The success of allogeneic hematopoietic stem cell transplantation, a key treatment for many disorders, is intertwined with T cell immune reconstitution. The thymus plays a key role post allogeneic hematopoietic stem cell transplantation in the generation of a broad but self-tolerant T cell repertoire, but it is exquisitely sensitive to a range of insults during the transplant period, including conditioning regimens, corticosteroids, infections, and graft-versus-host disease. Although endogenous thymic repair is possible it is often suboptimal, and there is a need to develop exogenous strategies to help regenerate the thymus. Therapies currently in clinical trials in the transplant setting include keratinocyte growth factor, cytokines (IL-7 and IL-22), and hormonal modulation including sex steroid inhibition and growth hormone administration. Such regenerative strategies may ultimately enable the thymus to play as prominent a role after transplant as it once did in early childhood, allowing a more complete restoration of the T cell compartment.
First successfully performed in 1959, allogeneic hematopoietic stem cell transplantation (allo-HCT) encompassed the earliest forms of stem cell therapy, cancer immunotherapy, and individualized treatment (1). More than half a century later, allo-HCT remains a key therapy for a variety of both non-malignant and malignant disorders. However, major challenges persist, including overcoming the hurdle of prolonged post-transplant immunodeficiency through successful immune reconstitution (2).
Innate immunity, encompassing monocytes, granulocytes, and NK cells, normally recovers in the first weeks to months after allo-HCT (3, 4). In contrast, restoration of adaptive cellular and humoral immunity occurs over a longer period of at least 1–2 y.
Reconstitution of the T cell compartment is of importance, beyond their key role against opportunistic pathogens (5). T cells are the key effectors of graft-versus-tumor immune responses, highlighted by the efficacy of donor lymphocyte infusions in mitigating disease relapse in certain conditions (6). They are also the central mediators of graft-versus-host disease (GVHD), borne out in the reduction of this complication through use of T cell–depleted grafts (7). However, complete quantitative and qualitative restoration of the T cell compartment may take many years, or in some cases may not fully recover, and is linked to increased infection, relapse, secondary malignancies, and overall mortality in transplant recipients (8). T cell reconstitution is thus a critical determinant to the success of allo-HCT, and will be the focus of this review.
It is important to note that much of the ensuing discussion may also have relevance to patients undergoing standalone chemotherapy or autologous HCT (auto-HCT), both of which can impart a degree of T cell depletion with thymic damage. Indeed, the anti-tumor effect of these cytoablative strategies is believed to be caused not only by direct effects on the malignant cells, but also in part by induction of new T cell responses, as T cells expand within the lymphopenic environment (9). Thus, many of the regenerative strategies detailed below may similarly have utility in these settings. However, unlike chemotherapy or auto-HCT, the strategy of allo-HCT is fundamentally based upon utilizing the beneficial effect of donor T cells, and we believe therapies to improve T cell reconstitution will be especially beneficial in this area.
Post-transplant T cell immune reconstitution
T cell recovery after allo-HCT occurs through two distinct pathways. In the early post-transplant period, there is peripheral expansion of adoptively transferred donor T cells in the graft or recipient T cells that survive conditioning. This expansion occurs predominantly in the CD8+ memory T cell population, which in comparison with naive T cells is more responsive to cytokines (IL-2, IL-15, and IL-17) and previously encountered pathogens (such as reactivated viruses), and less dependent upon recognition of self MHC–peptide complexes for survival (10). This initial wave of recovery leads to a contraction and skewing of the TCR repertoire, and thus precludes an effective immune response against a broad range of Ags. Furthermore, T cells expanding in the periphery are not subject to central tolerance mechanisms, leading to the potential for graft-versus-host T cell responses.
Full T cell immune reconstitution thus relies on a second pathway, in which there is de novo production of naive T cells in the thymus of the transplant recipient (11). During this process, the thymus is seeded by lymphoid progenitors derived directly from the graft or arising from donor hematopoietic stem cells. T cell development is subsequently tightly regulated by the bidirectional cross-talk between thymic stromal cells and developing thymocytes (12). Thymic epithelial cells (TECs), endothelial cells, fibroblasts, and dendritic cells within the thymic stroma play a critical role in guiding the differentiation of T cell progenitors through distinct developmental stages, leading to the generation of mature naive CD4+ or CD8+ T cells expressing an MHC-restricted, Ag-specific TCR. This forms a broad but self-tolerant T cell repertoire, which is vital to the development of a strong adaptive immune response against pathogens and tumors, without leading to GVHD.
