|Home | About | Journals | Submit | Contact Us | Français|
The pursuit of clinical transplant tolerance has led to enhanced understanding of mechanisms underlying immune regulation, including the characterization of immune regulatory cells, in particular antigen-presenting cells (APC) and regulatory T cells (Treg), that may play key roles in promoting operational tolerance. Dendritic cells (DC) are highly-efficient APC that have been studied extensively in rodents and humans, and more recently in non-human primates. Owing to their ability to regulate both innate and adaptive immune responses, DC are considered to play crucial roles in directing the alloimmune response towards transplant tolerance or rejection. Mechanisms via which they can promote central and peripheral tolerance include clonal deletion, the induction of Treg, and inhibition of memory T cell responses. These properties have led to the use of tolerogenic DC as a therapeutic strategy to promote organ transplant tolerance. In rodents, infusion of donor- or recipient-derived tolerogenic DC can extensively prolong donor-specific allograft survival, in association with regulation of the host T cell response. In clinical transplantation, progress has been made in monitoring DC in relation to graft outcome, including studies in operational liver transplant toleranceAlthough clinical trials involving immunotherapeutic DC for patients with cancer are ongoing, implementation of human DC therapy in clinical transplantation will require assessment of various critical issues. These include cell isolation and purification techniques, source, route and timing of administration, and combination immunosuppressive therapy. With ongoing non-human primate studies focused on DC therapy, these logistics can be investigated seeking the optimal approaches. The scientific rationale for implementation of tolerogenic DC therapy to promote clinical transplant tolerance is strong. Evaluation of technical and therapeutic logistic issues is an important next step prior to the application of “therapeutic” DC in clinical organ transplantation.
Reduced dependence on immunosuppressive drug therapy and the induction of donor – specific tolerance are major objectives for numerous studies aimed at improving clinical outcomes in transplantation. Much recent research has focused on better understanding of the cellular and molecular mechanisms underlying immune regulation, as well as exploring novel therapies to induce donor-specific tolerance. In recent years, considerable interest has focused on the prospective value of targeting specific immune regulatory cells (that include dendritic cells [DC], regulatory T cells [Treg], and myeloid-derived suppressor cells) or exploiting these cell populations to promote the long sought-after goal of clinical transplant tolerance.
DC are rare, uniquely well-equipped, functionally-diverse professional antigen (Ag)-presenting cells (APC) that are regarded as key instigators and regulators of innate and adaptive immunity. They play major roles in directing the immune response towards tolerance or immunity. DC are the best-characterized APC in the context of organ transplantation, where that favor graft acceptance are considered tolerogenic (tol DC), and those that induce graft rejection are viewed as immunogenic. Over the past two decades, beginning with recognition of their role in intra-thymic tolerance [1, 2], DC with tolerogenic properties have been studied extensively in small animal models and in human in vitro systems. Detailed background literature on both mouse and human tol DC and their relation to regulation of alloimmunity and the outcome of organ transplantation has been reviewed in [3–9]. Due to their inherent tolerogenicity, DC are considered potential therapeutic targets or cellular agents (negative cellular vaccines) in transplantation and autoimmune disease [10–12], contrasting with their potential immunostimulatory properties as anti-cancer vaccines . While in numerous experimental animal models, DC have proved to be effective tolerogenic APC, the current challenge is whether these benefits can be translated into effective strategies to promote transplant tolerance in humans (Figure 1).
DC constitute a heterogeneous rather than a discrete cell population, that arises from CD34+ cells within the bone marrow (BM). They are found both in the thymus and secondary lymphoid tissue, in the circulation, in peripheral tissues such as skin, mucosal surfaces, intestine, lung, liver, and in maternal decidua. As APC, they are specialized in Ag capture, processing, and presentation; upon activation (as occurs following organ transplantation), they display enhanced capacity to migrate from the periphery via afferent lymph or blood to regional lymphoid tissues, where they interact with T cells and regulate their function.
Multiple DC subsets have been described in various tissues, especially the spleen [14, 15], since the first description of myeloid DC in mouse spleen by Steinman and Cohn in 1973/74 [16, 17]. In humans, the two major and intrinsically distinct subpopulations of DC are the “conventional” myeloid DC (mDC) and plasmacytoid DC (pDC). mDC and pDC are distinguished by both cell surface markers and function [18, 19]. Under inflammatory conditions, and following Toll-like receptor (TLR) ligation and Ag uptake (e.g. pathogen or donor alloAg), mDC migrate to T cell areas of secondary lymphoid tissue to initiate adaptive immunity. On the other hand, pDC that express a genetic profile more closely resembling lymphoid cell development , recognize nucleic acids (both microbial and self) and produce interferon (IFN)-α, an important component of innate immune responses. Tolerogenic properties of both mDC and pDC (mouse and human) have been reported extensively, including the capacity of human mDC to induce Ag-specific tolerance in vivo, in healthy volunteers [21, 22].
