The thymus is one of the cornerstones of an effective immune system. It produces new T-cells for the naïve T-cell pool, thus refreshing the peripheral repertoire. As we age, the thymus atrophies and there is a decrease in the area of active T-cell production. A decline in the output of the thymus eventually leads to changes in the peripheral T-cell pool which includes increases in the number of cells at or near their replicative limit and contraction of the repertoire. Debate about the age-associated changes in the thymus leading to functional decline centres on whether this is due to problems with the environment provided by the thymus or with defects in the progenitor cell compartment. In mice, the evidence points towards problems in the epithelial component of the thymus and the production of IL-7 (interleukin 7). But there are discussions about how appropriate mouse models are for human aging. We have developed a simple system that utilizes both human keratinocyte and fibroblast cell lines arrayed on a synthetic tantalum-coated matrix to provide a permissive environment for the maturation of human CD34+ haemopoietic progenitor cells into mature CD4+ or CD8+ T-lymphocytes. We have characterized the requirements for differentiation within these cultures and used this system to compare the ability of CD34+ cells derived from different sources to produce mature thymocytes. The TREC (T-cell receptor excision circle) assay was used as a means of identifying newly produced thymocytes.
The thymus and thymic atrophy
In humans, the thymus is located in the anterior mediastinum at the base of the great vessels overlaying the heart. During the lifespan, it is seeded by blood-borne progenitors derived from the bone marrow. The amount of active bone-marrow-producing components of the haemopoietic system declines with age. At birth, active bone marrow is found in most bones, but, in later life, active sites may include spongy regions at the end of long bones, and also vertebral bones, sternum, ribs, flat bones of the cranium and pelvis. The progenitors which seed the marrow are one of the essential requirements for T-cell production in the thymus and the other is the stromal cell elements which, as a minimum, must include epithelial cells and fibroblasts (Figure 1).
Age-associated macroscopic changes in the thymus, noted in humans and many other mammals, include a reduction in the total volume occupied by the organ and some change in its colour which is often thought to be due to an increase in the fat content. At the microscopic level, there is a noted and significant reduction in cellularity with an accompanying loss of the distinct histological regions, the cortex and medulla. Accumulation of adipocytes is also a feature of this process. Comparative analysis from young and old animals has been at the centre of efforts to determine the cause of these age-related changes. Broadly, these causes may be placed into two camps, those who propose that changes in the stromal elements are the leading cause of age-associated atrophy and those who contend that changes in the number or function of the thymic seeding population are central to the decline seen.
The function of the thymus is to produce new T-cells each with a different receptor specificity such that the constant outflow of T-cells from the thymus contributes to the pool of receptor specificities of T-cells in the peripheral T-cell pool (the repertoire). New T-cells from the thymus enter the naïve T-cell pool. Exit from this pool is either though entering the memory T-cell pool following successful antigen-induced activation or through cell death. Atrophy of the thymus is coupled to a decline in the production of new T-cells and an age-associated reduction in the number of cells in the naïve T-cell pool. Consequently, as we age, our memory T-cell pool grows, our naïve T-cell pool shrinks and the overall T-cell repertoire becomes distorted.
With reduced output from the thymus, our peripheral T-cell pool may not be able to recognize new antigens. Additionally, the number of T-cells in the peripheral T-cell pool is closely regulated, such that a decline in the contribution from the thymus may be matched by the induced proliferation of peripheral T-cells. T-cells have limited replicative capacity and so every round of division brings them closer to their limit of replication. Both the reduced naïve T-cell pool and the subsequent homoeostatic proliferative response may have clinical consequences. One of these consequences may be an inability to cope with new vaccines or new infections. An example of the latter was the outbreak of West Nile virus in the New York metropolitan area in 1999 which was the first documented occurrence of this virus in the Western Hemisphere and where most (88%) of the hospitalized patients were at least 50 years of age . In addition to this susceptibility to new pathogens, clinicians recognize that older individuals often have difficulties in dealing with pathogens which they have previously overcome. Common problems include reactivation of herpes zoster virus  or the increased immune response to cytomegalovirus . This may be the result of a lifetime's response to these viruses leading to T-cell exhaustion or the limit of proliferation reached for the T-cells specific for those agents . In addition, this reduced immune function may lead to failure of an individual to make an adequate response to a vaccine such as influenza , leading to them being susceptible to infection with this organism.
