Author contributions: S.C.dB.: conception and design, collection, interpretation and analysis of data, editing and final approval of the manuscript; V.S.Z.; conception and design, financial support, collection, interpretation and analysis of data, editing and final approval of the manuscript; N.T.: conception and design, financial support, collection, interpretation and analysis of data, writing and final approval of the manuscript.
Disclosure of potential conflicts of interest is found at the end of this article.
first published online in STEM CELLSEXPRESS April 4, 2013.
Allogeneic hematopoietic stem cell (HSC) transplantation can cure patients suffering from diverse genetic and acquired diseases as well as cancers. Nevertheless, under conditions where T-cell reconstitution is critical, the entry of donor progenitors into the thymus remains a major bottleneck. It is assumed that following the intravenous injection of HSC, they first home to the BM. More committed progenitors can then be exported to the thymus in response to a myriad of signals regulating thymus seeding. Notably although, the thymus is not continually receptive to the import of hematopoietic progenitors. Furthermore, as stem cells with self-renewing capacity do not take up residence in the thymus under physiological conditions, the periodic colonization of the thymus is essential for the sustained differentiation of T lymphocytes. As such, we and others have invested significant efforts into exploring avenues that might foster a long-term thymus-autonomous differentiation. Here, we review strategic approaches that have resulted in long-term T-cell differentiation in immunodeficient (SCID) mice, even across histocompatibility barriers. These include the forced thymic entry of BM precursors by their direct intrathymic injection as well as the transplantation of neonatal thymi. The capacity of the thymus to support hematopoietic progenitors with renewal potential will hopefully promote the development of new therapeutic strategies aimed at enhancing T-cell differentiation in patients undergoing HSC transplantation. STEM Cells2013;31:1245–1251
Hematopoietic stem cells (HSCs) allow the blood system to be replenished and their ongoing differentiation is critical for maintaining long-term homeostasis in response to physiological and external demands. Transplantation of HSCs can be used to treat patients suffering from diverse genetic and acquired diseases, including cancers. This treatment requires the successful engraftment of HSCs/progenitors as well as a rapid and ongoing differentiation of all blood lineage cells, all in the absence of graft-versus-host disease. For patients where T-cell differentiation and/or function is affected, successful transplantation also necessitates that HSC or their progeny home to the thymus, the major site of T-cell differentiation. While the thymus was previously thought to be active only during infancy and childhood, we now know that the thymus continues to generate new T cells during adulthood; patients undergoing stem cell transplantation between 30 and 40 years of age have a high probability of thymic rebound while late thymic recovery has been detected in patients up to 50 years of age [1, 2].
ROLE OF THE THYMUS IN T-CELL DIFFERENTIATION
The thymus, derived from the third pharyngeal pouch endoderm during embryogenesis, is an organ that was already acknowledged by the Greeks. Its name may stem from the Greek word “thymos” which denotes life force or soul. Importantly although, the thymus was long considered to be obsolete relic of evolution. As recently as 50 years ago, Sir Peter Medawar, recipient of the Nobel Prize for his ground-breaking work on graft rejection and the acquisition of immune tolerance, stated “We shall come to regard the presence of lymphocytes in the thymus as an evolutionary accident of no very great significance” (1963) . Notably although, the pioneering studies performed by Francis Albert Pierre Miller in the early 1960s had already pointed to the critical role of the thymus in T-cell generation .
More recently, the treatment of patients with thymic dysgenesis by allogeneic thymus tissue transplant has resulted in a rescue of their T-cell deficiency, highlighting the importance of the thymus in human T-cell differentiation [5, 6]. This has also been extensively found to be the case in HIV-1 infection where immune reconstitution following highly active antiretroviral therapy has been shown to require an efficient thymopoiesis (reviewed in ). Furthermore, efficient T-cell regeneration is a critical parameter in the outcome of all patients undergoing allogeneic HSC transplantation, especially in patients outside the pediatric age group. Thus, today, in the beginning of the 21st century, the importance of a thymus, or more specifically, a thymus with an appropriate structure and architecture, is widely accepted to be a prerequisite for the optimal development of T precursors to mature functional T lymphocytes.
