Absence of thymus crosstalk in the fetus does not preclude hematopoietic induction of a functional thymus in the adult

Transcription

1 DOI /eji Cellular immune response 2395 Absence of thymus crosstalk in the fetus does not preclude hematopoietic induction of a functional thymus in the adult Natalie A. Roberts 1, Guillaume E. Desanti 1, David R. Withers 1 Hamish R. Scott 2, William E. Jenkinson 1, Peter J. L. Lane 1 Eric J. Jenkinson 1 and Graham Anderson 1 1 MRC Centre for Immune Regulation, Institute for Biomedical Research, University of Birmingham UK 2 Division of Molecular Pathology, Institute of Medical and Veterinary Science and The Hanson Institute, Adelaide, South Australia, Australia Cortical and medullary thymic epithelial cells provide essential signals for a normal programme of T-cell development. Current models of thymus development suggest that thymocyte-derived signals play an important role in establishing thymic microenvironments, a process termed thymus crosstalk. Studies on CD3etg26 mice lacking intrathymic T-cell progenitors provided evidence that normal development of the thymic cortex depends upon thymocyte-derived signals. Importantly, the reported failure to effectively reconstitute adult CD3etg26 mice raised the possibility that such crosstalk must occur within a developmental window, and that closure of this window during the postnatal period renders thymic epithelium refractory to crosstalk signals and unable to effectively impose T-cell selection. We have re-investigated the timing of provision of crosstalk in relation to development of functional thymic microenvironments. We show that transfer of either fetal precursors or adult T-committed precursors into adult CD3etg26 mice initiates key parameters of successful thymic reconstitution including thymocyte development and emigration, restoration of cortical and medullary epithelial architecture, and establishment of thymic tolerance mechanisms including maturation of Foxp3 1 Treg and autoimmune regulator-expressing medullary epithelium. Collectively, our data argue against a temporal window of thymocyte crosstalk, and instead demonstrates continued receptiveness of thymic epithelium for the formation of functionally competent thymic microenvironments. Key words: Developmental Immunology. Lymphoid Organs. Thymopoiesis Introduction Correspondence: Professor Graham Anderson g.anderson@bham.ac.uk The generation and selection of self-mhc-restricted T cells occurs within the thymus as a result of complex interactions with the thymic microenvironment. The thymus is anatomically compartmentalised into cortical and medullary regions, each containing specialised epithelial cell types that regulate positive and negative selection of the developing TCR repertoire [1]. Cortical thymic epithelial cells (ctec), characterised by expression of MHC class I and class II antigens and the proteosomal subunit b5t, are effective mediators of positive selection of CD4 1 CD8 1 thymocytes [2, 3]. In the thymic medulla, medullary thymic epithelial cells (mtec), including the subset that expresses the autoimmune regulator (Aire) gene, play a key role in purging newly

2 2396 Natalie A. Roberts et al. selected CD4 1 and CD8 1 thymocytes of potentially autoreactive TCR specificities [4, 5]. Thus, the generation of distinct and functionally competent ctec and mtec is key to the establishment of T-cell development and tolerance. It is now clear that ctec and mtec subsets arise during thymus ontogeny from a bipotent TEC progenitor population [6, 7]. Indeed, the development of bipotent progenitors towards the ctec and mtec lineages in the embryonic thymus occurs simultaneously with development of the first wave of developing T-cell precursors. This temporal pattern lends support to the idea that interactions between TEC and thymocytes leads to reciprocal signalling that is required for normal development of both cell types, a process termed thymus crosstalk [8, 9]. The importance of thymus crosstalk can be seen from studies analysing the thymic microenvironments of mice in which T-cell development is blocked at various stages. These studies revealed deficiencies in TEC development, most notably in relation to formation of the thymic medulla [9, 10]. Indeed, a molecular basis of crosstalk has recently been revealed, with RANKL expression by thymic haemopoeitic cells such as CD4 1 thymocytes and lymphoid tissue inducer cells being identified as an important regulator of mtec development [11 