Generally, human DC have been identified as major histocompatibility complex (MHC) class II+ (HLA-DR+) cells in peripheral tissues. In the circulation, they are lineage negative (Lin−) and express blood DC Ags (BDCA); while mDC are Lin−/MHC II+/CD11c+/CD123− (IL-3Rα)/BDCA-1+, pDC are Lin-/MHC II+/CD11c−/CD123+/BDCA-2+ and Ig-like transcript 7+. In mice, mDC are CD11c+CD11b+NK1.1−, whereas pDC are B220+CD11cloCD11b−Gr-1+ and express the murine pDC Ag-1/BM stromal cell Ag (BST)-2/CD317. Other mouse pDC markers include Siglec-H and the chemokine receptor CCR9. Based on their phenotypic and functional characteristics, DC can be identified at various stages of maturation, from immature through so-called `semi-mature' to mature. Immature DC express high endo-and phagocytic capacity, low levels of MHC class II and co-stimulatory molecules (CD40, CD80 and CD86), and low T cell stimulatory ability. These cells are associated with induction of T cell anergy and generation of Treg. By contrast, mature DC that express high levels of MHC class II, CD80 and CD86, have strong migratory capacity, associated with upregulation of the lymphoid homing receptor CCR7. They secrete the T helper (Th) 1 cell-driving cytokine IL-12. While in vitro, mature DC also expand Treg from naïve mice in the presence of IL-2, and increase their suppressive capacity  and ability to suppress autoimmunity , they also potently expand effector T cells.
Owing to their poor T cell stimulatory capacity, immature mDC have been considered the prototypic tolerogenic DC. The first reports to address the tolerogenic potential of immature DC (CD205+, MHC II+, CD40dim CD80dim, and CD86dim) in experimental (murine cardiac) allograft survival were by Fu et al  and Lu L et al . In these studies, immature donor BM-derived mDC propagated in vitro in granulocyte-macrophage colony stimulating factor (GM-CSF) and administered 7 days before vascularized heart graft transplantation, significantly prolonged graft survival, either alone or in combination with co-stimulation blockade. Similar observations were reported by Lutz et al , who showed that immature mDC, resistant to maturation, could prolong haplotype-specific heart allograft survival indefinitely (>100 days) also when given 7 days before transplantation. Subsequently, evidence accumulated that various phenotypically diverse DC could also promote transplant tolerance. Thus, semi-mature, maturation-resistant, `alternatively-activated' and genetically-modified DC have all been shown to prolong allograft survival and promote tolerance, often in conjunction with costimulation blockade, lymphocyte-depleting Abs or conventional immunosuppression .
DC generated in vitro in the presence of tumor necrosis factor (TNF)-α, IL-10 + transforming growth factor beta (TGF-β) or dexamethasone, display a “semi-mature” phenotype, with intermediate levels of MHC II, CD40, CD80, and CD86 expression. These DC can markedly prolong organ allograft survival and inhibit graft-versus-host disease (GVHD) following hematopoietic stem cell transplantation [28–30]. “Semi-mature” DC, modulated by vitamin D3 and dexamethasone, can express the inhibitory molecules immunoglobulin (Ig)-like transcript – 3 (ILT3) and the B7 family member programmed death ligand – 1 (PDL-1= B7-H1) that plays a role in the induction of Treg . E-cadherin is an epithelial adhesion molecule that is also expressed by epidermal DC, i.e. Langerhans' cells (LC). Surface ligation of E-cadherin inhibits LC maturation , whereas disruption of E-cadherin-mediated cell-cell contact by DC induces a “mature” phenotype with tolerogenic capacity. Such “mature” DC are able to induce CD4+T cells that produce IL-10 instead of IFN-γ, and provide protection against induction of experimental autoimmune encephalitis .
Apart from transplantation and autoimmunity, DC have been implicated in regulation of immune reactivity to the semi-allogeneic fetus. DC have been shown to be central to the control of immune tolerance at the maternal-decidual interface. In a study of the role of DC in maternal – fetal tolerance, bacterial lipopolysaccharide ( LPS) and IFN-α activation of human chorionic gonadotropin – treated, BM-derived and splenic DC, downregulated MHC II expression, but maintained high CD80 and CD86 expression, maintained high IL-12, but significantly increased IL-10 production , suggesting mechanisms by which these DC might regulate immune reactivity directed against the fetus.
Thus, while the correlation between DC maturation and functional characteristics is well recognized, these data highlight the complexity of the relationship between DC phenotype and their tolerogenic potential.