T cell reconstitution in mice
Similar to humans, mice show a fall in thymic output with age, associated with an expansion of memory T cells, and a constriction of the TCR repertoire. More recently, however, it has been demonstrated that despite such parallels, key differences also exist. Using in vivo kinetic labeling with heavy isotopes, it was shown that whereas naive T cells are very long lived in humans (13), their lifespan in mice is much shorter (14).
Furthermore, thymectomy experiments and analysis of recent thymic emigrants (RTEs), using both signal joint TCR excision circles (sjTRECs) and CD31 as markers, demonstrated that the underlying mechanism of T cell maintenance with aging differed significantly between species. Whereas in mice thymic output continues to act, even in old age, as the almost exclusive source of naive T cells, humans in contrast maintained their naive T cells pool in adulthood almost entirely through peripheral expansion (14).
This lack of need for thymic function in healthy human adults with aging may seem at odds with the concept that full T cell immune reconstitution after allo-HCT requires both an initial peripheral expansion of T cells, as well as T cell production by the thymus. These scenarios can be reconciled, however, if we consider that states of immense lymphodepletion, such as that induced by allo-HCT, diminish the T cell compartment to a greater degree than normal aging, and thus necessitate that the thymus is brought to the fore. This concept is supported by a study of adult HIV-infected patients, deficient in CD4+ naive T cells, treated with highly active antiretroviral therapy. In response to treatment, recovery of CD4+ naive T cells occurred in association with an increase in sjTREC levels, which is highly suggestive that adult thymic output was playing a key role in highly active antiretroviral therapy–induced immune reconstitution (15).
Thus, where there is greater dependency on increased thymic contribution, such scenarios align human T cell naive production much more closely with that of the mouse.
Thymus damage during allo-HCT
The importance of the thymus to T cell immune reconstitution must be reconciled, however, with the exquisite sensitivity of the gland to a range of acute insults. Several of these insults may be encountered during standalone chemotherapy or auto-HCT. The greater intensity of allo-HCT conditioning, combined with the ensuing broader complication rate (including those caused by allo-reactive T cells), means that allo-HCT may behave like a perfect storm, where many of these insults may occur in parallel, leading to significant acute thymic injury and reduced function.
The conditioning regimen of allo-HCT is used to ensure space for the donor graft, provide sufficient immunoablation to prevent graft rejection, and reduce tumor burden. The intensity of the conditioning regimen can vary and is classified as myeloablative (high dose), reduced intensity, and non-myeloablative (16). This has been traditionally achieved through a combination of the appropriate doses of chemotherapy and radiotherapy.
Both chemotherapeutic agents, such as cyclophosphamide, and radiation can induce acute thymic damage and loss of cellularity. All thymocyte populations are typically affected, but the double positive (DP) thymocyte population is particularly sensitive to irradiation. Although traditionally thought to be resistant to such damage, there is now increasing evidence to suggest that chemotherapy and radiation can also directly damage the thymic stroma, particularly TECs. Medullary TECs expressing the highest levels of MHC class II are especially sensitive to these insults, likely due to their higher rate of proliferation (17). Because these cells play a key role in negative selection (18), this selective depletion may in theory lead to profound defects in central tolerance. It is notable that certain cell populations, such as endothelial cells and innate lymphoid cells (ILCs), are relatively radio resistant (19, 20), and these cells can play a role in endogenous thymic regeneration.
Abs targeting T cells, such as ATG and alemtuzumab, are commonly incorporated into conditioning regimens as an additional therapy to reduce graft rejection, and lead to significant depletion of thymocytes. Although such therapies do not directly target the thymic stromal cells, the function of the stromal compartment may also be altered through loss of lymphoid-stromal cross-talk interactions (12).
Glucocorticoid hormones, acting through the nuclear glucocorticoid receptor, exert a wide range of immunosuppressive and anti-inflammatory effects, providing the rationale for their use in many conditions including autoimmune disease, allergic and inflammatory disorders, and lymphoid malignancies (21). Specific to the post-transplant setting, glucocorticoids are used as first-line therapy for GVHD (see below). Levels of endogenous glucocorticoids may also be elevated following allo-HCT in response to a wide range of stressors, including starvation, infection, and psychological stress, similarly leading to immunosuppression (22). Within the thymus, glucocorticoids induce apoptosis of DP thymocytes, which preferentially express the glucocorticoid receptor (23), in an Apaf-1 and caspase-9 dependent manner, leading to acute involution (24, 25). In view of the wide range of conditions associated with their elevated levels, glucocorticoids thus represent a common pathway mediating many episodes of thymic damage following allo-HCT.
