Thymic epithelium provides an essential cellular substrate for T cell development and selection. Gradual age-associated thymic atrophy leads to a reduction in functional thymic tissue and a decline in de novo T cell generation. Development of strategies tailored toward regeneration of thymic tissue provides an important possibility to improve immune function in elderly individuals and increase the capacity for immune recovery in patients having undergone bone marrow transfer following immunoablative therapies. In this study we show that restriction of the size of the functional thymic epithelial progenitor pool affects the number of mature thymic epithelial cells. Using an embryo fusion chimera-based approach, we demonstrate a reduction in the total number of both embryonic and adult thymic epithelium, which relates to the initial size of the progenitor cell pool. The inability of thymic epithelial progenitor cells to undergo sufficient compensatory proliferation to rescue the deficit in progenitor numbers suggests that in addition to extrinsic regulation of thymus growth by provision of growth factors, intrinsic factors such as a proliferative restriction of thymic epithelial progenitors and availability of progenitor cell niches may limit thymic epithelial recovery. Collectively, our data demonstrate an important level of regulation of thymic growth and recovery at the thymic epithelial progenitor level, providing an important consideration for developing methods targeted toward inducing thymic regeneration.
Age-related changes in the immune system contribute to an overall reduction in immune responsiveness, leading to increased susceptibility to infectious disease (1). Of these changes, age-related thymus atrophy has a profound effect on the continued generation of naive T cells and results in a peripheral T cell pool containing dominant memory T cell clones, a process that limits diversity and ultimately reduces the capacity for vaccination-based strategies (2, 3, 4, 5). Additionally, the capacity for T cell-mediated immune recovery following ablative therapies and bone marrow transfer declines with age in humans (6), most likely occurring as a result of a reduced capacity of the aged thymus to process T cell progenitors. Several lines of evidence indicate that a reduction in the epithelial compartment of the aged thymus is a major factor in its ability to efficiently process and select mature T cells. For example, a reduction in the proportion of medullary thymic epithelium essential for negative selection of autoreactive T cell clones occurs with increasing age (7), while limiting the availability of thymic epithelial cell niches directly affects thymic T cell production (8). Importantly, while the regenerative capacity of the aged thymus can be revealed in a variety of models (9, 10), neither the mechanisms regulating this process nor the cell types being targeted are fully understood.
Against this background, it is important to understand the mechanisms regulating the growth and size of the thymic epithelial compartment and its potential for regeneration following aging or damage. A key factor in this is defining the nature and persistence of epithelial stem or progenitor cell activity throughout the life of the thymus. During initial thymus organogenesis in the murine embryo, both cortical and medullary epithelial lineages are known to arise from a common progenitor population exclusively derived from the endodermal germ layer (11, 12, 13). Within the thymic primordium, initial differentiation of thymic epithelium from a bipotent progenitor to cortical and medullary lineages is dependent on the action of the transcription factor FoxN1, and this can occur independently of thymocyte interactions within the embryo (14, 15). Additionally, mice bearing a hypomorphic allele of FoxN1 have demonstrated an ongoing role for this transcription factor in controlling thymocyte-dependent stages of thymic epithelial maturation and maintenance (16, 17). However, the ongoing persistence and identity of progenitor activity at later developmental stages into the adult remain unclear (18). Recently, two different mechanisms regulating organ size and regenerative capacity have been defined and shown to apply to different organs, even when such organs are of the same embryonic germ layer origin (19). Thus, in liver, reducing the epithelial progenitor pool during organogenesis is rapidly compensated by increased expansion of the remaining progenitors so that the organ still achieves normal size, consistent with its known regenerative capacity. In contrast, reducing the functional progenitor pool allocated to the pancreas-forming domain of embryonic endoderm is not rescued by compensatory proliferation, resulting in a smaller organ, suggesting that the proliferative capacity of pancreatic epithelial progenitors is restricted and finite.