DC are essential for the maintenance of both central and peripheral tolerance in the normal steady state, where depletion of all subsets DC (including pDC and LC) in mice results in fatal autoimmunity . Central tolerance is achieved through negative selection of self- or foreign Ag-reactive thymocytes and is a highly-efficient process mediated by APC and the induction of Treg in the thymus. Circulating DC have been shown to migrate to the thymic medulla through a three-step adhesion cascade involving P-selectin, interactions of the integrin very late Ag 4 (VLA-4) with its ligand vascular cell adhesion molecule -1 (VCAM-1), and pertussis toxin-sensitive chemoattractant signaling . Using a thymic transplantation system, Proietto et al  have demonstrated that DC in the periphery can migrate into the thymus, where they efficiently induce Treg generation and negative selection. In diabetes-prone, non-obese diabetic (NOD) mice, BM-derived DC treated with thymic stromal lymphopoietin (TSLP) produced by Hassall's corpuscles in the thymic medulla, acquire a tolerogenic phenotype, induce the conversion of naïve T cells into functional CD4+CD25+Foxp3+ Treg, and provide protection from autoimmune diabetes . Similarly to thymic mDC, thymic pDC can drive nTreg development and also induce Treg [39, 40]. The Treg induced by pDC are more efficient producers of IL-10 than those induced by thymic mDC. These data emphasize the role of DC in both the development and education of T cells in the thymus.
In the absence of inflammation, tol DC maintain peripheral tolerance to self Ags through various inter-related mechanisms, including T cell deletion, induction of T cell anergy and induction of Treg, expression of immunomodulatory molecules, and production of immunosuppressive factors (e.g. IL-10; TGFβ; and indoleamine dioxygenase [IDO]). APC expressing IDO play a critical role in maintaining peripheral tolerance. Prolongation of pancreatic islet allograft survival in mice by soluble cytotoxic T lymphocyte-associated Ag 4 (CTLA4) fusion protein (CTLA4-Ig) requires intact tryptophan catabolism in the recipient, likely due to the ability of CTLA4-Ig to upregulate IDO production by host DC . In the mouse spleen, an IDO+CD19+ DC subset exhibits a mature phenotype in the steady-state, and synthesizes high amounts of IDO that mediate T cell suppression. IDO+ CD19+ DC increase their production of IDO following CD80/86 ligation by CTLA4 [42, 43] or TLR9 ligation , and require autocrine release of IFN-α for signal transducer and activator of T cells (STAT) 1 activation and IDO up-regulation. Inhibition of IDO expression by silencing RNA (siRNA) inhibits DC-mediated suppression of Tcell proliferation . Tryptophan-deprived DC show reduced capacity to stimulate T cells, express high ILT3 and LT4, and induce suppressive CD4+CD25+Foxp3+ Treg . In rodents, CD103+ DC in mesenteric lymph nodes and intestinal mucosa have been shown to express IDO. Inhibition of IDO activity induces Th1 and Th17 T cells in vivo, and prevents the development of Treg specific for oral Ags . In an autocrine manner, murine tolerogenic CD8+ DC secrete TGF-β which maintains IDO activation . In a mouse model of collagen-induced arthritis, LPS-stimulated DC upregulate IDO expression, induce markers for Treg (Foxp3, TGF-β1 and CTLA-4) in vivo, and improve arthritis scores when injected after immunization .
The programmed death-1 (PD-1) receptor is an inhibitory molecule expressed on activated T cells and its ligands, PD-L1 and PD-L2, contribute to the negative regulation of T lymphocyte activation and peripheral tolerance. PD-1-PD-L1 interactions maintain peripheral tolerance by mechanisms distinct from those of CTLA-4. Ab-mediated blockade of PD-1 or PD-L1 results in enhanced T cell-DC interaction and autoimmune diabetes . It has been reported that PD-L1 and PD-L2 are expressed at very low levels on immature LC. Mature epidermal LC lack PD-1 expression, but express high levels of PD-L1 and PD-L2. Also, blockade of PD-L1 and/or PD-L2 on dermal DC results in enhanced T cell activation . Thus, LC not only have tolerogenic properties, but also have regulatory functions that can counteract the pro-inflammatory activity of surrounding keratinocytes.
Heme oxygenase-1 (HO-1) is an intracellular enzyme that degrades heme and inhibits inflammation in vivo. Human DC dramatically decrease HO-1 expression during their maturation in vitro. By contrast, cobalt protoporphyrin induction of HO-1 expression is associated with downregulation of LPS-induced human DC maturation, secretion of anti-inflammatory cytokines, and inhibition of alloreactive T-cell proliferation. Also, induction of HO-1 inhibits the production of reactive oxygen species (ROS) induced by LPS in human DC, although their ability to produce IL-10 is maintained .
The non-classical MHC class I Ag HLA-G is expressed on human DC and T cells, and is involved in tolerance induction . HLA-G is a key molecule in the regulation of allo-immune responses at the human fetal-maternal interface. It interacts with three inhibitory receptors: ILT2, ILT4, and killer cell Ig--like receptor 2DL4 (KIR2DL4). HLA-G modulates immune responses through protection from NK cell- and cytotoxic T-lymphocyte-mediated cytolysis, inhibition of allogeneic T-cell proliferation, and modulation of DC function. HLA-G/ILT4 interaction promotes the development of tol DC [54, 55].
