Regulatory Myeloid Cells in Transplantation

Abstract

Regulatory myeloid cells (RMC) are emerging as novel targets for immunosuppressive (IS) agents and hold considerable promise as cellular therapeutic agents. Herein, we discuss the ability of regulatory macrophages, regulatory dendritic cells, and myeloid-derived suppressor cells to regulate alloimmunity, their potential as cellular therapeutic agents, and the IS agents that target their function. We consider protocols for the generation of RMC and the selection of donor- or recipient-derived cells for adoptive cell therapy. Additionally, the issues of cell trafficking and antigen (Ag) specificity after RMC transfer are discussed. Improved understanding of the immunobiology of these cells has increased the possibility of moving RMC into the clinic to reduce the burden of current IS agents and to promote Ag-specific tolerance. In the second half of this review, we discuss the influence of established and experimental IS agents on myeloid cell populations. IS agents believed historically to act primarily on T cell activation and proliferation are emerging as important regulators of RMC function. Better insights into the influence of IS agents on RMC will enhance our ability to develop cell therapy protocols to promote the function of these cells. Moreover, novel IS agents may be designed to target RMC in situ to promote Ag-specific immune regulation in transplantation and to usher in a new era of immune modulation exploiting cells of myeloid origin.

Despite excellent short-term outcomes due to the prevention and successful treatment of acute rejection, late graft failure remains an important problem in organ transplantation (¹). Moreover, current nonspecific suppression of the immune system using antirejection drugs carries significant risks, including infection, malignancy, and drug toxicity (²). Currently, there is increasing interest in the potential of regulatory innate or adaptive immune cells to control allograft rejection (³). Targeting myeloid cells with the goal of minimizing dependency on immunosuppressive (IS) drugs and promoting donor-specific tolerance represents a promising approach.

Herein, we discuss strategies to target regulatory myeloid cells (RMC) in situ and prospects for cell therapy in transplantation using RMC. Three RMC populations—regulatory macrophages (Mreg), regulatory dendritic cells (DCreg), and myeloid-derived suppressor cells (MDSC)—will be the focus of this review. Mreg will be discussed in the context of studies on peripheral blood mononuclear cell (PBMC)–derived cells differentiated in macrophage colony-stimulating factor (M-CSF) and then stimulated with interferon (IFN)-γ, because most work on Mreg in the field of transplantation has been focused on this population (^{4, 5}). DCs are innate professional antigen (Ag)–presenting cells (APC) that serve as critical initiators and regulators of innate and adaptive immunity (^6–8). For in-depth analysis of DC ontogeny and the mechanisms that underlie their immune regulatory capacity, see recent comprehensive reviews (^8–12). MDSC are a heterogeneous population of immature myeloid cells and myeloid progenitors that regulate antitumor immunity and share the ability to suppress effector T cell responses. The origin and suppressive mechanisms of MDSC have been reviewed in detail (^{13, 14}).

RMC AS CELLULAR IMMUNOTHERAPEUTIC AGENTS

In vitro Generation of RMC

RMC generated in vitro for therapeutic evaluation are propagated typically from rodent bone marrow (BM) cells or human PBMC (Fig. 1). Although differentiation procedures between species are similar, distinct starting cell populations make the translation of findings from rodents to humans difficult (¹⁵). Moreover, RMC therapy lacks standard differentiation protocols because the optimal immune regulatory properties of each RMC population are unknown (¹⁶). Although MDSC have not been evaluated for immune regulatory function in humans, protocols for the propagation and administration of Mreg and DCreg have been described in human renal transplantation and in healthy volunteers or type 1 diabetics, respectively (Table 1). Importantly, no adverse effects of RMC therapy have been reported in these limited clinical studies to date.

TABLE 1: Influence of RMC administration in humans

FIGURE 1: Generation of RMC in vitro from rodent BM cells or human PBMC. Mreg, DCreg, and MDSC can be generated in vitro from precursors in rodent BM or human PBMC exposed to specific growth factors. In some cases, RMC (Mreg and MDSC) are also activated in vitro by the addition of inflammatory cytokines or other soluble factors. DCreg are often generated in the presence of anti-inflammatory cytokines or agents that suppress their activation into stimulatory DC.