Because the thymus does not contain an endogenous source of cell progenitors, the T-lymphocyte generation cannot proceed without seeding of the thymus with progenitor cells from external sources. In view of this, atrophy linked to a decline in the number of seeding cells was suggested more than 30 years ago  and, more recently, work has been carried out to phenotype and enumerate the number of early T-cell progenitors and showed that these decline both in number and in function with age . Although this decline may contribute to atrophy in mice, the extent of its contribution is difficult to assess since the number of progenitor cells needed to seed the thymus is extremely small and there is no evidence to suggest that the number of progenitor cells falls below the necessary levels. Evidence to suggest competency in the progenitor compartment while indicating defects in the thymic stromal milieu come from experiments using transplantation of fetal thymus under the kidney capsule of old mice. These showed that normal thymopoeisis could be generated in the transplanted lobes by progenitors from the old host. When early thymic progenitors from young animals were injected into old thymuses, they were unable to restore thymopoeisis . This was also complemented by a study showing that progenitor cells were recruited equally well by the thymus of young and aged mice . Both studies suggest that the age-related defect is located within the stromal element of the thymus, and several studies have indicated that aging is accompanied by a phenotypic change in the stromal elements of the thymus including loss of expression of MHC II , changes in the expression of Notch1 and Delta , and alterations in the expression of some cytokines produced by the stromal elements [12,13].
Aging is associated with a reduction in the amount of IL (interleukin)-7 produced within the thymus of mice [12,13]. This cytokine has a central role in the development and survival of T-lymphocytes at several stages of development, and treatment of old animals with IL-7 has been reported to reverse thymic atrophy, enhance thymopoiesis and consequently improve immune responses compared with age- and sex-matched control animals [14–16]. IL-7 is now in clinical trials, and the therapeutic approach used is repeated injection to ensure a steady maintenance of the drug within the tissues in order to be effective [17–19]. Compliance on drug use is often linked to issues such as side effects and delivery systems, and, with IL-7, which has to be delivered by injection over a period of weeks, it often means either that the drug has to be delivered in a clinical setting or that the recipient has to be trained for self-injections. With these modes of delivery, there is often low compliance and a limitation in its progression to being widely used. Although IL-7 could be beneficial if it could be delivered to individuals within the community to improve their immune systems, this would require a more acceptable route of delivery and the route chosen should have few, if any, side effects.
Our recent study using animal models showed that IL-7 could be inhaled in the form of an aerosol, entering the body through the lungs and going directly into the bloodstream and from there being transported around the organs and tissues. Comparison between aerosol-delivered and injected IL-7 showed that within the first 48 h of delivery, the amount in the blood was indistinguishable between the two routes of delivery, but that with delivery using the inhalation route, there was better distribution around the body specifically to the lymphoid organs . In terms of the effectiveness of the drug, our results showed that aerosol delivery was better than delivery by injection, so animals receiving IL-7 by aerosol showed a greater effect on their immune systems particularly when they were challenged with influenza infection (R. Aspinall, S. Govind, A. Kok and P.O. Lang, unpublished work). Furthermore, and of considerable concern to us was whether there would be any adverse effects associated with delivery via the lungs. Our results would suggest that this is not the case .
Translating this mode of therapy from animals to humans still requires several steps, and among the most important is the determination of whether all of the elements which contribute to thymic decline in animal models are mirrored in humans.
Mice versus humans
Studies in humans have revealed a pattern of timing of thymic atrophy, which parallels that seen in rodent models. Although the exact time point at which atrophy is triggered is variable and dependent on several factors, many would agree that after birth, the thymus increases in weight and cellularity, and at a time previously thought to be associated with puberty, the thymus begins to decline in size. The rate of decline is biphasic, with a rapid decline seen in the period up to middle age followed by a slower phase thereafter.