14]. Thus, formation of an intact medulla for the establishment of T-cell tolerance depends upon interactions between thymus-resident haemopoietic cells and mtec progenitors. Paralleling these studies on the thymic medulla, crosstalk-dependent mechanisms have also been reported to be required for establishment of the thymic cortex and ctec development. Thus, mice lacking intrathymic T-cell precursors as a result of insertion of multiple copies of a human CD3e transgene were shown to have severely disorganised cortical thymic architecture [15]. However, unlike mice harbouring blocks in T-cell development at later stages (e.g. Rag-1-deficient mice), restoration of thymic architecture and TEC development in CD3etg26 mice appeared to involve a temporally restricted window [15, 16]. Thus, it was proposed that induction of the thymic cortex required interactions between fetal stromal cells and early T-cell precursors. We [17] and others [18] have shown that initial specification of bipotent TEC progenitors into the ctec and mtec lineages can occur independently of thymocyte crosstalk, suggesting that any role for such signals is in the maintenance of progenitor population competence and/or survival. Of equal importance, it was also reported that attempts to reconstitute adult CD3etg26 mice after closure of this window resulted in an inflammatory colitis-like disease [19], suggesting failure of the thymus to support effective T-cell selection and/or failure of mechanisms that normally impose T-cell tolerance. Such observations are important as they suggest potential limitations for strategies aimed at boosting thymic function, for example in the elderly or following ablative therapy, where restoration of T-cell production and maintenance of self-tolerance may be compromised as a result of long periods of thymic inactivity. In this study, we have analysed the ability of the adult thymus to undergo reconstitution following the continued lack of thymocyte crosstalk in the fetal, neonatal and post-weaning periods. We show that an absence of T-cell precursors during these stages does not render the thymic epithelial microenvironment refractory to normal reconstitution. Importantly, we also show that establishment of thymic microenvironments is accompanied by the development of cellular mediators of T-cell tolerance, including Foxp3 1 Treg and Aire-expressing mtec. In conclusion, our data argue against a developmentally regulated window for the establishment of thymic microenvironments and instead demonstrate that thymic epithelial cells retain their receptivity to thymocyte crosstalk. Results Absence of fetal/neonatal thymus crosstalk does not prevent thymic reconstitution in adult mice CD3etg26 mice represent an effective mouse model to study the role of crosstalk in thymic epithelial cell development. Presumably as a consequence of transgene integration, these mice lack T-cell progenitors, with the thymus instead harbouring B cells [20]. To investigate whether thymus crosstalk in the developing thymus during the fetal/neonatal periods is required for successful thymic reconstitution in the adult, we injected nonirradiated adult CD3etg26 mice with syngeneic Lineage-negative fetal liver (Lin FL) progenitors from WT E15 mouse embryos to provide a source of T-cell progenitors. At the indicated time point, typically 9 wk post reconstitution, mice were sacrificed and lymphoid tissues harvested for analysis. In contrast to the thymus of uninjected age-matched control mice that remained small, cystic and hypotrophic (Fig. 1A), the thymus of CD3etg26 mice receiving fetal liver progenitors grew in size, did not contain cysts and showed an increase in cellularity (Fig. 1A). When analysed for evidence of T-cell development, CD3etg26 mice receiving Lin FL showed a pattern of CD4, CD8 and abtcr expression similar to that of WT thymocytes, including the generation of CD4 1 CD8 1 thymocytes and mature CD4 1 CD8 and CD4 CD8 1 subsets expressing high levels of abtcr (Fig. 1B). In addition, analysis of secondary lymphoid tissue of reconstituted CD3etg26 mice showed successful recruitment of T cells to the normally T-celldeficient T zone of the spleen, demonstrating that T-cell production in the thymus is accompanied by thymic emigration and effective T-cell homing in the periphery (Fig. 2). Thus, although the thymic microenvironment of adult CD3etg26 mice has been deprived of thymocyte crosstalk during thymus development and in adult life, it retains its receptivity to thymus crosstalk signals and is capable of supporting a full programme of T-cell development from fetal liver progenitors, including progenitor recruitment and differentiation culminating in the appearance of mature T-cell subsets in peripheral lymphoid organs. One possibility to explain the efficacy of thymus reconstitution in adult CD3etg26 mice by fetal liver progenitors is they are specialised in their ability to promote effective thymus crosstalk because of their ability to generate fetal-specific cell types such as fetal gd T cells [21] and fetal lymphoid