In this study we have investigated whether the thymus, where the epithelium is also of an endodermal origin, conforms to either of these models of organ size determination. To do this, we have utilized embryo fusion chimeras formed between wild-type (WT)3 mice expressing enhanced yellow fluorescent protein (eYFP) in all tissues under the control of the Rosa26 promoter and FoxN1-deficient nude mice (non-eYFP), allowing the contribution of each partner to various tissues in resultant individuals to be traced at defined stages of development. In FoxN1-deficient nude mice, normal development of thymic epithelial progenitors is blocked at an early stage of thymus organogenesis, although initial specification of epithelial progenitors and capacity to form a thymic rudiment occur via a FoxN1-independent mechanism (14). This system provides a model in which “sterile” developmentally incompetent progenitors can compete with WT progenitors for allocation into the thymic-forming domain of endoderm in chimeric individuals. Due to the random variation in chimerism occurring between individual mice, this provides a mechanism for determining the size of the functional progenitor pool against its impact on the eventual size of the thymic epithelial compartment.
Using this approach, we show that restricting the availability of a functionally competent FoxN1-expressing population within the thymic epithelial progenitor pool results in a decrease in the size of the thymic epithelial cellular compartment during embryonic stages of organogenesis and that this reduction persists into the adult period. Our data support the notion that the size of the endodermal progenitor population allocated to the initial formation of the thymus has a finite capacity for expansion and can influence events in both embryonic and adult thymus, a finding that has important implications for the development of strategies aimed at regenerating the thymus following the onset of involution.
Materials and Methods
C57BL/6 nude, eYFP, CD1, and BALB/c mice were bred at the University of Birmingham, and all experiments were performed in accordance with the U.K. Home Office regulations. Adult mice were used at 4 wk of age.
Generation of embryo fusion chimeras
Generation of embryo fusion chimeras was performed in line with previous reports (20). Briefly, 8-cell-stage embryos were stripped of zona pellucida using pronase. One denuded 8-cell embryo from each partner was added to a microwell and cultured in vitro at 37°C overnight in M16 media until a fused blastocyst was observed. Fused blastocysts were subsequently transferred into 2.5-day-postcoitus pseudopregnant CD1 females. Day of blastocyst transfer was taken as day 3 of gestation. Generation of adult chimeric mice was confirmed by coat color phenotype and analysis of chimerism in thymus and/or pancreatic and submandibular salivary gland epithelium.
The following Abs were used for flow cytometry: anti-CD45-PE (clone 30-F11), anti-CD4-PE (clone GK1.5), anti-CD4-allophycocyanin (clone: L3T4), anti-CD44-PE-Cy7 (clone IM7), anti-CD25-biotin (clone PC61), anti-CD3e-allophycocyanin (clone 1 45-2C11), anti-CD19-PE (clone 6D5), anti-CD8a-allophycocyanin (clone 53-6.7) (all eBioscience), streptavadin-PE (BD Pharmingen), and anti-EpCAM-1-Alexa Fluor 647 (epithelial cell adhesion molecule, clone G8.8).
Embryonic thymi were digested using trypsin (0.25%) and EDTA (Sigma-Aldrich) to give single-cell suspensions (21). Adult thymi were cut into small pieces and incubated with collagenase dispase (2.5 mg/ml) and DNase I (1.5 μg/ml) (Sigma-Aldrich) at 37°C, pipetting regularly to ensure disaggregation. A subsequent brief incubation of adult thymic cells with trypsin (0.25%) at 37°C was performed to ensure full dissociation of tissue (22). Digestion of pancreas and submandibular salivary gland was performed using collagenase dispase (2.5 mg/ml) and DNase I (1.5 μg/ml).
Flow cytometry was performed using a dual laser LSR I machine (BD Biosciences), with forward/side scatter gates set to exclude nonviable cells. FACS data were analyzed using FlowJo software (Tree Star). Quantitation of total adult organ cell numbers was performed using AccuCount blank particles (Spherotech).