The immunosuppressive nature of allo-HCT renders patients vulnerable to infections, including those caused by opportunistic organisms. Although previously thought to be an immune-privileged site (26) protected from infections and immune responses, increasing evidence over the last decade has highlighted that the thymus is a target for a range of pathogens, including bacteria, viruses, fungi, and parasites (27), leading to acute involution and impaired function.
In the context of allo-HCT, infection-induced thymic damage is most commonly caused by bacterial LPS released from gut–derived gram-negative bacteria. LPS leads to severe acute thymic atrophy that peaks within 3–5 d, characterized by loss of DP thymocytes (28). Other pathogens may directly invade the thymus. For example, CMV, which may reactivate following allo-HCT, infects TECs, leading to disruption of thymic architecture and degradation of the thymic microenvironment (29).
GVHD is a complication of allo-HCT, characterized by three distinct phases: 1) tissue damage from conditioning therapy or other causes; 2) activation of alloreactive donor T cells by host APCs; and 3) target tissue damage mediated by soluble and cellular effectors (7).
Although traditionally associated with damage to the gut, liver and skin, murine models of GVHD have demonstrated that the thymus is an extremely sensitive target of alloreactive T cells. Thymic GVHD leads to loss of the large DP subset, as a consequence of both a block in differentiation in early DN thymocytes and also increased apoptotic cell death in the DP population (30). TECs have likewise been identified in these models as direct targets for GVHD, and, through high levels of MHC expression, may act as APCs sufficient in themselves to prime alloreactive T cells (31). This raises the distinct possibility that GVHD can be restricted to the thymus even during subclinical GVHD, and thus in many cases may not be detected in the clinical setting yet still have detrimental consequences for T cell reconstitution. These preclinical models are reinforced by patient samples, in which GVHD has been shown to induce loss of cortical and medullary thymocytes, loss of TECs, decreased demarcation of the cortico-medullary junction, and disrupted thymic architecture (32, 33).
Such changes have been shown to lead to a decreased thymic output of naive T cells, as indicated by sjTRECs, and a distorted TCR repertoire (34). Furthermore, murine models of GVHD have also highlighted that thymic damage in such cases is associated with impaired negative selection of thymocytes and Treg development, which may be potential contributory factors to the development of chronic GVHD and autoimmunity post allo-HCT (35, 36).
Although a standard first-line treatment for GVHD, the ability of corticosteroids to induce thymic atrophy adds further complexity to the treatment of thymic GVHD, and alternate approaches may prove more beneficial to improve T cell reconstitution.
Enhancing thymic T cell reconstitution after allo-HCT
In the face of these protean insults, the acutely damaged thymus is capable, to a degree, of undergoing endogenous regeneration, allowing for spontaneous restoration of its function. A wide range of pathways has been implicated in such intrinsic repair, notably keratinocyte growth factor (KGF) (37) and, more recently, IL-22 signaling (19), both of which are discussed in more detail below.
The ability of the thymus to endogenously regenerate declines with age and with following repeated insults (38). This is particularly pertinent to allo-HCT, which is now being increasingly used in older patients, many of whom have had several previous lines of treatment. Thus, there is a need to develop exogenous strategies to enhance thymic function following allo-HCT to facilitate a more complete T cell reconstitution. We outline several different approaches below, focusing on those strategies that have entered clinical trials. These strategies are broadly applicable to allo-HCT performed across a range of conditions, but it is likely their success will depend, at least partly, on the degree of damage to the thymus caused by the intensity of the transplant and the degree of ensuing complications.
Keratinocyte growth factor.
KGF, also known as fibroblast growth factor 7, is a 22.5 kDa member of the fibroblast growth factor family. Within the thymus, it is produced by mesenchymal cells, αβ thymocytes, and γδ T cells, and acts in a paracrine manner to promote proliferation of epithelial cells (39). Specifically, KGF binds to an epithelial cell–specific splice variant of the fibroblast growth factor receptor-2 (FgfR2-IIIb), expressed on TECs (40). KGF knockout mice do not have an obvious T cell phenotype at baseline, but show a deficit in thymic recovery after immune insults, such as sublethal radiation and HCT (37).
It therefore follows that exogenous administration of recombinant KGF has been tested for its ability to enhance thymic regeneration. KGF protects epithelial cells from several thymic injuries, including radio- and chemotherapy, allogeneic or syngeneic HCT and GVHD (41–43). Mechanistically, activation of FgfR2-IIIb on TECS promotes stimulation of the p53 and NF-κB pathways, results in the upregulation of BMP2, BMP4, Wnt5b, and Wnt10b, and ultimately induces TEC proliferation in mice (44). As a result of thymic cross-talk, this is associated with thymocyte expansion and increased T cell export.