Human Mreg are differentiated from donor PBMC acquired by leukapheresis, in recombinant human M-CSF for 6 days followed by 24 hr stimulation with IFN-γ (²³). Human DCreg are typically differentiated from PBMC or purified monocytes in the presence of granulocyte-M-CSF (GM-CSF) and interleukin (IL)-4, with the addition of one or more factors that promote their tolerogenicity (reviewed in (^{11, 24})). DCreg are typically immature myeloid DC (mDC) and maturation resistant or “alternatively activated” (e.g., exposed to IL-10 and transforming growth factor-β during propagation then stimulated with lipopolysaccharide [LPS]), so that they maintain expression of major histocompatibility complex (MHC) molecules but display low levels of costimulatory molecules and proinflammatory cytokines. Vitamin D3 (vitD3) and dexamethasone promote DCreg (^{25, 26}). Thus, activation of human DC cultured in vitD3/dexamethasone with LPS results in stable, “alternatively activated,” semimature DC (²⁷). Addition of IL-10 (“DC-10”) (²⁸) or the mechanistic/mammalian target of rapamycin (mTOR) inhibitor, rapamycin (RAPA) (²⁹), to human monocyte cultures also produces DCreg. Non human primate (NHP) monocyte-derived DCreg can be generated using vitD3 and IL-10 (^{30, 31}). DCreg are also made using low-dose GM-CSF in the absence of IL-4 (³²). Thus, although Mreg differentiation is comparatively well defined, there is significant variability in methods to generate DCreg. Importantly, generation of recipient-derived RMC for clinical use must be validated with PBMC from patients with preexisting disease (³³). In this regard, DCreg generated from patients with rheumatoid arthritis (³⁴) or relapsing-remitting multiple sclerosis (³⁵) exhibit a similar phenotype and function to DCreg generated from healthy controls.

MDSC exhibit considerable phenotypic heterogeneity and are subdivided into those that resemble monocytes or are similar phenotypically to neutrophils (³⁶). They require factors to induce their activation in addition to their expansion (¹³). Thus, mouse monocytic MDSC are generated from BM cells in G-CSF, GM-CSF, or both and activated with IL-6 or IL-13 (^{37, 38}). Table 2 outlines adoptive MDSC therapies that have been evaluated in mouse models of skin or pancreatic islet cell transplantation, graft-versus-host disease (GVHD), and type 1 diabetes. Human MDSC generated from PBMC with GM-CSF+IL-6 appear to exert the most potent suppressive capacity, but GM-CSF+IL-1β, prostaglandin E2 (PGE2), tumor necrosis factor-α, or vascular endothelial growth factor also induce suppressive MDSC (⁴⁵). Similarly, GM-CSF and IL-6 can be used to generate suppressive human BM-derived MDSC (³⁷). Addition of PGE2 to GM-CSF and IL-4–stimulated human PBMC cultures blocks DC differentiation and promotes MDSC generation (⁴⁶).

TABLE 2: MDSC transfer for immune modulation or cell therapy of allograft rejection

RMC therapies need to be designed in conjunction with current IS protocols due to the success of the latter in achieving high short-term organ allograft survival rates (¹⁵). Thus, experimental RMC therapy needs to be undertaken with an appropriate IS agent(s) that maintains their tolerogenic properties. In rodent organ transplant models, Mreg (⁴⁷) and DCreg (¹¹) synergize with pharmacologic agents, antilymphocyte serum, or costimulation blockade, but the impact of IS agents on MDSC is largely unknown.

Selection of Donor or Recipient RMC for Therapy

Mouse Mreg prolong allograft survival only when donor derived (⁴⁷). Although there is a potential risk of sensitizing the recipient to donor, this has not been observed in the human renal transplant recipients given Mreg to date (^{48, 22}). The risk is mitigated by infusing the cells 1 week before transplantation (to avoid surgically induced inflammation) and choosing IS agents that are likely to maintain the tolerogenic properties of RMC in the face of inflammation.

Both donor- and recipient-derived DCreg have been investigated extensively in rodent transplant models (^{11, 15, 24, 49}). Although allogeneic DC trafficking from rodent organ grafts may survive in lymphoid tissue for several days in unmanipulated hosts or even weeks in immunosuppressed recipients (^{50, 51}), these donor DC may be also killed by host natural killer cells (⁵²) and reprocessed by endogenous DC able to present donor alloAg (⁵³). Donor Mreg survive in humans for at least 30 hr in the spleen, liver, and BM (²³) and 2 weeks in mice in the lung (⁴⁷). Although DCreg can be generated from the graft recipient at any time, the optimal method of loading donor alloAg (donor cell lysate, exosomes, and apoptotic cells) has not been established (²⁴). One group has used unpulsed autologous DCreg to promote long-term rodent allograft survival; thus, maturation-resistant DCreg are given in the peritransplantation period and acquire donor alloAg in situ (^54–56).