Confocal microscopy was performed as described (23). Frozen tissue sections were stained with the following: anti-pan cytokeratin-FITC (clone C-11, Sigma-Aldrich), rabbit anti-keratin 5 (polyclonal MK5, Covance Research Products), anti-rabbit Ig-biotin (DakoCytomation), anti-CD4-Alexa Fluor 647 (clone L3T4), anti-CD8-biotin (clone CT-CD8b, both eBioscience), and streptavadin-Alexa Fluor 555 (Invitrogen). Confocal images were obtained using an LSM 510 Meta microscope (Zeiss) using Zeiss LSM software.
Data were evaluated using Pearson product-moment correlation coefficient to determine correlation between datasets. r2 values represent coefficients of determination. A p-value of <0.05 was considered significant in all analyses of the significance of correlation.
Nude-derived tissues in fusion chimeras contribute normally to most tissues but not the thymus
The establishment of cortical and medullary thymic epithelial cells from bipotent progenitors represents an essential step in the generation of thymic microenvironments that support the generation of a self-tolerant T cell repertoire. While it is clear that thymic epithelial cell (TEC) progenitor development involves both differentiation and phases of expansion, the processes acting on the progenitor pool to ensure normal thymus growth are not clear. To study the influence of the size of the initial thymic epithelial progenitor pool on thymus development, we adopted a strategy based on FoxN1-deficient embryo fusion/blastocyst complementation (24, 25). Importantly, FoxN1 regulates thymic epithelial differentiation in a cell-autonomous fashion (20). Moreover, while FoxN1-deficient TEC progenitors are still able to undergo initial stages of thymus organogenesis (26), in the postnatal thymus it is proposed that they may represent blocked progenitors unable to continue their differentiation program (13, 20). Thus, by generating chimeric mice from a mixture of eYFP-marked WT (FoxN1-sufficient) and nude (FoxN1-deficient) embryos (hereafter termed WT/FoxN1−/− mice), we assessed the impact of the presence of sterile (FoxN1-deficient) progenitors within the TEC progenitor pool on thymus development.
Embryo fusion chimeras were generated by fusing FoxN1−/− or FoxN1+/+ WT 8-cell embryos 1:1 with eYFP-expressing 8-cell embryos to form individual fused blastocysts (Fig. 1,A), which were transferred to pseudopregnant foster mothers and allowed to develop to the required stage for analysis. Overt chimerism was evident at a gross level in both embryonic stages and in adults (Fig. 1,B) and could be quantified in individual organs following disaggregation and flow cytometry for eYFP expression in conjunction with markers of epithelial (EpCAM-1) or hematopoietic (CD45) cell populations. As described previously (25), variation in the level of chimerism was observed between individual mice in FoxN1+/+ WT/eYFP chimeras (data not shown). However, within any one individual FoxN1+/+ WT/eYFP mouse, chimerism was readily apparent in all tissues analyzed, including the thymus (Fig. 1,C). Importantly, consistent with the inability of FoxN1−/− thymic epithelial cells to develop beyond the early progenitor stage, the proportion of cells of this origin in the adult thymic epithelial compartment of WT/FoxN1−/− mice was minimal even when the proportion of FoxN1−/− contribution to other tissues was high, as illustrated by comparison between chimerism seen in submandibular salivary gland epithelium, an organ being independent of FoxN1 function (Fig. 1,D) and thymic epithelium (Fig. 1,E) within a single WT/FoxN1−/− mouse. Importantly, these findings confirm previous observations that while FoxN1−/− cells are unable to contribute normally to differentiated thymic tissue, they are able to contribute normally to other FoxN1-independent organs such as the submandibular salivary gland (25), which can therefore be used as an indicator of the level of chimerism in individual mice. Additionally, analysis of chimerism within CD45+ thymocytes and CD45−EpCAM− thymic mesenchyme within the same WT/FoxN1−/− chimera (Fig. 1,E) revealed no effect of loss of FoxN1 function within these cellular compartments and also demonstrated a similar trend of chimerism as observed within submandibular salivary gland epithelium (Fig. 1 D). In support of this, comparison of chimerism between thymic epithelium and pancreatic epithelium in individual FoxN1+/+ WT/eYFP mice revealed a similar trend in the degree of chimerism between different organs of the same mouse (data not shown). These findings are in agreement with previous studies indicating that individual embryo fusion chimeras display chimerism within organs at a level comparable with the relative contributions of individual embryo partners to skin and pigment formation (25).