3 tissue inducer cells [22]. To investigate this possibility, we isolated CD3e CD4 CD8 CD251CD44 triple negative 3 (TN3) progenitors, representing T-lineage-restricted cells [23], from Cellular immune response adult WT thymus, and transferred them intravenously into adult CD3etg26 mice. Figure 1C shows that adult thymic TN3 cells are also able to successfully reconstitute the thymus of adult CD3etg26 mice including an increase in thymic cellularity (data not shown), and generation of CD41CD81 thymocytes and CD41 CD8 abtcrhi and CD4 CD81abTCR1 cells, the latter also being capable of undergoing thymic export and migration to the splenic T-zone (data not shown). Thus, specific signals from fetal progenitors are not required to rescue thymic function following a prolonged lack of crosstalk. Establishment of cortical and medullary intrathymic microenvironments in adult CD3etg26 mice To investigate whether the programme of T-cell development occurring in the thymus of adult CD3etg26 reconstituted mice is accompanied by changes in thymic epithelial microenvironments, we compared thymic architecture in uninjected and fetal liverinjected CD3etg26 mice. As previously reported [15], thymic epithelial cells in control CD3etg26 mice are disorganised, and attempts to define ctec and mtec subsets using antibodies to keratin-8 and keratin-5 shows little Keratin-81 cortical and Keratin-51 medullary separation using these markers (Fig. 3A). In marked contrast, analysis of CD3etg26 mice 9 wk after fetal liver reconstitution showed clearly defined keratin-81 cortical areas and keratin-51 medullary areas separated by a distinct cortico-medullary junction (Fig. 3A). Moreover, thymocyte compartmentalisation successfully occurs in reconstituted mice, with CD41CD81 cells present within cortical areas, and more mature CD41 and CD81 subsets residing in medullary areas (Fig. 3B). Interestingly, when we analysed thymic microenvironments of control and reconstituted mice with the ctec and mtec markers CD205 and CD40 [2], we found a different pattern of staining as compared with using keratin-8 and keratin-5 to identify ctec and mtec. Thus, discrete CD2051CD40 and CD205 CD401 areas were found to be present in unreconstituted mice (Fig. 3C), which may represent rudimentary cortical and medullary microenvironments that are devoid of thymocytes. This difference in keratin and CD40/CD205 staining raises the interesting possibility that patterns of keratin expression in TEC subsets are more reflective of the physical organisation of epithelial cells, rather than their lineage commitment. Collectively, these findings suggest that the emergence of cortical and medullary epithelial lineages occurring during embryonic thymus development independently of thymocyte crosstalk [17, 18] can persist in the adult thymus despite the prolonged absence of thymocytes. Figure 1. Successful thymic reconstitution does not require prior thymocyte crosstalk during the fetal period. Thymuses from unreconstituted and fetal liver-reconstituted (9 wk post injection) were analysed grossly (A) (arrow: cystic structure of unreconstituted thymus) and by flow cytometry for expression of CD4, CD8 and abtcr (B). (C) Successful reconstitution of T-cell development in CD3etg26 mice, 16 days after i.v. injection of adult TN3 thymocytes. Data in (A) mean1sd of four independent experiments. Scale bars in (A) indicate 1 mm. Restoration of thymic function includes generation of cellular mechanisms of T-cell tolerance Earlier studies demonstrating a failure of normal thymic reconstitution in adult CD3etg26 mice also reported, around 5 8 wk after transplant, the onset of an inflammatory wasting 2397