During pregnancy, a population of immature decidual DC appears to differentially express specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN; a C-type lectin). Human DC-SIGN+ cells are CD83+CD25+ mDC, with a high capacity to stimulate naive T-cell proliferation. DC-SIGN mDC appear during early pregnancy in the human endometrium and are closely associated with inter-cellular adhesion molecule -3 - expressing large granular lymphocytes (LGL), which have been shown to produce high local concentrations of GM-CSF and IL-10. Due to the high affinity of DC-SIGN for ICAM-3, it has been suggested that interaction between these two cell populations (DC-SIGN+ and LGL) may prevent the interaction of DC-SIGN+ cells with T cells and further maturation of DC-SIGN into potent immunostimulatory DC .
Naturally-occurring Treg are thymic-derived CD4+ CD25+ Foxp3+ T cells that constitute a significant population (5–10%) of peripheral T cells in normal mice. In humans, their incidence among CD4+ T cells declines with age, starting from a range of 4–10% in cord blood, decreasing to 1–4% in young adults and to 0.5–1.5% in healthy elderly donors . In addition to their role in T cell anergy and deletion, DC in the periphery are pivotal in Treg induction, where their ability to induce and interact with Treg is critical for their tolerogenic effect. Also, conversely, Treg can promote the tolerogenic phenotype and capacity of DC. CD4+CD25+ Treg activated with anti-CD3 overexpress CTLA4 and can condition DC to express IDO functional activity. This effect requires B7 expression and IFN-γ production by DC. Additionally, in the same study, CD4+CD25+Treg activated with LPS could induce tryptophan catabolism in DC through a CTLA4-independent but cytokine-dependent mechanism .
DC regulate the suppressive ability, expansion and/or differentiation of Treg in vitro . This dynamic interaction between DC and Treg has been confirmed in vivo, where loss of DC results in a loss of Treg, increases IFN-γ- and IL-17-producing T cells, and is associated with increased risk of autoimmune disease . A “feedback” regulatory loop has been proposed, where, under steady-state conditions, the decrease in Treg results in a fms-like tyrosine kinase-3 (Flt3)-dependent increase in DC, which leads to increased homeostatic expansion of Treg. Additionally, absence of CTLA4 on Treg reduces the frequency of DC expressing IDO in mesenteric lymph nodes. Moreover, mice genetically deficient in CCR4 (CCL22 receptor) on Treg show markedly reduced IDO expression by mesenteric lymph node DC .
IL-10 contributes to the induction of anergy and the development of suppressive T cells and plays a fundamental role in the reciprocal effects of tol DC and Treg. Costimulation of CD4+T cells through the inducible costimulatory (ICOS) molecule results in IL-10 production. On the other hand, blocking of the ICOS/ICOS-L interaction abolishes anergy induction in human CD4+T cells by immature DC. CD4+T cells from ICOS-deficient patients are completely resistant to anergy induction and differentiation into suppressor T cells, even after supplementation of IL-10 . Tol DC secreting high levels of IL-10 (DC10) in the absence of IL-12 induce adaptive IL-10-producing regulatory type-1 (Tr1) T cells .
Unlike mDC, maturing pDC increase their expression of ICOS-L that is involved in de novo differentiation of IL-10-producing Treg . Moreover, in mice, pDC can down-regulate T cell responses through upregulation of IDO, a phenomenon that requires engagement of CD200 or glucocorticoid-induced TNF receptor ligand on the APC and autocrine release of type-I IFNs [65, 66]. Furthermore, human pDC matured by TLR9 ligation promote Treg differentiation , and when activated with IL-3 plus CD40 ligation, induce CD8+ Treg . CD40 ligand stimulation of mature human pDC but not mDC, primes naive CD8+T cells to become IL-10-producing Treg . In parallel, human CD8+CD28− T suppressor cells can also inhibit the function of mDC by increasing expression of the inhibitory receptors ILT3 and ITL4, that interfere with nuclear factor -κB-mediated activation of the DC .
Heterogeneity in the memory cell population of the human immune system can lead to donor cross-reactive memory T cells in transplant recipients. High frequencies of Ag-experienced memory T cells correlate with enhanced rejection risk and are a major hurdle to tolerance induction in humans. Human memory T cells are resistant to depletion (e.g. by Campath-1H) . Also, there is growing evidence that effector memory T cells are relatively resistant to costimulation blockade. Tolerance induction protocols based on CTLA-4–Ig have been found to be ineffective in recipients that possess cross-reactive, virus-elicited donor-reactive effector memory T cells. These recipients show lack of attenuation of donor-reactive CD8+ T-cell responses . Significantly, human `alternatively-activated', monocyte-derived DC (moDC) (conditioned by exposure to vitamin (Vit) D3 and dexamethasone, then LPS-stimulated), induce anergy in allogeneic memory CD4+ T cells which is IL-10-independent . The induction of anergy in these memory T cells can be prevented by exogenous IL-12p70, but not by depleting CD4+CD25hi T cells. Moreover, the tol DC do not expand Foxp3+ T cells preferentially from memory T cells. Similarly, earlier studies showed that co-cultures of dexamethasone-conditioned human moDC and alloreactive CD4+ memory T cells rendered the T cells hyporesponsive to restimulation with mature DC. Memory CD8+ T cells are a potent barrier to transplant tolerance induction. Adoptive transfer of fully-differentiated, long-lived CD8+ memory T cells genetically-targeted in vivo by steady-state DC, results in initial expansion, followed by partial deletion, where residual T cells are unresponsive to further antigenic stimulation . TLR-stimulated allogeneic pDC induce CD8+ Treg that inhibit memory T cell alloimmune responses, . The important ability of DC to inhibit memory T cell responses underscores the potential of these cells to regulate alloimmune responses in vivo.