Nude epithelial progenitors effectively compete for space in the allocation of epithelial progenitors to the thymic anlage
To confirm that FoxN1−/− progenitors do initially contribute to the developing thymus in fusion chimeras and compete for “space” in the progenitor pool, we investigated the level of chimerism seen in the thymus during the early stages of thymus development. In nonchimeric FoxN1−/− mice, a thymic anlage is established from the third pharyngeal pouch endoderm. This consists of epithelial cells displaying blocked or abnormal development and persists as a discrete alymphoid structure until at least embryonic day (E)16 with no increase in size (27, 28), indicating that FoxN1−/− cells should be readily detectable in chimeric rudiments at these early stages. We therefore disaggregated WT/FoxN1−/− chimeric thymic lobes at E15 and looked for the presence of both FoxN1−/− (eYFP−) and FoxN1+/+ WT (eYFP+) contributors within the epithelial compartment. FoxN1−/− cells were readily detectable in chimeric lobes at both E12 and E15, with the range of chimerism varying between individual embryos from 10 to 50% (Fig. 2 A). These findings provide direct evidence that developmentally sterile FoxN1−/− epithelial progenitors can be effectively incorporated into the developing thymic rudiment in chimeric animals.
To determine whether incorporation of developmentally blocked FoxN1−/− cells into the thymic anlage restricts the space available for WT epithelial progenitors capable of normal proliferative expansion, we next assessed the relationship between the total number of embryonic CD45−EpCAM+ epithelial cells and the degree of FoxN1−/− contribution in individual embryonic thymi. As shown in Fig. 2,B, there was a strong negative correlation between the percentage contribution of FoxN1−/− cells to the thymic epithelial pool and the overall number of thymic epithelial cells detected in E15 thymic lobes. A similar correlation between the total number of epithelial cells and the extent of FoxN1−/− contribution was also seen in E12 lobes (Fig. 2 C), indicating that increased proliferation by WT cells cannot fully compensate for the reduction in competent progenitor number between E12 and E15. These observations confirm that the presence of FoxN1−/− cells does reduce the number of WT progenitors in the early rudiment and provide a basis to determine whether initial reductions in the number of developmentally competent epithelial progenitors in the early thymus are compensated by the time the adult stages are reached.
Functional thymic epithelial progenitor pool size determines the size of the thymic epithelial compartment
Recent studies have shown that organs displaying a high level of regenerative capacity, such as liver, undergo compensatory proliferation in response to a reduction in the size of the initial progenitor pool (19). To investigate whether a similar compensatory mechanism might operate within the thymus, to restore the reduction in the size of the thymic epithelial compartment observed in fetal WT/FoxN1−/− fusion chimeras, we analyzed TEC numbers in a range of adult chimeras.
As FoxN1-deficient TECs are able contribute to the initial formation of the embryonic thymic rudiment, but do not undergo normal programs of growth and differentiation as shown above (also see Ref. 14), chimerism in the adult thymus does not reflect chimerism levels when the thymus is first established. To overcome this problem, we compared the number of TEC in chimeric adults to the degree of chimerism in a range of other tissues. As demonstrated above, this provides a reflection of the degree of chimerism expected in the thymus when both partners can contribute normally, reflecting the initial degree of chimerism in the thymus anlage. Since thymic epithelium is derived from endoderm (26), we first compared TEC numbers with the level of chimerism seen in the epithelial compartment of other endoderm-derived organs (Fig. 3, A and B). A significant correlation was observed between the total number of TECs and the contribution of FoxN1-deficient cells to both submandibular salivary gland epithelium (r2 = −0.98, p = <0.00001) and pancreatic epithelium (r2 = −0.67, p = 0.01) (Fig. 3, A and B, respectively), with the number of TECs decreasing as the degree of FoxN1-deficient nude chimerism increased. In support of this, a similar correlation was also observed when TEC numbers were analyzed in relation to chimerism within nonendodermal tissues including both FoxN1-deficient nude-derived thymocytes (Fig. 3,C, r2 = −0.88, p = 0.0005) and nonepithelial thymic stromal cells (Fig. 3 D, r2 = −0.92, p = 0.0001).