This work has been extended to non-human primate models, where adult rhesus macaques, receiving KGF after autologous HCT, had accelerated hematopoietic recovery, improved thymopoiesis (as evaluated by TCR excision circle analysis) and enhanced naive T cell recovery following transplant (45). However, functionally this was not associated with improved immunity against CMV reactivation nor an improved response to tetanus toxoid vaccination (45).
Human recombinant KGF (Palifermin) is a Food and Drug Administration-approved drug for the prevention of mucositis in recipients of high dose chemotherapy (46). Although there are as yet no clinical studies designed to investigate the sole efficacy of KGF on thymic function, the combination of KGF and leuprolide to enhance immune recovery after allo-HCT is being evaluated as part of a clinical trial (NCT01746849).
The 25-kDa cytokine IL-7 is implicated in several adaptive immune processes including thymopoiesis, B cell development, and lymph node organogenesis (47). It also represents a prosurvival factor for ILCs. The IL-7 receptor (IL-7R) is a heterodimer complex consisting of the CD132 common γ-chain receptor, expressed on lymphoid and myeloid cells, and the IL-7Rα (CD127), the expression of which is restricted to lymphoid cells. Produced within the thymus predominantly by cTECS, and in part by fibroblasts, IL-7 acts to promote proliferation, survival (through the PIK3/Akt pathway) and differentiation (through Jak-Stat signaling) of the developing thymocytes (48). The importance of its role is highlighted by the development of severe combined immunodeficiency syndrome in patients with a defect in the IL-7Rα or the common γ-chain (49, 50).
Several preclinical studies have demonstrated the beneficial effects of exogenous administration of IL-7, which enhances thymopoiesis and export of RTEs, in addition to increasing the homeostatic proliferation of mature peripheral T cells (51, 52). With respect to HCT, IL-7 can lead to accelerated T cell recovery in syngeneic and allogeneic murine models (53–55), and can also increase Ag-specific T cell responses to vaccination and viral infections (56, 57).
Extending this work to humans, several early studies, often focusing on lymphopenic patients, have revealed an ability of recombinant human IL-7 (rhIL-7) to enhance CD4+ and CD8+ T cell subsets, associated with preferential expansion of the RTE pool and a broadening of the TCR repertoire. In the transplant setting, a phase 1 clinical trial of recombinant human IL-7 (CYT107) in recipients of T cell–depleted allo-HCT (NCT00684008) demonstrated an increase in CD3+, CD4+, and CD8+ T cells in hIL-7 treated patients. Although no significant effects were reported on thymic output as measured by analysis of RTEs, patients receiving hIL-7 did show a broader TCR β repertoire diversity compared with untreated patients (58).
IL-22, a member of the IL-10 cytokine family, has gained increasing interest for its potential beneficial effect in promoting epithelial integrity and antimicrobial immunity at mucosal surfaces (59). Produced by Th17 cells and ILCs, IL-22 binds to a heterodimeric cell surface receptor, IL-22R, composed of IL-10R2 and IL-22R1 subunits (60–63). IL-22R is expressed on cells of epithelial origin, such as keratinocytes, intestinal or lung epithelial cells, and hepatocytes, and is absent on the cells of the immune system (59, 64).
IL-22 plays an important role in mediating endogenous recovery of thymus function after acute damage (19). Depletion of DP thymocytes triggers upregulation of IL-23 by dendritic cells, which in turn stimulates the production of IL-22 by intrathymic ILCs. IL-22 directly promotes the proliferation and survival of TECs, leading to regeneration of the supporting microenvironment and, ultimately, to rejuvenation of thymopoiesis. A similar effect is observed after exogenous administration of murine recombinant IL-22. The corollary of this endogenous repair pathway is the possibility that IL-22 may thus have utility as a potential therapeutic option to stimulate thymic recovery in immunocompromised patients.
In relation to its ability to exert protective effects on the gastrointestinal epithelium (64), a phase IIa clinical study has recently been initiated to evaluate the safety and tolerability of human recombinant IL-22 (hrIL-22) in conjunction with systemic corticosteroids in the treatment of gastrointestinal acute GVHD in patients receiving HCT (NCT02406651). Importantly, peripheral T cell counts will be evaluated as a part of the study, allowing for further investigation of hrIL-22 as a thymus boosting therapy, which may impact T cell reconstitution.
Sex steroid inhibition.