4 2398 Natalie A. Roberts et al. colitis-like disease involving colonic infiltration by activated CD41 T cells [19]. Such findings suggested a failure of normal mechanisms of T-cell tolerance, perhaps as a result of defective generation of intrathymic microenvironments. To investigate this possibility, we analysed two aspects of thymus reconstitution relevant to T-cell tolerance, namely generation of mtec expressing Aire, and Foxp31 Treg [24]. Analysis of mice at time points up to 9 wk after reconstitution with either Lin FL (Fig. 4A) or adult TN3 thymocytes (data not shown), showed that, in contrast to unreconstituted mice (data not shown and [25]), keratin-51 Aire1 mtec were readily detectable in thymus sections of reconstituted adult CD3etg26 mice (Fig. 4). In addition, flow cytometric and confocal analysis showed the presence of CD41Foxp31 Treg in both the thymus and spleen of reconstituted mice (Fig. 4B and C) that were absent in unreconstituted mice (data not shown). In addition, we found that reconstituted adult CD3etg26 mice were healthy, and showed no signs of disease or weight loss at any time after Lin FL reconstitution (latest time point analysed 9 wk, data not shown). Moreover, analysis of colon tissue from reconstituted mice demonstrated no pathology (latest time point after reconstitution analysed 9 wk, data not shown). Thus, in our hands, thymic reconstitution of adult mice that lacked thymocyte crosstalk in the fetal/neonatal periods is accompanied by the establishment of cellular mechanisms of T-cell tolerance and is not associated with inflammatory disease. Discussion Figure 2. Thymic reconstitution is accompanied by thymic emigration and migration to the T-zone of secondary lymphoid tissue. Splenic tissue from CD3etg26 mice, 9 wk after fetal liver reconstitution, was analysed by flow cytometry for expression of CD19, CD3, CD4 and CD8. Note the presence of CD31CD41 and CD31CD81 T cells in reconstituted but not control CD3etg26 mice. Confocal analysis of frozen tissue sections demonstrates the presence of CD31 T cells within the CCL211 splenic T zone. Note the expression of CCL21 and the presence of CD41 CD3 cells in unreconstituted mice, the latter representing lymphoid tissue inducer cells [34]. Scale bars represent 100 mm. Data are representative of at least four reconstituted mice. The development of bipotent thymic epithelial cells to form functionally mature cortical and medullary epithelial microenvironments is essential for T-cell production and the establishment of tolerance. While discrete subsets of TEC are beginning to be defined, little is known about the mechanisms that regulate their development. Several studies have highlighted the importance of thymus crosstalk in TEC development, a process in which haemopoietic cells including thymocytes influence the development of thymic microenvironments. Importantly, current models of thymus development and crosstalk include data from studies on mice in which thymic T-cell progenitors are absent, such as adult CD3etg26 mice. Such studies suggested that contact between fetal stromal cells and developing thymocytes is essential for further thymic development, particularly the thymic cortex [15,16,19]. Furthermore, as this window was reported to be developmentally regulated and limited to the fetal period, evidence was provided indicating that failure of thymus crosstalk early in thymus development renders the thymus refractory to crosstalk signals at later stages and incapable of supporting effective T-cell development and selection [15,16,19]. Here, and in contrast to these earlier studies, we show that adult CD3etg26 mice, despite being deprived of thymocyte crosstalk signals during the fetal, neonatal and adult periods, undergo successful thymic reconstitution as indicated by the production and thymic emigration of both conventional and Foxp31 Treg, and the

5 Cellular immune response Figure 3. Restoration of normal thymic architecture accompanies thymic reconstitution of adult CD3etg26 mice. Thymus sections from unreconstituted and fetal liver-injected (9 wk after transfer) were analysed for ctec and mtec development and organisation using the markers keratin-5 and keratin-8 (A), CD40 and CD205 (C), together with analysis of thymocyte organisation using CD4 and CD8 (B). Note the restoration of cortical and medullary architecture in reconstituted mice. in (A) indicate thymic cysts. Scale bars represent 50 mm. Data are representative of at least four reconstituted mice. establishment of functional cortical and medullary microenvironments including the Aire1 mtec subset that is important for T-cell tolerance. To date, we have injected progenitors into 13 adult mice, aged up to 9 wk at the time of cell transfer, and observed successful