Collectively, these data demonstrate that, in the adult thymus, there is a direct correlation between the degree of FoxN1-deficient chimerism seen across a range of tissues and the total number of TECs. This argues that the reduction of TEC numbers observed in the embryonic period as a result of reducing the number of developmentally functional TEC progenitors allocated to the thymic anlage is not overcome by compensatory proliferation by the time the adult stage is reached.
Impact of thymic epithelial compartment size on thymocyte production
To investigate whether the reduced TEC frequency observed in WT/FoxN1−/− chimeras affects thymocyte development, we analyzed total thymocyte numbers and compared them to total epithelial cell numbers in individual thymi during embryonic development and in the adult. At E15 (Fig. 4 A) a strong correlation (r2 = 0.64, p = <0.0001) was apparent between the total number of TECs and total thymocyte numbers, with the latter decreasing in line with reductions in the number of developmentally competent epithelial cells, a finding in agreement with previous observations suggesting that thymic epithelial cell numbers limit the availability niches for thymocyte development (8, 29).
The relationship between reduced thymocyte numbers and reduced epithelial cell numbers demonstrated a much less dramatic correlation in the adult (Fig. 4,B) (r2 = 0.15). Although the correlation followed the same general trend as that observed in the embryo, it was not found to be significant (p = 0.3). Additionally, assessment of thymus mass and total thymocyte cellularity in adult WT/FoxN1−/− mice compared with the degree of FoxN1−/− chimerism, as determined by analysis of pancreatic epithelium, did not demonstrate a significant relationship at this stage (Fig. 4, C and D, respectively).
Analysis of the developmental distribution of thymocyte subsets in the adult thymus of WT/FoxN1−/− chimeras demonstrated normal patterns of both CD4−CD8− double-negative I–IV subsets compared with WT (Fig. 5, A and B) and also progression through CD4+CD8+ double-positive to CD4+CD8− and CD4−CD8+ single-positive stages (Table I). Additionally, confocal analysis of WT/FoxN1−/− chimeric thymi revealed normal organization and differentiation of cortical and medullary thymic epithelium as assessed by cytokeratin staining (Fig. 5, C and D), as well as normal distribution patterns of developing thymocytes as defined by CD4 and CD8 (Fig. 5, E and F). Taken together, these data suggest that despite a reduced initial thymic epithelial progenitor pool, WT/FoxN1−/− chimeric thymi demonstrate normal differentiation and organization of thymic microenvironments and support a normal program of T cell development.
|Thymocyte Subset .||WT .||WT/FoxN1−/− .|
|% CD4−CD8−||3.27 ± 0.4||3.68 ± 0.2|
|% CD4+CD8+||81.2 ± 0.5||81.37 ± 5.4|
|% CD4+CD8−||13.11 ± 0.8||12.54 ± 0.7|
|% CD4−CD8+||2.42 ± 0.2||2.41 ± 0.1|
|Thymocyte Subset .||WT .||WT/FoxN1−/− .|
|% CD4−CD8−||3.27 ± 0.4||3.68 ± 0.2|
|% CD4+CD8+||81.2 ± 0.5||81.37 ± 5.4|
|% CD4+CD8−||13.11 ± 0.8||12.54 ± 0.7|
|% CD4−CD8+||2.42 ± 0.2||2.41 ± 0.1|
Thymocyte developmental subset distribution as determined by CD4 and CD8 expression in adult WT and WT/FoxN1−/− thymi (mean ± SE: WT, n = 5; WT/FoxN1−/−, n = 6).