Similar events could accompany cell therapy with MDSC, because these cells are also able to process and present Ag (^{57, 58}). As precursors of myeloid cells, MDSC can differentiate into DC and macrophages (^{37, 39, 59, 60}), but MDSC have not been found to potentiate immunity after their adoptive transfer (Table 2) and retain immune regulatory function, even if they do differentiate (^{37, 39}). On the contrary, cyclooxygenase-2 activation by inflammatory mediators, such as IL-1β and IFN-γ, prevents the differentiation of MDSC into DC (⁶¹), whereas IFN-γ is an important stimulator of MDSC suppressive function (⁶²). These properties resemble those of Mreg that are activated by IFN-γ (⁴⁸) and provide the advantage that inflammatory conditions such as occur in organ transplantation may reinforce the suppressive activity of MDSC. Thus, selection of donor or recipient RMC presents its own distinct challenges, such as circumventing allosensitization, and the need for/nature of Ag pulsing.

Ag Specificity

The ability of RMC to regulate immune responses in an Ag-specific manner is an important consideration to avoid global immunosuppression. Mouse (⁴⁷) and human (²³) Mreg suppress mitogen-activated CD4⁺ and CD8⁺ T cell proliferation, and mouse Mreg delete alloreactive T cells specifically in vitro (⁴⁷). Moreover, donor-, but not recipient-, or third party-derived Mreg prolong mouse cardiac allograft survival (⁴⁷), suggesting that Mreg can regulate alloAg-specific immunity in vivo. Administration of transplant acceptance-inducing cells (TAIC), that is, unpurified Mreg, to human renal transplant recipients has been reported to promote donor-specific hyporesponsiveness even in a presensitized recipient (^{21, 22}).

Donor- and host-derived DCreg promote long-term allograft survival or donor-specific tolerance in rodent transplant models when combined with antilymphocyte serum, anti-CD40L (CD154) monoclonal antibody (mAb), or cytotoxic T lymphocyte Ag 4 (CTLA4)-Ig (^63–67). Importantly, local administration of immature autologous DC to healthy human volunteers results in inhibition of Ag-specific CD8⁺ T cell effector function (¹⁷) and generation of regulatory CD8⁺ T cells (¹⁸). These latter findings provide proof-of-principle that DC have the capacity to regulate Ag-specific responses in humans. Recently, donor-derived DCreg have been shown to prolong organ allograft survival in a robust preclinical NHP renal transplant model accompanied by reduction in donor-reactive T memory cell responses (³¹).

The Ag specificity of MDSC suppressive function depends on the model, microenvironment, and activation of target lymphocytes (⁶⁸). MDSC can inhibit both CD4⁺ and CD8⁺ T cell reactivity (^{40, 57, 62, 69, 70}). They can suppress Ag-specific CD8⁺ T cell responses (⁵⁷), but it is not known whether they are capable of Ag-specific CD4⁺ T cell suppression (¹³), especially in view of their low or absent MHC class II expression (⁷¹). Importantly, MDSC generated in vitro can promote Ag-specific CD8⁺ T cell hyporesponsiveness (³⁷). In a mouse model of cardiac allograft tolerance induced by donor-specific transfusion and anti-CD40L mAb, suppression of T cells by graft-infiltrating MDSC was nonspecific, and BM and splenic monocytes did not suppress (⁷²). Taken together, DCreg and Mreg have Ag-specific regulatory capacity in transplantation, but the conditions under which MDSC suppress alloimmunity in an Ag-specific manner need to be better understood to harness these cells for therapeutic application.

Trafficking and Migration of RMC Under Inflammatory Conditions and After Their Adoptive Transfer

There is evidence that human Mreg administered via central venous access migrate to the lungs and then distribute to the liver, spleen, and BM within 30 hr of their infusion (²³). Murine Mreg demonstrate a similar distribution pattern after intravenous (i.v.) injection and notably do not migrate to lymph nodes (⁴⁷). Little is known about chemokine receptor expression on Mreg and the location(s) where they exert their regulatory function in vivo is not known (⁴).