Many important studies have therefore focused on reversing the normal process of thymic atrophy during aging as well as in the context of cytoreductive treatments such as chemotherapy and radiation. These latter treatments exacerbate thymic atrophy by injuring thymic epithelial cells (TEC), thereby slowing T-cell reconstitution (reviewed in [8, 9]). Notably although, several strategies can improve the situation. Sex steroid ablation, used for the treatment of prostate cancer, improves thymic output  as does a hypogonadal state . Other agents such as keratinocyte growth factor (KGF) stimulate TEC proliferation and can reduce TEC injury when given prior to conditioning [12, 13]. Moreover, when administered in conjunction with a p53 inhibitor, a subset of thymic fibroblastic reticular cells is restored and there is an increased expression of chemokine ligands (CCL21) on thymic lymphoid stroma, at least in irradiated mice . KGF-induced thymopoiesis can also be accelerated by agents such as leuprolide acetate (Lupron), a testosterone inhibitor . Finally, other hormonal treatments such as insulin-like growth factor-I and growth hormone/ghrelin have been shown to promote the expansion of TEC [8, 9, 16].
As regards thymic recovery, results of a very recent clinical trial in patients undergoing T-cell-depleted allogeneic HSC transplantation are of much interest; while the IL-7 cytokine was found to function primarily to expand pre-existing effector memory T cells, it also had a thymopoietic effect resulting in an increased T-cell diversity . Moreover, IL-23-induced IL-22 production has recently been shown to promote the entry of early thymic progenitors (ETPs) into the thymus . Importantly, this study found that administration of recombinant IL-22 significantly improved thymic recovery after irradiation .
It is therefore likely that optimal treatments aimed at enhancing thymic activity will result from a combination approach; promoting TEC stimulation, bone marrow (BM) hematopoietic function, and enhanced entry as well as expansion of progenitors. This goal can be fostered by therapeutic strategies that could include, but are not limited to, the use of cytokines, hormones, and sex steroid ablation together with approaches aimed at improving the function of T-cell progenitors themselves. In the remainder of this review, we will focus on the steps regulating the migration of hematopoietic progenitors to the thymus, their persistence within the thymus, and their capacity for T-cell differentiation.
MIGRATION OF HEMATOPOIETIC PROGENITORS AND ENTRY INTO THE THYMUS
In the context of HSC transplantation, the generation of T cells requires that the injected HSC themselves, or more likely precursor progeny of these HSC, migrate to the thymus. The processes controlling migration are likely to have evolved concomitantly with the differentiation of the thymus, in jawed vertebrates approximately 500 million years ago . Indeed, in zebrafish, the migrational journey of individual hematopoietic precursors into the thymus can be visualized over more than 12 hours . This migration is largely dependent on the presence of chemokine/chemokine receptor pairs dating back to cartilaginous fish . In mammals, there is a high redundancy with multiple molecules regulating entry of hematopoietic precursors into the thymus, including CCL19-CCL21/CCR7, CCL25/CCR9, and P-selectin/PSGL-1. Consistent with this overlap, abrogation of any one of these receptor-ligand interactions only modestly affects thymus colonization [22–27]. However, elimination of two interactions (CCR7 together with CCR9) can dramatically reduce the entry of ETPs [28, 29].
It is although important to note that even under conditions where the potential for T-cell differentiation is optimal including: (a) optimal expression of chemokine/receptor pairs on migrating hematopoietic cells and thymic tissue, (b) high level engraftment of transplanted HSCs in the BM, and (c) a thymus architecture which is conducive to hematopoietic precursor differentiation, the entry of progenitors into the thymus still remains a major bottleneck in T-cell differentiation. At least in mice, the thymus is not continually receptive to the import of hematopoietic progenitors, alternating between refractory and responsive periods . During refractory periods, representing approximately 3 weeks out of 4, donor hematopoietic progenitor cells do not appear to be capable of seeding the thymus and only undergo T-cell differentiation if they are directly injected into the thymus [30, 31]. This “gated” entry of thymocyte precursors into the murine thymus appears to be regulated by the absolute number of immature cells in the double negative 3 (DN3) population, a subset that varies with a cyclic period of approximately 4 weeks . While it is not clear whether this cycling also occurs in humans, if it is the case, it would signify that the timing of an HSC transplant is critical and by analogy, that there are significant time periods wherein intravenous (IV)-injected hematopoietic progenitors will only minimally “find” their way into the thymus.
HEMATOPOIETIC PRECURSOR DIFFERENTIATION WITHIN THE THYMUS
To better understand the differentiation of hematopoietic progenitors within the thymus, extensive experiments were carried out wherein physiologic thymocyte progenitors were isolated and directly injected into recipient thymi. These experiments showed that the thymic settling progenitors that normally settle the thymus are only able to give rise to a single round of thymocyte differentiation. These experiments led to the supposition that long-term thymocyte differentiation requires an on-going migration of donor progenitors from the BM to the thymus, with new BM precursors replacing resident thymocytes [32–35].