thymus reconstitution in 12. Despite using mice at similar ages and similar time points of analysis, the reasons for the discrepancies between our study and these earlier studies are currently not clear. In addition to successful reconstitution, we have not observed any symptoms of inflammation or colitis in mice at 9 wk post reconstitution (data not shown). While our study involved transfer of either fetal liver progenitors or adult thymocytes without preconditioning, earlier studies involved transfer of T-depleted bone marrow following either 5-fluorouracil or irradiation treatment, raising the possibility that damage to the thymic microenvironment caused by conditioning, as has been reported in other studies [26 28], could perturb thymic reconstitution and promote inflammation. Alternatively, while our experiments involved reconstitution with syngeneic (C57BL/6) lymphoid progenitors, earlier studies involved an allogeneic donor host combination where T-depleted marrow from C57BL/6xCBA/J (H-2bxk) F1 mice was transferred into H-2k mice, where residual T cells in the F1 inoculum could be related to the experimental outcome. Whatever the case, our data argue against a developmental window for thymic reconstitution, and instead demonstrate that thymic epithelial cells remain receptive to thymocyte crosstalk signals up until adulthood. While it is clear that CD3etg26 mice are devoid of T-cell progenitors, the thymus of these mice contains significant numbers of B cells ([20], and data not shown), and it is not clear whether B cells are capable of influencing TEC microenvironments in vivo. However, it should be noted that B cells have recently been shown to fail to induce the three-dimensional organisation of TEC in vitro [29]. It is also important to note that the findings from our studies on CD3etg26 mice are in line with an earlier study of Rodewald and Fehling [30], who reported a lack of requirement for fetal thymus crosstalk in c-kit / gc / 2399

6 2400 Natalie A. Roberts et al. mice where thymocytes and T-cell development are virtually absent [31], and with human studies documenting T-cell reconstitution of SCID patients following the postnatal period [32, 33]. We believe these observations have implications for current models of thymocyte crosstalk in development of thymic microenvironments, and also for strategies aimed at rejuvenating thymus function and T-cell production following extended periods of thymic hypotrophy. Materials and methods Mice All mice were housed under specific pathogen free conditions at the Biomedical Services Unit, University of Birmingham under UK Home Office guidelines. CD3etg26 mice were obtained from The Jackson Laboratories, and used for reconstitution between 5 and 9 wk old. For the generation of timed C57/Bl6 embryos, the day of detection of the vaginal plug was designated as day zero. Antibodies and flow cytometry The following antibodies were used for flow cytometric analysis (all obtained from ebioscience unless stated otherwise): anti- CD4PECy7 (clone GK1.5), anti-cd8fitc (clone ), anti-tcrb PE (clone H57-597), anti-cd19apc (clone 1D3, BD Pharmingen), anti-cd3e PE (clone 145-2C11, BD Pharmingen), anti-foxp3 FITC (clone FJL-16s, BD Pharmingen). Intracellular staining for Foxp3 was performed according to the manufacturer s instructions. Streptavidin-conjugated PE was from BD Pharmingen. Flow cytometry was performed on a BD-LSR dual laser machine, with forward and side scatter gates set to exclude non-viable cells. Cell preparations and transfers Figure 4. Cellular aspects of T-cell tolerance following thymic reconstitution. (A) Frozen thymus sections of reconstituted (9 wk after transfer of fetal liver) and WT mice were analysed for the presence of keratin-5 1 Aire 1 mtec. (B) The presence of CD4 1 Foxp3 1 Treg was also determined by confocal analysis of thymus and spleen sections from reconstituted CD3etg26 mice. (C) Flow cytometric analysis of thymocytes from reconstituted CD3etg26 mice for Foxp3 expression, gated on CD4 1 CD8 thymocytes. Note the presence of Aire 1 mtec and Foxp3 1 Treg in reconstituted mice, both of which are absent from unreconstituted mice (data not shown). Scale bars represent 10 mm. Data are representative of at least four reconstituted mice. Fetal liver progenitors were prepared from E15 C57/Bl6 mouse embryos by sorting lineage-negative cells using a cocktail of antibodies to the following: TER119, Gr-1, B220, CD3e (all BD Pharmingen). Adult CD3e CD4 CD8 CD25 1 CD44 thymocytes were