DC present alloAg to T cells through the direct, indirect or `semi-direct' pathways of allorecognition [76–79]. Following transplantation, donor DC migrate from the graft and present donor MHC molecules to allospecific T cells via the direct pathway. Recipient DC that have processed donor alloAg present allopeptides on self (recipient) MHC molecules to donor-reactive T cells through the indirect pathway. Through the semi-direct pathway, T cells recognize donor MHC molecules transferred, intact, from donor cells to the surface of recipient DC . Also, DC can acquire MHC molecules or allopeptides from other cells by the transfer of vesicles, or fragments of plasma membrane,- a process known as cell `nibbling' . The direct pathway is considered the main mechanism that leads to acute graft rejection, which decreases in influence with time after transplantation. On the other hand, the indirect pathway becomes the main mechanism of allorecognition at later times after transplantation, and is strongly implicated in chronic rejection.
In vitro propagation of tolerogenic donor- or recipient-derived DC has been used extensively as an experimental approach to target the pathways of allorecognition, with the aim of prolonging transplant survival, while reducing dependency on immunosuppressive drugs. The development of techniques to propagate large numbers of tol DC in vitro has provided the basis for ascertaining the ability of these cells to downregulate both host-versus-graft and graft-versus-host immune responses mediated by T cells (in the context of hematopoietic stem cell transplantation). There have been numerous reports of indefinite murine allograft survival following infusion of either donor- or recipient-derived tol DC (reviewed in ), administered either alone or in combination with conventional immunosuppressive drugs, costimulation blockade or T cell-depleting agents (Figure 2).
Intravenous infusion (7 days before transplantation) of in vitro-generated donor immature mDC significantly prolongs the survival of murine vascularized (heart) or non-vascularized (pancreatic islet) allografts [25, 82]. When combined with CD40–CD154 blockade, anti-T-cell therapy or, by using `maturation-resistant' mDC, mouse cardiac allograft survival can be prolonged extensively or indefinitely, without affecting acute rejection of third-party transplants [26, 27]. Alternatively, administration of donor blood monocyte-derived immature DC (MHC II+, CD80low, CD86−), together with post-transplant total lymphoid irradiation and anti-thymocyte globulin, prolongs rat heart allograft survival indefinitely .
The role of pDC in experimental transplant tolerance has recently been examined. Intravenous administration of donor mouse pre-pDC before transplantation prolongs the survival of vascularized cardiac allografts significantly . This effect was not found to be Ag-specific however, as third-party BM-derived pDC also significantly prolonged transplant survival, suggesting a non-specific immunosuppressive effect. Flt3L-mobilized donor splenic pDC also significantly prolong cardiac allograft survival in the absence of immunosuppression, an effect that is enhanced by combination with anti-CD154 mAb therapy . In mice, pDC precursors (pre-pDC) facilitate allogeneic haematopoietic stem cell engraftment and promote tolerance to skin allografts .
A recent study in mice compared the thymus-homing capacity and tolerogenic function of in vitro-propagated allogeneic BM-derived mDC and pDC. Only Flt3L-induced DC (containing both mDC and pDC) and not GM-CSF-propagated mDC migrated to the recipient thymus after intravenous injection, and induced negative selection of donor-reactive CD4 and CD8 single positive thymocytes. Notably, central and peripheral tolerance induced by Flt3L-mobilized donor DC infusion led to prolonged donor-specific skin allograft survival .
Donor DC may also impact the indirect pathway by “transporting” alloAg to recipient DC in host secondary lymphoid organs. Following subcutaneous footpad injection of BALB/c-derived DC (IAd, IE+) into C57BL/6 recipient (IAb, IE−) mice, host DC in the draining lymph nodes express the IAb–IEα52–68 MHC–peptide complex . In a recent study, maturation-resistant donor-derived DC were injected systemically and shown to be processed and donor Ag presented to recipient T cells by CD11chi host DC. The injected donor DC were successful in prolonging cardiac allograft survival through delivery of donor Ag to recipient APC . These donor-derived DC may not act directly through presentation of donor Ag. Rather that may serve as a source of donor Ag for recipient DC for presentation to indirect pathway T cells, down-regulation of the anti-donor response, and prolongation of allograft survival, similar to donor-specific transfusion. These conclusions raise important points regarding the underlying tolerogenic mechanisms of both donor versus recipient-derived DC. In the context of human organ transplantation, induction of donor-specific tolerance using donor-derived DC would require precise assessment of appropriate associated immunosuppressive therapy.