What is the phenotype(s) of the hematopoietic progenitor cells that take up residence in the thymus under physiological conditions? This remains somewhat of an open question as multiple phenotypes have been described and most of the markers are species specific. In mice, multipotent progenitors (MPP), retaining both myeloid and lymphoid potential, are generated from HSC and are characterized by the upregulation of the Fms-like tyrosine kinase receptor 3 (Flt3/CD135; resulting in LSK/CD135+ cells) . Subsets of MPP, characterized by expression of molecules such as RAG-1  and L-selectin , can also give rise to thymic precursors. Common lymphocyte precursors (CLPs) are more committed in that they lack myeloid potential and these progenitors can be distinguished as CLP-1 (Lin–Sca-1+CD117+/loCD127+CD135+)  and CLP-2 (Lin–Sca-1+CD117–CD127+CD135+B220+) . CLP-2 are thought to be the most differentiated population with T-cell potential before commitment to the B cell lineage . As regards the kinetics of thymopoiesis, it appears that MPP are capable of a more extended thymopoiesis than CLP, but are themselves not able to support a long-term thymopoiesis [22, 23, 37, 38, 41, 42]. In humans, the phenotype of early progenitors is quite different from mice with the presence of CD34 and the absence of CD1a, a member of the CD1 family of MHC-like glycoproteins, defining the earliest progenitors. Recently, expression of the CD10 and CD24 markers has been shown to be associated with lymphoid differentiation potential; CD34+CD10+CD24− cells have the potential to generate B, T, and NK lineages but not myeloid lineage cells . Nevertheless, it appears that this population of CLPs is skewed toward a B lymphocyte fate, whereas cord blood progenitors with a CD34+CD7+CD45RA+ phenotype express T/NK cell lineage-affiliated genes and show enhanced T lineage differentiation potential . Furthermore, recent studies have demonstrated that following ex vivo Notch ligand stimulation of CD34+ progenitors, it is the differentiated CD34+CD7+ CD45RA+ CLP subset that displays augmented in vivo thymus colonization and a more rapid T-cell differentiation than conventional CD34+ progenitors [45–47]. Thus, the precursors that settle the human thymus under physiological conditions are likely to share this CD34+CD7+CD45RA+ phenotype.
Once progenitor cells enter into the thymus, the thymic environment generally results in their acquisition of a short-lived T-cell precursor phenotype (as has been previously shown for common lymphocyte progenitors [23, 48]). It was initially thought that thymic settling progenitors lost myeloid potential before acquiring either T or B potential but a few recent studies showed that some progenitors within the thymus had lost B potential while retaining myeloid potential [49–51]. These data raised some controversy as it was questioned as to whether this myeloid differentiation occurs under physiological conditions within the thymus itself [52, 53] but it was very recently reported that ETPs give rise to the majority of thymic granulocytes . Thus, it appears that the progenitors that normally seed the thymus, while incapable of supporting long-term T-cell differentiation, can give rise to multiple blood cell lineages.
OBTAINING LONG-TERM T-CELL DIFFERENTIATION FROM THYMIC PROGENITORS WITH SELF-RENEWAL POTENTIAL
As described above, the absence of “bona fide” HSC in the thymus has been one of the holy grails in the field of thymocyte differentiation. This postulate dictates that either an HSC with self-renewing capacity cannot take up residence in the thymus, potentially due to a phenotype that is not conducive to its migration, or alternatively, an HSC that reaches the thymus cannot “settle” there or be maintained in an undifferentiated state.
It was therefore of interest to determine whether an HSC could sustain long-term thymopoiesis if it was forced into the thymus. In these experiments as well, BM precursors were not found to give rise to long-term thymopoiesis but it is important to note that in the vast majority of these studies, the thymus was perturbed by irradiation prior to transplantation [32–34, 38, 42, 55, 56]. Furthermore, even in the absence of irradiation, elegant experiments showed that intrathymically injected BM progenitors do not sustain long-term thymopoiesis in wild-type (WT) mice where there is a normal importation of precursors . However, recent studies, performed by our group and others, have shown that in immunodeficient mice, under conditions where competitive BM progenitors and/or early thymocyte progenitors are restricted, long-term thymus-autonomous T-cell differentiation can occur [57–59]. Thus, long-term thymopoiesis from injected progenitors can occur in the thymus under conditions where there is an available progenitor thymic “niche” (Fig. 1A).
