells can give rise to a functional thymus means that MTS24 can no longer be considered a universal marker of TESC/TEPC. While this is a disappointing setback, a perhaps even greater challenge lays ahead – the demonstration of self-renewal of thymic epithelial stem cells. Unlike the situation with HSC, serial transplantation of thymic epithelial stem cells may not be readily demonstrable with re-aggregate thymic organ culture (RTOC).
The Rossi et al.9 paper outlines an important limitation of this assay. When MTS24+ cells were sorted from older thymi (E16), they failed to re-aggregate efficiently in vitro, and therefore the progenitor capacity of these cells could not be readily determined. Furthermore, Bennett et al. 14 reported in their study that whole transplanted thymic lobes were more efficient at supporting T cell development than thymic lobes which had been disaggregated and re-aggregated. Therefore, RTOC-based transplantation systems may have important limitations for the identification of TESC, as the failure of sorted cells to give rise to a functional thymus on transplant may result from a failure to generate the correct microenvironment (niche) in which progenitor cells can fulfill their potential, rather than an absence of a stem cell population. One way of overcoming these limitations may be an assay recently described in a previous study by Rossi et al. 11, in which single E12 thymic epithelial cells are microinjected into an intact fetal thymic lobe, which is then transplanted under the kidney capsule as a whole organ. This assay is feasible with only small numbers of individual cells, and may therefore be more sensitive than RTOC-based approaches for identifying stem cell populations. One potential shortcoming of this assay however, is that unlike the RTOC-based assays it may be difficult to determine the functional attributes of cells derived from transplanted TESC, due to the presence of a host epithelial cell population.
For such a single cell reconstitution experiment to succeed, new methods for prospectively isolating TESC/TEP will need to be developed. Now that MTS24 can no longer be considered an exclusive marker for progenitor cells in the thymus, new sorting strategies will need to be devised. Encouragingly, enrichment of mammary 16 and prostate 17 epithelial stem cells to frequencies of 1/65 and 1/35, respectively, has recently been achieved, demonstrating that prospective isolation of stem cells from epithelial tissues can be achieved on the basis of surface marker expression. It is possible that markers shared by both mammary and prostate stem cells, such as CD49f (integrin-α6), may be of use to home in on TESC populations. Additionally, the expression of transcription factors associated with “stemness” such as Nanog and Oct4 18 may be of use in identifying TESC populations. Indeed, Nanog and Oct4 have both been observed in adult stem cell populations 19, and transcripts for each of these molecules have been detected in thymic epithelial cells 20.
The ability to halt or reverse declining thymic output would be an attractive therapeutic option in a number of clinical settings, including age-associated involution of the thymus. TESC are likely key to such efforts, and indeed have the attractive feature that they are able to initiate thymopoiesis in a self-organizing fashion, at least under favourable conditions 12. Additionally, given the importance of the thymic epithelium in central tolerance, the ability to identify and manipulate TESC may open up new avenues for the treatment of autoimmune disease. The Rossi et al.9 paper is a timely reminder of the long road ahead.