Most experimental protocols that have used donor-derived DC to prolong organ allograft survival require preparation of the DC several days before transplantation, which is applicable to live-donor renal and liver transplantation, but not to organ transplantation from deceased donors. Once injected, these donor-derived DCs may be eliminated by host natural killer (NK) cells, as shown in mice . The importance of reprocessing donor alloAg by host DC following administration of DC is yet to be clarified and has significant therapeutic implications for the use of donor versus recipient DC to improve the outcome of clinical organ transplantation (see below). Administration of recipient-derived (autologous) tol DC loaded with donor alloAg be more advantageous, since these DC can be generated in vitro at any time, have an inherent potential to regulate the indirect pathway and thus potentially impact the development of chronic rejection, and may be associated with less risk of host sensitization.
Recipient-derived DC pulsed with donor cell-derived Ags, including allopeptides, have been used to down-regulate the anti-donor T cell response and promote transplant tolerance. Intrathymic inoculation of recipient BM-derived mDC pulsed with a donor allopeptide, in combination with a single dose of anti-lymphocyte serum, one week before transplantation, induced donor-specific tolerance to heart and pancreatic islet allografts in rats [91, 92]. This approach is potentially relevant to pediatric heart transplantation, where the young thymus is accessible during transplant surgery. By contrast, intravenously-injected mDC did not home to the thymus, but the presence of the thymus was critical for the induction of allograft tolerance. This contrasts to the aforementioned capability of donor-derived pDC and mDC (Flt3L-mobilized) to home to recipient thymus after their systemic injection.
We have reported that repetitive (×2) intravenous administration of recipient-derived, maturation-resistant rapamycin-conditioned BM-derived mDC, pulsed with cell-free lysate from donor splenocytes, prolongs cardiac allograft survival indefinitely in 40% of fully MHC-mismatched recipient mice . Interestingly, the tolerogenic potential of non-pulsed, recipient-type immature BM-derived DC, administered the day before transplantation  to promote donor-specific tolerance, is potentiated by suboptimal immunosuppression with LF15-0195, a deoxyspergualin (DSG) analogue that inhibits DC maturation in vivo . Recipient BM-derived mDC generated in vitro with GM-CSF, IL-10 and TGFβ1, and then stimulated with LPS (`alternatively-activated' DC, that produce IL-10 but little IL-12), prevent lethal GVHD following allogeneic BM transplantation in sublethally-irradiated mice . By contrast, the tol DC do not restrain the ability of CD8+ T cells to mediate a graft-versus-leukemia effect, indicating the immunomodulatory capacity of alternatively-activated mDC to selectively target T cell responses in vivo.
Since Ag presentation via the indirect pathway is thought to play a major role in the development of chronic rejection, activation of donor-specific T cells through the indirect pathway is considered a principal target for innovative tolerogenic therapies in organ transplantation. In a recent study , Ab-mediated targeting of recipient mDC in situ with intact MHC class I monomers inhibited the development of indirect alloresponses and the generation of IgG alloAbs by interfering with T cell-dependent B cell activation, resulting in MHC-mismatched skin allograft survival. This was achieved in combination with temporary inhibition of the direct alloimmune response using anti-CD8-depleting Ab.
It has been reported that host-derived pDC are integral to the development of vascularized organ allograft tolerance induced by DST and anti-CD154 mAb . Thus, PDCA-1+ pDC acquired donor MHC II-derived allopeptide from the graft and migrated to LN (but not spleen), where they induced allo-specific CD4+Foxp3+ Treg. In this model, pDC depletion or prevention of pDC LN homing inhibited Treg development and tolerance induction. However, in a rat model of cardiac allograft tolerance induced by CD40Ig, pDC accumulated in the graft and the spleen, but not LN, and induced CD8+ Treg that suppressed CD4+ effector cells . pDC regulation of alloreactive CD4+ T cells was either direct, through an IDO-dependent mechanism, or indirect through CD8+ Treg, in a contact-dependent manner. Recently, it has been shown that complete absence of pDC within LN of CCR7−/− mice prevents the successful induction of tolerance to cardiac allografts (with DST and anti-CD154 mAb), whereas adoptive transfer of syngeneic pDC is sufficient to rescue graft survival in CCR7−/− recipients in a dose-dependent manner .
Tol DC bearing both self and donor MHC molecules have been generated from BM precursors of semi-allogeneic (donor × recipient) F1 rodents or by the fusion of DC of different MHC haplotypes. Negative vaccination of (Lewis) (LEW) rats with dexamethasone-treated, semi-allogeneic BM-derived immature DC (from LEW × AUG [August] F1 rats) led to the indefinite survival of AUG-derived kidney grafts in LEW rats. No transplant vasculopathy was detected when this treatment was combined with CTLA4–Ig and a short, postoperative course of calcineurin inhibition (cyclosporine A; CsA) . The data suggest that, in this model, pre-treatment of graft recipients with semi-allogeneic DC leads to indirect pathway-mediated regulation by Treg, the activity of which is dependent on the presence of the donor MHC molecules. In mice, injection of F1 (BALB/c × C57BL/6) DC expressing transgenic viral IL-10 and CCR7 (to enhance homing capacity to secondary lymphoid tissue) 7 days before transplant, results in indefinite survival of F1 heart grafts in >80% of C57BL/6 recipients .