The ability of sex steroids to modulate the structure and function of the thymus has been recognized for more than a century (65, 66), and surgical or chemical sex steroid inhibition (SSI) has been shown to promote thymic growth in several preclinical and clinical studies (67, 68). Mechanistically, recent studies have demonstrated that SSI can directly promote the expression of CCL25 (69) and the Notch ligand DLL4 (70), as well as promoting the function of hematopoietic stem and progenitor cells (71–73).
In the context of immune insults including radiotherapy, chemotherapy, and HCT, studies in mice have shown that SSI accelerates thymic recovery leading to improved T cell reconstitution (74–76). Clinical studies using luteinizing hormone-releasing hormone (LHRH)-agonists, which desensitize the LHRH-receptor and ultimately lead to the inhibition of luteinizing hormone and follicle stimulating hormone release, have demonstrated accelerated engraftment and enhanced T cell reconstitution (as evidenced by a broader TCR repertoire and improved T cell function) in HCT recipients (77). Currently, clinical trials are ongoing to test the effects of the LHRH-agonist (Lupron), either alone (NCT01338987) or in combination with KGF (Palifermin) (NCT01746849), in promoting immune recovery of allo-HCT patients.
The anterior pituitary hormone growth hormone (GH) may modulate several aspects of immunity (78, 79), either directly or through its principal mediator, insulin-like growth factor 1 (IGF-1). With regards to the thymus, the receptors for both GH and IGF-1 are expressed by both TECs and thymocytes (78, 80, 81).
GH-deficient mice exhibit defects in DP thymocyte development and thymic hypoplasia, indicating a key role for this hormone in thymic function. Likewise, GH or IGF-1 administration promotes improved thymic cellularity, increased TCR diversity and enhanced recovery of the hematopoietic compartment in immunocompromised and aged animals (80–83). Several intrinsic and extrinsic mechanisms have been identified for the beneficial effects of GH on thymic function such as enhanced proliferation of TECs and trafficking of common lymphoid progenitors into the thymus (81, 84, 85).
These murine studies have been translated to early encouraging results in the clinical setting. Recombinant human GH enhances thymic recovery and immune reconstitution in both HIV-infected patients (86, 87) and GH-deficient adults (88). More recently, a phase 1 clinical trial was performed to test the safety and efficacy of recombinant human GH in improving immune reconstitution post unrelated cord blood transplant (NCT00737113), which should offer greater insight into the role of this hormone in the allo-HCT setting.
Although yet to be tested in humans, numerous novel preclinical approaches offer further promise of improving thymic T cell reconstitution. For example, tyrosine kinase activity may be targeted for beneficial effect. The cytokine FMS-like tyrosine kinase 3 ligand enhances thymopoiesis in mice (89), and the use of sunitinib to inhibit tyrosine kinase in early thymic progenitors has been postulated to improve thymic niche accessibility (90). Cellular therapy, such as the use of ex vivo generated T cell precursors, has also been shown to enhance thymopoiesis in mouse models of allo-HCT (91).
The last two decades have seen a surge in our understanding of the interplay between allo-HCT, immune reconstitution, and thymic biology (Fig. 1). This has been paralleled by the development of novel strategies to rejuvenate the transplant-damaged thymus. The challenge will be to translate these advances into a definitive clinical therapy, allowing the thymus to play a key role in posttransplant T cell reconstitution.
This work was supported by National Institutes of Health Awards R00-CA176376 (J.A. Dudakov), R01-HL069929 (M.R.M.v.d.B.), R01-AI080455 (M.R.M.v.d.B.), R01-AI101406 (M.R.M.v.d.B.), P30 CA008748 (C.B. Thompson), Project 4 of P01-CA023766 (M.R.M.v.d.B.), R01-HL124112 (R.J. Jenq), and R01HL123340-01A1 (Cadwell). Support was also received from The Lymphoma Foundation, The Susan and Peter Solomon Divisional Genomics Program, and the Memorial Sloan-Kettering Cancer Center Cycle for Survival. This project has received funding from the European Union’s Seventh Framework Programme for research, technological development, and demonstration under Grant Agreement 602587. M.S.C. was supported by a Lady Tata Memorial Trust International Award for Research in Leukaemia.
Abbreviations used in this article:
allogeneic hematopoietic stem cell transplantation
insulin-like growth factor 1
innate lymphoid cell
keratinocyte growth factor
luteinizing hormone-releasing hormone
recent thymic emigrant
signal joint TCR excision circle
sex steroid inhibition
thymic epithelial cell.
The authors have no financial conflicts of interest.