Measurement of thymic output has been achieved by following the change in the number of TRECs (T-cell receptor excision circles) in T-cells in the blood as we age . TRECs are formed during T-cell development when rearrangement of the DNA to produce a functional αβ T-cell receptor leads to excision of a stretch of DNA within the genes coding for the α chain and the formation of a closed circle of DNA within the nucleus . Because of the mechanism associated with selection of the receptor, there may be up to two of these signal joint closed circles within each thymocyte. These circles are thought to be relatively stable and maintained within the cells and so are present in higher amounts in populations of recent thymic emigrants . However, since these closed circles do not contain an origin or replication when the cell divides, the circles are passed to only one of the daughter cells, effectively diluting their number within the overall population. Analysis of the blood of human volunteers reveals that the percentage of individuals with detectable TRECs declines with age and are still detectable in many individuals over 90 years of age, but are not detectable in many centenarians [21,23].
The similarity in the tempo of atrophy across several mammalian species led many authors to suggest that there would be similar causes, with the underlying expectation that therapies developed in rodent models would translate to humans. The identification of defects in the thymic stromal elements in animal models led to the examination of the stromal component of the human thymus from normal human subjects up to 78 years of age. The results from these studies showed an age-related increase in expression of LIF (leukaemia-inhibitory factor), oncostatin M, SCF (stem cell factor), IL-6 and MCSF (macrophage colony-stimulating factor) . Although these results differ from those found in rodent experimental systems, they still show age-associated changes in the stromal elements in line with expectations from the murine models.
The principal question that remains concerns the T-cell progenitors seeding the thymus. In old mice, as we have seen, the question has been resolved and there would appear to be sufficient T-cell progenitors to seed a thymus when the defects in the stroma have been reversed. But the difference between the murine and human situation may be up to or above 80 years, a time period that could have a considerable impact on the functional capacity of the human T-cell progenitor populations.
There are two areas of concern: the first is whether there are sufficient numbers of T-cell progenitors circulating in the blood of older individuals, and the second is the functional capacity of any of these progenitors. Since the preservation of numbers of progenitors must be dependent on cell division, foremost among the concerns about maintenance of progenitor function must include the possibility of accumulation of DNA damage in stem cells during the aging process . Multiple hypotheses have been proposed to address these concerns, and recent studies have suggested that the regenerative capacity of stem cells decline with age. Specifically, presumptive progenitors appear to be myeloid-biased and exhibit less inclination to differentiate into lymphoid lineages [26,27] or alternatively there may be a loss of a specific subsets with the capacity to form lymphoid progeny [28,29].
Building a human thymus
If a therapy based around supplementing the age-related thymic stromal cell deficiencies is to translate to human trials, then we need to be certain that there would be sufficient functional progenitor cells to repopulate an aging thymus. We have approached this question in a novel manner taking a lead from synthetic biology by trying to construct a human thymus in the laboratory using commonly available cell lines. Experiments in mice revealed that epithelial cells and fibroblasts were critical in the construction of the thymic environment  and both were needed to engineer the development of seeded precursors towards the T-cell lineage. Work in humans showed that skin-derived cells were able to support T-lymphocyte development from bone-marrow-derived CD133+ cells . Building on these studies, we chose to construct a human thymic stroma using readily available cell lines, including a human keratinocyte cell line and human fibroblasts from a commercial source.
Monolayer versus three-dimensional cultures
Although monolayer cultures of thymic stromal cells will not support T-cell development , monolayer cultures have been described which permit the differentiation of haemopoietic precursor cells along the T-cell developmental pathway and notable among these are the OP9-Dl1 cells [33,34]. These cells express Notch Dll-1 (Delta-like ligand 1), one of the five Notch ligands, the others of which are Dll-3, Dll-4, Jag (Jagged)-1 and Jag-2. The Dll-4 ligand is normally highly expressed on thymic epithelial cells and it is this ligand which provides the signal for progenitors to adopt the fate of differentiation along the T-cell lineage . Experiments with our human cell lines when grown in monolayer cultures showed that when these monolayers were seeded with cord-blood-derived CD34+ cells in medium supplemented with IL-7, IL-15 and Flt-3L (Fms-like tyrosine kinase 3 ligand) , there was no differentiation of these CD34+ cells and they failed to survive.