Expression of CCR7 by DC directs them to secondary lymphoid organs where they interact with T cells. Adoptively transferred, IL-10–expressing DC require CCR7 to prolong mouse cardiac allograft survival (⁷³), suggesting that DCreg, and likely Mreg, must traffic to secondary lymphoid for their regulatory function. Notably, IL-10 reduces DC CCR7 expression and lymph node homing ability (⁷⁴). Up-regulation of CCR7 after activation of DCreg by Toll-like receptor (TLR) ligation in vitro may be required to improve the migratory function of these cells (⁷⁵). After i.v. injection, rodent host-derived DCreg migrate rapidly to the spleen (^{76, 77}), whereas RAPA-conditioned DC migrate to the lymph nodes after intramuscular injection (⁷⁸). The route of DCreg administration may be critical to optimize their function in vivo (⁷⁵). Whereas i.v. DCreg injection prolongs cardiac allograft survival in mice, subcutaneous (s.c.) injection of the same DCreg does not affect graft survival (⁷⁹). Similarly, in a NHP model, i.v. administration of DCreg results in immune regulation (³⁰), whereas intradermal injection may boost the immune response (¹⁵). In human cancer patients, intradermal injection increases the migration of immature DC to the draining lymph nodes compared with s.c. administration (⁸⁰); however, s.c. administration of immature DC has been shown to regulate CD8⁺ T cell responses to model Ags in humans (^{17, 18}). Together, these studies suggest that optimization of delivery route is critical to DCreg function and that directing their migration to secondary lymphoid organs is important.

MDSC express chemokine receptors, such as CX3CR1 (⁴¹) or CCR2 (^{57, 81}), which direct them toward sites of inflammation, but they can also be directed toward secondary lymphoid organs by expression of CD62L (^{38, 57}) and CCR7 (³⁸). It is unknown whether MDSC migration to the allograft, secondary lymphoid organs, or both is preferable after their adoptive transfer; however, MDSC are required to migrate to the graft and not lymph nodes for experimental transplant tolerance induced by donor-specific infusion and anti-CD154 mAb (⁷²). The complement component C5a participates in the recruitment of MDSC to tumors and peripheral lymphoid organs in mice (⁸²). Thus, it will be of interest to determine whether C5a plays a similar role in transplant rejection, because C5 is integral to Ab-mediated rejection (⁸³). In vitro–generated MDSC traffic to peripheral lymphoid tissue and sites of inflammation in GVHD, including the liver and spleen (⁴⁴) or spleen and lymph nodes (³⁸). MDSC expanded in vivo in response to LPS that inhibited alloimmunity migrated to the spleen when transferred to skin transplant recipients, but their migration to the graft was not assessed (⁴²). MDSC accumulate within tumors (³⁹) and at sites of inflammation in murine experimental allergic encephalomyelitis (⁶²) and chronic contact eczema (⁷⁰). They also accumulate within the spleen (^{39, 70, 84}) and lymph nodes (^{39, 70}) in inflammatory disease and cancer. After transplantation, rodent MDSC are found in the allograft and peripheral blood (^{72, 85, 86}) as the result of migration from the BM (⁷²). Although human MDSC were reported to be elevated in the peripheral blood of renal transplant recipients, they were not assessed in biopsy tissue (⁸⁷).

In summary, RMC therapies have demonstrated promising immune regulatory capacity. However, it will be necessary to rationally design protocols in transplantation that optimize in vitro generation of RMC whose in vivo migration (to the appropriate sites) and function are supported by the IS regimen. Further preclinical studies are warranted to optimize each parameter in increasingly stringent models from rodent to NHP while also continuing to progress RMC therapy in human transplant recipients.

TARGETING RMC WITH THERAPEUTIC AGENTS

This section summarizes reports concerning the influence of IS drugs, specific therapeutic Abs, and novel immunoregulatory strategies on DC, macrophages, and MDSC (Table 3).

TABLE 3: Influence of IS drugs, biological agents, and novel immunoregulatory agents on mDC, macrophages, and MDSC in vivo

Conventional IS Drugs

Transplant recipients receive pharmacologic and biological agents to control graft rejection, and although the principal mechanism of action of these agents is inhibition of T cell responses, they also modulate RMC. The influence of anti-inflammatory agents, IS drugs, and biological IS on DC function in vivo has been reviewed in detail elsewhere (^{10, 123, 124}). Studies of their influence on Mreg and MDSC are limited.

The most extensively studied IS drugs that target DC in vivo are glucocorticoids (GC), calcineurin inhibitors (CNI), RAPA (sirolimus), and mycophenolate mofetil (MMF) (¹⁰). The in vivo effects of GC on DC have been reviewed by Kooten et al. (¹²⁵). Specifically, GC reduce peripheral DC numbers and inhibit their maturation and production of proinflammatory cytokines while enriching for Mreg (^{88, 126}). Endogenous GC promote the expansion of MDSC in a murine model of trauma (⁸⁹), and exposure of monocytes to GC induces CD11b⁺Gr-1⁺CD124⁺Ly6C^med MDSC (¹²⁷). Administration of dexamethasone to glioblastoma patients increases circulating CD14⁺HLA-DR^lo/negCD80^- IS cells that resemble MDSC (¹²⁸).