Experiments performed in our lab, in mice with an immunodeficiency due to the absence of the ZAP-70 protein, showed that the intrathymic (IT) administration of allogeneic HSC progenitors results in significant increases in the absolute numbers of c-Kithigh ETPs and is associated with the generation of a normal medullary structure [57, 60] (Fig. 1C). Impressively, the ETP derived from intrathymically injected progenitors had a much more expanded lineage potential as compared to the BM-derived thymic settling cells that normally seed the thymus , with the former harboring T, B, granulocyte-macrophage (GM), and megakaryocyte-erythroid potential . Furthermore, it has recently been elegantly shown that thymus-autonomous T-cell differentiation can be sustained in mice with combined mutations in Rag2 together with either c-Kit or γc [58, 59]. These studies differed from those described above in that thymic lobes from newborn WT mice, rather than HSC, were transferred into immunodeficient mice with a paucity of their own bone marrow progenitors (Fig. 1B). Thymomas and lymphomas were detected in a subset of the thymus-transplanted mice [58, 59] but not following IT transplantation of HSC [57, 60, 61]. The precise mechanisms accounting for these important differences will need to be evaluated but it is possible that the transfer of a thymus under conditions where the progenitor cells are not capable of responding to growth factors such as stem cell factor (c-Kit ligand) or IL-7 may have selected for transformed progenitors. Conversely, upon IT transfer of progenitors under conditions where both host and donor cells are capable of responding to trophic signals, it appears that there is no increased pressure of transformation. The capacity to maintain long-term unbiased TCR repertoires for differentiated thymocytes will depend on the long-term persistence and contribution of early DN precursors to thymopoiesis. Harnessing this potential will be critical to strategies aimed at translating the potential of an IT progenitor cell transplantation approach to the clinic. As the majority of patients undergoing HSC transplantation are not characterized by severe combined immunodeficiency (SCID), it will be important to determine whether a thymus-autonomous renewal can occur in non-SCID patients under certain conditions of conditioning wherein distinct host cell subsets are eliminated, especially when the transplanted HSC are directly injected into the thymus. Furthermore, an HSC transplantation approach, combining IV and IT injections, could be advantageous in promoting a more rapid T-cell reconstitution, especially as thymic recovery following BMT has been shown to be limited by the number of progenitors entering the thymus . This possibility is discussed below.
In the context of IT transplantations performed in mice, the injected HSC and/or progenitors give rise to CD4 as well as CD8 T cells. However, other hematopoietic lineage cells, including B cells, monocytes, and megakaryocytes are either not generated at all or are present at only very low levels. Thus, in order to optimize T-cell differentiation from transplanted progenitors, in conjunction with differentiation of other blood lineage cells, one possibility would be to administer allogeneic HSC by both IV and IT routes (Fig. 2A). It will be necessary to determine whether the kinetics and/or dosing of HSC modulate the differentiation of one or another lineage cell.
The clinical applicability of an IT stem cell transplantation approach may be complicated by the fact that the phenotypes of hematopoietic precursor cells in mice and humans differ significantly. However, it is important to note that while these differences add a layer of complexity to this type of translational research, the factors regulating hematopoiesis as well as thymopoiesis are extremely well conserved. As such, it should be possible to translate an IT cell transfer strategy from mice to man. Indeed, we have shown that IT injection can be easily and safely used in anesthetized macaques, with thoracoscopy resulting in accurate targeting of the thymus and requiring a total procedure time of less than 15 minutes  (Fig. 2B). Furthermore, other groups have also reported on nonsurgical ultrasound-guided approaches that do not require exposure of the thoracic cavity [69–71]. These technological developments will foster the feasibility of such an approach. Other findings also strongly support continued translational studies for the development of IT -based HSC transplantation strategies including: (a) a significantly more rapid kinetics of T-cell reconstitution following IT as compared to IV administration of HSC and (b) long-term thymopoiesis by IT-injected precursors in nonconditioned recipients, even across histocompatibility barriers (Fig. 2) [23, 41, 57, 60–62]. It will therefore be important to determine the clinical settings in which an IT transplantation approach, potentially enhancing HSC differentiation to a T lineage fate, will improve short- as well as long-term patient outcome.
We thank all members of our lab for their scientific critique and support. We are grateful to G. Podevin, Y. Cherel, L. Dubreil, and Ph. Moullier for their collaboration on intrathymic injections in a primate model and thank G.P. and Ph.M. for permitting the use of their photograph showing endoscopic-mediated intrathymic injections in Figure 2B. S.d.B. was supported by a fellowship from the Association Française contre les Myopathies (AFM). V.Z. and N.T. are supported by CNRS and INSERM, respectively. This work was funded by Grant R01AI059349 from the National Institute of Allergy and Infectious Diseases, the AFM, the European Community (contract LSHC-CT-2005-018914 “ATTACK”), ANR, ARC, and INCA.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.