The OP9-Dl1 system is not as efficient in producing mature thymocytes and naïve T-cells as the fetal thymic organ culture system, and, with the thought that three-dimensional cultures may be better than two-dimensional ones, we attached our keratinocytes and fibroblasts to a three-dimensional tantalum-coated matrix. When these cultures were seeded with CD34+ cord-blood-derived cells, they provided the environment for the production first of immature thymocytes expressing both CD4 and CD8 and then of mature thymocytes expressing either CD4 or CD8 along with the CD3 molecule and clear expression of the αβ T-cell receptor. Differentiation to this level was often noted by 14 days after seeding .
Characteristics of the culture which provide the correct signal
The need for a complex multidimensional structure for the efficient production of T-cells from their progenitors was shown very early in the murine system with the development of the fetal thymic organ culture system which, as mentioned above, was more supportive of T-cell development than monolayer sheets of thymic stromal cells. Later work revealed that sheets of thymic stromal cells in monolayer cultures fail to maintain expression of the Notch ligands Dll-1 and Dll-4 and so cannot support T-lymphopoiesis. In comparison, fetal thymic organ cultures showed up-regulation of these Notch ligands, suggesting that levels of Dll-1 and Dll-4 were maintained through the close interactions between cells in a three-dimensional arrangement . In our human model system, we see the up-regulation of the Notch ligand Dll-4 within the epithelial cell line when cultured in the matrix. For the progenitor cell, this may mean that signalling from multiple directions may be a more efficient means of ensuring the translation of the signal into differentiation.
The induction of changes in expression of Dll-4 as well as other genes within the epithelial cells was of considerable interest. Changes in gene expression are often considered to be associated with external stimuli acting through ligand–receptor interactions. For example up-regulation of the Notch ligand Dll-4 in response to VEGF (vascular endothelial growth factor) has been reported previously . The idea that a change in cellular configuration from attachment to a planar surface, where large areas of the cell are in contact with the substratum, to attachment to a three-dimensional matrix, where there may be a much smaller area of the cell in contact with the substratum, could lead to such considerable change in expression and would suggest several pos-sible stimuli. Although it remains plausible that growth of the epithelial cells on the tantalum matrix could lead to the change in production of a factor which acts through a cell-surface receptor and changes the expression of Dll-4, we must also consider the possibility that it could be an internally generated signal leading to a change in gene expression. Altered gene expression in cells cultured in monolayers compared with their culture in three dimensions has been noted previously for the human liver hepatocellular carcinoma cell line HepG2 with changes associated not only with genes concerned with the cytoskeleton and the cell structure, but also with genes concerned with metabolism [38,39].
The production of a simple model system containing human cells which permits the efficient differentiation of progenitor cells along the T-cell lineage provides an opportunity to answer a number of critical questions. Foremost among these is the frequency of thymic seeding cells in the blood of older adults and how this frequency is affected by age. Strategies directed at reversing thymic atrophy and inducing new T-cell development in older adults will be dependent on progenitor cells transiting from the bone marrow to the thymus via the bloodstream and if these are not present at high frequency, then, in addition to rejuvenating or supplementing missing components of the thymic stromal element, there may be an additional requirement to increase the abundance of thymic progenitors in the bloodstream.
Biochemical Determinants of Tissue Regeneration: Biochemical Society Annual Symposium No. 81 held at Shrigley Hall Hotel, Macclesfield, U.K., 11–13 December 2013. Organized and Edited by Adam Giangreco (University College London, U.K.) and Catherine Merry (University of Manchester, U.K.).
Our work is supported by Research into Ageing [grant number 234] and also by the Myrtle Peach Trust.