prepared from mechanically dissociated 4 6 wk old adult thymus lobes using the following antibodies (all BD Pharmingen): CD3e-PE (clone: 145-2C11), CD4-FITC (clone: GK1.5), CD8a-FITC (clone: ), CD44-PEcy7 (clone: IM7), CD25-APC (clone: 7D4). All sorting was performed on a MoFlo cell sorter with forward/side scatter gates set so as to exclude non-viable cells. For cell transfers, freshly purified cells (greater than 95% purity, data not shown) were resuspended in a 200 ml volume of PBS and injected into the tail vein of adult CD3etg26 mice (aged up to 9 wk old) without prior irradiation or conditioning. Up to adult TN3 or Lin FL cells were transferred in a single injection per animal. Lymphoid tissues were harvested from

7 Cellular immune response 2401 injected and control mice, days (TN3 injection) and 5 9 wk (Lin FL injection) after reconstitution, and analysed by confocal microscopy and flow cytometry. Confocal microscopy Confocal analysis was performed as described previously [34] using the following additional antibodies: anti-cd40 biotin (clone 3/23, BD Pharmingen), anti-cd205 (clone NLDC-145, Serotec), rabbit anti-keratin-5 (Covance), biotinylated mouse anti-keratin-8 (Progen), anti-aire [35], anti-b220 biotin (RA3-6B2, BD Pharmingen), anti-cd4 Alexa-Fluor 647 (clone GK1.5, BD Pharmingen) anti-cd3 FITC (clone 145-2C11, BD Pharmingen), anti-foxp3 FITC (BD Pharmingen). CCL21 was detected using a four-step amplification staining method involving sequential incubations of goat antimouse CCL21 (R&D Systems), donkey anti-goat-fitc (Jackson ImmunoResearch Laboratories), rabbit anti-fitc (BioSource International), and goat anti-rabbit-fitc (Southern Biotechnology Associates). Additional second steps were streptavidin Alexa-Fluor 555 (Invitrogen), anti-rabbit AlexaFlour 647 (Invitrogen). Acknowledgements: This work was supported by an MRC programme grant to E.J.J. and G.A. We thank Roger Bird for cell sorting, and the Biomedical Research Unit and Birmingham University for provision of animals. Conflict of interest: The authors declare no financial or commercial conflict of interest. References 1 Kyewski, B. and Klein, L., A central role for central tolerance. Annu. Rev. Immunol : Shakib, S., Desanti, G. E., Jenkinson, W. E., Parnell, S. M., Jenkinson, E. J. and Anderson, G., Checkpoints in the development of thymic cortical epithelial cells. J. Immunol : Murata, S., Sasaki, K., Kishimoto, T., Niwa, S., Hayashi, H., Takahama, Y. and Tanaka, K., Regulation of CD8 1 T cell development by thymusspecific proteasomes. Science : Mathis, D. and Benoist, C., Aire. Annu. Rev. Immunol : Anderson, M. S., Venanzi, E. S., Klein, L., Chen, Z., Berzins, S. P., Turley, S. J., von Boehmer, H. et al., Projection of an immunological self shadow within the thymus by the aire protein. Science : Rossi, S. W., Jenkinson, W. E., Anderson, G. and Jenkinson, E. J., Clonal analysis reveals a common progenitor for thymic cortical and medullary epithelium. Nature : Bleul, C. C., Corbeaux, T., Reuter, A., Fisch, P., Monting, J. S. and Boehm, T., Formation of a functional thymus initiated by a postnatal epithelial progenitor cell. Nature : van Ewijk, W., Shores, E. W. and Singer, A., Crosstalk in the mouse thymus. Immunol. Today : Shores, E. W., Van Ewijk, W. and Singer, A., Disorganization and restoration of thymic medullary epithelial cells in T cell receptor-negative scid mice: evidence that receptor-bearing lymphocytes influence maturation of the thymic microenvironment. Eur. J. Immunol : Palmer, D. B., Viney, J. L., Ritter, M. A., Hayday, A. C. and Owen, M. J., Expression of the alpha beta T-cell receptor is necessary for the generation of the thymic medulla. Dev. Immunol : Rossi, S. W., Kim, M. Y., Leibbrandt, A., Parnell, S. M., Jenkinson, W. E., Glanville, S. H., McConnell, F. M. et al., RANK signals from CD4(1)3(-) inducer cells regulate development of Aire-expressing epithelial cells in the thymic medulla. J. Exp. Med : Irla, M., Hugues, S., Gill, J., Nitta, T., Hikosaka, Y., Williams, I. R., Hubert, F. X. et al., Autoantigen-specific interactions with CD4 1 thymocytes control mature medullary thymic epithelial cell cellularity. Immunity : Hikosaka, Y., Nitta, T., Ohigashi, I., Yano, K., Ishimaru, N., Hayashi, Y., Matsumoto, M. et al., The cytokine RANKL produced by positively selected thymocytes fosters medullary thymic epithelial cells that express autoimmune regulator. Immunity : Akiyama, T., Shimo, Y., Yanai, H., Qin, J., Ohshima, D., Maruyama, Y., Asaumi, Y. et