Donor–recipient DC–DC hybrids have been created in vitro by fusing a CD95L (FasL)-transduced DC line with allogeneic splenic DC, with the goal of inducing selective apoptosis of donor-reactive T cells that recognize alloAg though each possible pathway of allorecognition . These CD95L-transduced DC–DC hybrids suppressed allospecific DTH responses against both parental strains, but did not affect responses to third party Ag. Repetitive intravenous administration resulted in delayed onset of GVHD in mice, due to MHC-restricted, CD95L-mediated apoptosis of donor anti-host CTL.
In humans, repetitive stimulation with monocyte-derived immature mDC pulsed with an HLA-A2-derived allopeptide has been used to expand alloAg-specific Treg ex vivo . mDC can also down-regulate CD8+ T cell responses. Proof of principle that human immature DC can exhibit tolerogenic properties in vivo has been provided by studies of human volunteer responses to model Ags. Thus, in humans, subcutaneous injection of autologous, monocyte-derived immature DC pulsed with an influenza matrix peptide induced Ag-specific regulatory CD8+ T cells secreting IL-10, and memory CD8+ T cells with impaired IFN-γ secretion and cytolytic function [21, 22].
Analysis of peripheral blood DC in organ or BM transplant patients may be a helpful immune monitoring tool [103, 104]; reviewed in ref . Higher incidences of circulating pDC relative to mDC are found in operationally tolerant pediatric liver allograft recipients and in patients on low dose immunosuppressive drug therapy undergoing prospective drug weaning, compared with patients on maintenance immunosuppression. These differences cannot be ascribed to the dose of immunosuppression or to the whole blood level of calcineurin inhibitor. In addition, high PD-L1/CD86 ratios on pDC correlate with elevated CD4+CD25hiFoxp3+ Treg in transplant patients who exhibit a state of operational tolerance , which is consistent with evidence that the balance between expression of inhibitory PD-L1 and costimulatory CD80/CD86 ligands regulates the outcome of their interaction with T cells . Patients treated with B7-CD28 costimulatory pathway blocking agent Lea29Y (Belatacept) display significantly higher soluble HLA-G (sHLA-G) plasma concentrations than patients treated with calcineurin inhibitors or healthy donors. In these investigations, DC were identified as one of the cellular sources of sHLA-G in the CTLA4-Ig-treated patients . As indicated above, interaction of HLA-G with its receptor ILT-4 on DC down-regulates their T-cell stimulatory capacity. In stable liver transplant patients off all immunosuppression, circulating mDC (but not pDC) express higher cell surface HLA-G, which correlates significantly with that of Foxp3 on Treg .
The large amount of information generated from many pre-clinical rodent studies and in vitro work with human DC provides the basis for the development of a protocol for utilizing therapeutic DC in a clinical setting with special relevance to organ transplantation. It is clear that human DC can regulate immune reactivity by a variety of mechanisms, e.g. induction and expansion of Treg (such as CD4+CD25+ T cells or Tr1 cells that make IL-10), anergy and/or deletion. Initial studies have demonstrated the feasibility of pharmacologic or genetic approaches to enhance and maintain the tolerogenicity of either ex vivo-generated donor or recipient DC. In transplantation, one of the potential concerns relating to the administration of tol DC is that, in the context of endogenous and exogenous danger signals, and in the presence of pro-inflammatory cytokines, the administered DC may mature and lose their immunomodulatory potential, or even accelerate graft rejection. This would be minimized/avoided if DC were to be administered sufficiently in advance of transplantation (an approach adopted in most experimental small animal models). Such an approach is possible in live donor transplantation. In the deceased donor setting, adequate control of DC maturation would be essential. Notably, in rodents, maturation-resistant host DC administered post-transplant under cover of low dose immunosuppression can promote indefinite allograft survival , demonstrating their efficacy under these conditions.
Multiple pharmacologic and biologic agents, including anti-inflammatory and immunosuppressive drugs, as well as irradiation techniques have been evaluated for their potential to induce human tolerogenic DC (Table 1). These studies have evaluated phenotypic characteristics, cytokine production and T cell immunostimulatory capacity of the manipulated DC. The DC were either tested directly in vitro with or without the agent, or obtained from transplant patients already receiving the agent [110–115], in which case the DC were differentiated from isolated cells (blood monocytes) and evaluated for maturation and functional capacity.
Extracorporeal photopheresis (ECP) has been shown to be an effective treatment for, solid organ graft rejection, GVHD and other T cell-mediated diseases. ECP is an immunomodulatory therapy in which leukocytes are exposed to 8-methoxypsoralen (8-MOP) and ultraviolet A radiation (PUVA). PUVA-treated or –untreated, highly-purified CD14+ cells (monocytes) were incubated with immunosuppressive drugs in the presence of IL-4 and GM-CSF. Neither immunosuppressive drugs at the lowest clinical concentration, nor their combination with PUVA, affected myeloid DC generation; however, a tolerogenic phenotype was induced. This approach may be useful for the generation of immature tol DC in vitro.