Some endodermal organs such as the liver can rapidly recover from reductions in the size of the progenitor pool from which they arise to produce a normal-sized organ by the time of birth. This ability is associated with the known regenerative capacity of the adult liver and suggests that progenitor populations with a relatively unrestricted capacity for proliferation persist into the adult stages. Conversely, the pancreas, another endoderm-derived organ, is not able to compensate for reductions in its progenitor pool and lacks noticeable regenerative ability in the adult, suggesting that it arises from a progenitor pool with restricted capacity for proliferation that is mostly exhausted during organ formation (19). Our present findings demonstrate that the endoderm-derived compartment of the thymus more closely resembles the pancreas than the liver in its ability to compensate for reductions in the size of its epithelial progenitor pool. These findings demonstrate that the epithelial progenitor pool of the thymus has a limited capacity for proliferation within our model. Whether this limited capacity is entirely intrinsic to the progenitor population or is influenced by the availability of stem cell niches, some of which may remain occupied by developmentally blocked FoxN1-deficient cells, remains to be determined.
Our findings also demonstrate that a reduction in the number of functional epithelial cells in intact animals has consequences for the number of thymocytes produced without affecting the normal pattern of thymocyte development. This is in agreement with previous observations where restricting the extent of epithelial progenitor cell proliferation by removing mesenchymal support (8) or by knockout of the fibroblast growth factor receptor (FGFR)2iiib on these cells (30) resulted in an overall smaller thymus with fewer thymocytes. This correlation between reduced functional epithelial cell numbers and reduced thymocyte numbers in WT/FoxN1−/− chimeric thymi was most obvious in embryonic as compared with adult thymus where the reduction in thymocyte numbers relative to epithelial cells was less marked. This difference may reflect the fact that at E15, the thymocyte population is almost exclusively comprised of cells at the double-negative 1–3 stages, which are thought to be particularly dependent on the availability of epithelial niches or products for their development (29). Additionally, it is possible that the requirement for thymic epithelium-derived chemokine signals required for recruitment of hematopoietic precursors to the thymus within the embryo (31) may be more pronounced in embryonic vs adult settings, leading to a more pronounced defect in thymocyte numbers in embryonic WT/FoxN1−/− chimeras. In contrast to the reliance of DN thymocyte subsets on niche availability, there is evidence that thymocyte proliferation and differentiation post β-selection is less dependent on epithelial availability (29) and this may lead to a less obvious relationship between epithelial and thymocyte numbers as later stages of thymocyte maturation accumulate to give the dominant cortical CD4+CD8+ population seen in the adult. It remains to be determined whether the restriction of numbers before β-selection has consequences for the extent of the αβ TCR repertoire even though numbers may be subsequently amplified.
Our findings have important implications for strategies for thymic regeneration following age-related involution or for thymic recovery after ablative therapy since they suggest that the potential for expansion of thymic epithelial populations in the adult may be limited. However, our findings do not exclude the possibility that cells with some proliferative capacity survive into the adult stages, and some studies have indicated the ability of thymic epithelial cells to respond to external stimuli and undergo expansion with apparent reversal of thymic atrophy in adult mice (7, 32). In this context, although bipotent progenitors for cortical and medullary thymic epithelium have been demonstrated within embryonic and postnatal thymus (11, 13), it is as yet unclear whether such progenitors persist within the aged thymus (33). Similarly, the existence and possible persistence of lineage-restricted progenitors (cortex or medulla) are still to be defined. Thus, it will be important to consider the development of alternative strategies for the restoration of thymic function, including grafting of donor-derived thymus tissue (34, 35) or of thymic epithelial progenitors generated de novo from embryonic stem cells.
The authors have no financial conflicts of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by a Leverhulme Early Career Fellowship (to W.E.J.) and an Medical Research Council Programme Grant (to E.J.J. and G.A.).
Abbreviations used in this paper: WT, wild type; E, embryonic day; EpCAM, epithelial cell adhesion molecule; eYFP, enhanced yellow fluorescent protein, TEC, thymic epithelial cell.