al., The tumor necrosis factor family receptors RANK and CD40 cooperatively establish the thymic medullary microenvironment and self-tolerance. Immunity : Hollander, G. A., Wang, B., Nichogiannopoulou, A., Platenburg, P. P., van Ewijk, W., Burakoff, S. J., Gutierrez-Ramos, J. C. and Terhorst, C., Developmental control point in induction of thymic cortex regulated by a subpopulation of prothymocytes. Nature : van Ewijk, W., Hollander, G., Terhorst, C. and Wang, B., Stepwise development of thymic microenvironments in vivo is regulated by thymocyte subsets. Development : Jenkinson, W. E., Rossi, S. W., Jenkinson, E. J. and Anderson, G., Development of functional thymic epithelial cells occurs independently of lymphostromal interactions. Mech. Dev : Klug, D. B., Carter, C., Gimenez-Conti, I. B. and Richie, E. R., Cutting edge: thymocyte-independent and thymocyte-dependent phases of epithelial patterning in the fetal thymus. J. Immunol : Hollander, G. A., Simpson, S. J., Mizoguchi, E., Nichogiannopoulou, A., She, J., Gutierrez-Ramos, J. C., Bhan, A. K. et al., Severe colitis in mice with aberrant thymic selection. Immunity : Tokoro, Y., Sugawara, T., Yaginuma, H., Nakauchi, H., Terhorst, C., Wang, B. and Takahama, Y., A mouse carrying genetic defect in the choice between T and B lymphocytes. J. Immunol : Havran, W. L. and Allison, J. P., Developmentally ordered appearance of thymocytes expressing different T-cell antigen receptors. Nature : Mebius, R. and Akashi, K., Precursors to neonatal lymph nodes: LT beta 1 CD45 1 CD4 1 CD3 cells are found in fetal liver. Curr. Top. Microbiol. Immunol : Ceredig, R. and Rolink, T., A positive look at double-negative thymocytes. Nat. Rev. Immunol : Ramsdell, F. and Ziegler, S. F., Transcription factors in autoimmunity. Curr. Opin. Immunol : Zuklys, S., Balciunaite, G., Agarwal, A., Fasler-Kan, E., Palmer, E. and Hollander, G. A., Normal thymic architecture and negative selection are associated with Aire expression, the gene defective in the autoimmune-

8 2402 Natalie A. Roberts et al. polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED). J. Immunol : Adkins, B., Gandour, D., Strober, S. and Weissman, I., Total lymphoid irradiation leads to transient depletion of the mouse thymic medulla and persistent abnormalities among medullary stromal cells. J. Immunol : Chung, B., Barbara-Burnham, L., Barsky, L. and Weinberg, K., Radiosensitivity of thymic interleukin-7 production and thymopoiesis after bone marrow transplantation. Blood : Rossi, S., Blazar, B. R., Farrell, C. L., Danilenko, D. M., Lacey, D. L., Weinberg, K. I., Krenger, W. and Hollander, G. A., Keratinocyte growth factor preserves normal thymopoiesis and thymic microenvironment during experimental graft-versus-host disease. Blood : Masuda, K., Germeraad, W. T., Satoh, R., Itoi, M., Ikawa, T., Minato, N., Katsura, Y. et al., Notch activation in thymic epithelial cells induces development of thymic microenvironments. Mol. Immunol : Rodewald, H. R. and Fehling, H. J., Molecular and cellular events in early thymocyte development. Adv. Immunol : Rodewald, H. R., Ogawa, M., Haller, C., Waskow, C. and DiSanto, J. P., Pro-thymocyte expansion by c-kit and the common cytokine receptor gamma chain is essential for repertoire formation. Immunity : Cavazzana-Calvo, M. and Fischer, A., Gene therapy for severe combined immunodeficiency: are we there yet? J. Clin. Invest : Speckmann, C., Pannicke, U., Wiech, E., Schwarz, K., Fisch, P., Friedrich, W., Niehues, T. et al., Clinical and immunologic consequences of a somatic reversion in a patient with X-linked severe combined immunodeficiency. Blood : Withers, D. R., Kim, M. Y., Bekiaris, V., Rossi, S. W., Jenkinson, W. E., Gaspal, F., McConnell, F. et al., The role of lymphoid tissue inducer cells in splenic white pulp development. Eur. J. Immunol : Hubert, F. X., Kinkel, S. A., Webster, K. E., Cannon, P., Crewther, P. E., Proeitto, A. I., Wu, L. et al., A specific anti-aire antibody reveals aire expression is restricted to medullary thymic epithelial cells and not expressed in periphery. J. Immunol : Abbreviations: Aire: autoimmune regulator ctec: cortical thymic epithelial cells Lin FL: lineage-negative fetal liver mtec: medullary thymic epithelial cells TN: triple negative Full correspondence: Professor Graham Anderson, Floor 4, Institute for Biomedical Research, Medical School, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK Fax: g.anderson@bham.ac.uk Received: 9/4/2009 Revised: 12/5/2009 Accepted: 8/6/2009