A number of logistic issues need to be evaluated prior to implementation of therapeutic human tol DC (Figure 3). For example, cell isolation and culture techniques for ex vivo preparation of tol DC can vary from one center to another, which can result in significant variation in cell yield. Use of Ficoll isolation and cryopreservation can result in a higher pDC/mDC ratio than with fresh cells . Also, it has been shown that initial culture of human DC can be associated with variability in surface molecule expression as well as cytokine production . Additionally, materials used for clinical scale production of human DC in a “closed-system” have been shown to affect cytokine production by DC, e.g. IL-10 and IL-12, which might affect their therapeutic potential in vivo after administration . This warrants technical standardization of DC preparation for clinical use.
In small animal models, BM or blood monocytes have been used successfully to generate tol DC that significantly prolong allograft survival. In humans, blood monocytes are the most common source for tol DC, in addition to other sources, e.g. cord blood or leukopheresis products. When recipient tol DC are considered, alloAg presentation by these cells requires targeting the DC with donor Ag. In rodents, several methods have been described for pulsing tol DC with donor Ag with successful therapeutic outcomes. These include cell-free lysate, MHC peptides, early apoptotic cells, or exosomes. The optimal approach for human in vivo use has not been determined.
The optimal route of administration of tol DC can be a critical step in tolerance induction, taking into consideration the desired migration of the injected tolDC to T cell areas in secondary lymphoid tissues. In a clinical trial of DC for cancer therapy, it was verified that radiolabeled DC injected intradermally had a higher migration rate to lymph nodes than DC injected subcutaneously. Also, mature DC showed greater migratory activity than immature DC . In these studies, DC were detected in lymph nodes 20–60 min after injection, and the maximum concentration was reached after 48–72 h. Although the purpose of this study was to induce immunostimulatory (rather than tolerogenic) effects, the observations highlight the possible effect of phenotypic maturation on the migratory capacity of the injected DC. Additional issues regarding DC administration are timing, dose and frequency. In a non-human primate model, a single intravenous infusion of donor-derived tol DC (VitaminD3/IL-10-treated to induce maturation resistance) together with a single CTLA4-Ig dose, resulted in modulation of allogeneic T cell responses and induction of hyporesponsiveness to donor and third party alloAgs .
Optimum immunosuppressive agents for combination therapy with tol DC need to be determined (Table 1). The immunosuppressive regimen administered to patients following transplantation may critically affect the therapeutic potential of tol DC. For example, a 4-day course of corticosteroids in healthy humans induces a fall in pDC numbers that begin to recover immediately after steroid withdrawal . Mammalian target of rapamycin inhibitors such as rapamycin and calcineurin inhibitors can have differential effects on DC subsets in the circulation after transplantation . T cell-depleting agents can also deplete DC. Thus Campath-1H (anti-CD52) mAb caused a strong, sustained reduction in the total number of peripheral DC, with a significant shift in the pDC/mDC ratio towards pDC as early as one month following kidney transplantation . Other issues include stable DC immaturity/tolerogenicity/specificity, ensuring minimal risk of host sensitization; outcomes that can be monitored adequately in the laboratory and clinic; overcoming heterologous immunity/memory/late graft failure; regulatory issues: safety (GMP compliance); standardization/quality control; commercialization of therapeutic product.
As yet, there are no reports of tol DC therapy in NHP organ transplantation. While small animal models provide important insight into mechanisms underlying tolerance induction, translation into the clinical setting has proven to be much more difficult, possibly due to differences between rodent and human leukocyte biology, and the barrier of heterologous immunity/T cell memory that is more prevalent in humans. Pre-clinical studies in NHP can provide a great amount of information that is needed to establish tol DC in relation to cell and organ transplantation. NHP models (especially rhesus macaques) have proven valuable tools in immunological studies utilizing DC therapy to mount an immune response to control infection and cancer, i.e. as a vaccine therapy. From a different perspective, tol DC can potentially be used to ameliorate the adaptive immune response to an allograft, i.e. negative vaccination. NHP tol DC have been characterized where vitD3/IL-10 treated (maturation-resistant) tol DC were propagated and infused safely in rhesus macaques. . Systemic administration of these cells, in conjunction with CTLA4Ig, was well-tolerated and associated with significant hyporesponsiveness to donor allo-Ags . Additionally, co-culturing these cells with CD4+CD127−/lo T cells resulted in the generation of anergic T cells with allo-Ag specific suppressive capacity in vitro . In a recent study, after culturing BM-derived rhesus macaque DC in vitro, adherent (but not non-adherent DC) were found to be poor stimulators of T cells and inhibited T-cell proliferation, i.e. tol DC, via a heme oxygenase-1 (HO-1)-dependent mechanism . Further studies of tol DC in the NHP model are likely to provide important insights into strategies for potential clinical application (Figure 3).
The authors' work is supported by National Institutes of Health grants R01 AI67541, U01 AI51698; U01 AI91197 and P01 AI81678.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.