CNI (i.e., cyclosporine A [CsA] and tacrolimus [FK506]) are frontline antirejection agents used in combination with an antiproliferative agent, particularly MMF. CsA and tacrolimus, but not RAPA, inhibit MHC-restricted Ag presentation by DC in vitro (¹²⁹) and in vivo (¹⁰²). Tacrolimus treatment of mice reduces responsiveness of macrophages and DC to LPS (¹⁰³). Numbers of thymic DC and macrophages are decreased in rats during CsA treatment (^96–98); however, their function appears to be unaffected (⁹⁶). On the contrary, increased numbers of DC have been reported in NHP with long-surviving renal allografts treated with both tacrolimus and sirolimus (¹³⁰). CsA combined with CCR5 blockade increases cardiac graft survival in NHP, an effect that is associated with generation of alternatively activated macrophages through activation of the peroxisome proliferator-activated receptor-γ (⁹⁹). Additionally, CsA inhibits the phenotypic maturation, endocytic activity, and allostimulatory function of human peripheral blood DC (¹³¹). CsA or tacrolimus increases the incidence of mDC in peripheral blood of human heart transplant recipients, but no difference in expression of the DC maturation marker CD83 is observed (¹⁰⁰). To our knowledge, direct effects of CNI on MDSC have not been studied; however, expression of the immunophilin FK506 binding protein 51 is increased in monocytic and granulocytic MDSC isolated from tumor-bearing mice and regulates their suppressive function (¹³²). Additionally, calcineurin and nuclear factor of activated T cells signaling are negative regulators of myelopoiesis, and CsA augments numbers of differentiated DC in vitro (¹³³). Therefore, it appears likely that CNI impact MDSC.

MMF is an antiproliferative prodrug of mycophenolic acid (MPA) that inhibits B- and T cell proliferation (¹³⁴). MPA also suppresses DC maturation and reduces Ag presentation to T lymphocytes (^{109, 135–137}). As MPA has been reported to suppress granulopoiesis, it is possible that it also affects MDSC.

RAPA inhibits the serine/threonine kinase mechanistic target of RAPA (mTOR) (¹³⁸). Its administration to mice impairs DC costimulatory molecule up-regulation, production of proinflammatory cytokines, and T cell allostimulatory function (^{106, 139, 140}). Moreover, RAPA induces apoptosis in DC but not in monocytes or macrophages (¹⁰⁷). Haidinger et al. (¹⁰⁸) found that DC in kidney transplant patients treated with RAPA displayed increased immunostimulatory potential compared with those in patients treated with CNI and in healthy controls. Interestingly, RAPA prevents the anti-inflammatory effects of GC on human monocytes as well as mDC (¹⁴¹). Moreover, RAPA conditioning augments IL-12 production by mouse BM-derived DC or human monocyte-derived DC stimulated with LPS or proinflammatory cytokines, respectively (^{142, 143}). Thus, under different circumstances, RAPA can exert proinflammatory or anti-inflammatory effects on DC. mTOR is required for DC development, so it will be interesting to determine whether RAPA affects MDSC due to its ability to inhibit myelopoiesis (¹⁴⁴).

Thus, in addition to the ability of conventional IS agents to inhibit B- and T cell activation, these drugs exert profound, but variable, effects on macrophage and DC differentiation and function.

Experimental IS Agents

Histone deacetylase (HDAC) inhibitors (including suberoylanilide hydroxamic acid, trichostatin A, and valproic acid) are antitumor agents that also have anti-inflammatory properties. HDAC inhibitors reduce TLR-induced costimulatory molecule expression and proinflammatory cytokine release by DC and their T cell allostimulatory activity in vitro and in vivo (^{111, 112, 145}). HDAC inhibition blocks GM-CSF–dependent function in macrophages and their differentiation to DC (¹⁴⁶), but there are contradictory reports regarding its influence on cytokine secretion (^{147, 148}), which may reflect the specific HDAC inhibitor or dose used. We have demonstrated recently (¹⁴⁹) that HDAC inhibitors augment GM-CSF–mediated murine MDSC expansion in vitro and in vivo and that these MDSC exhibit similar suppressive potency to control MDSC.

Proteasome inhibitors, such as bortezomib, are believed to block the activation and nuclear translocation of nuclear factor (NF)-κB, a transcription factor central to DC maturation and inflammatory responses (¹⁵⁰). In experimental hematopoietic stem cell transplantation, bortezomib attenuates GVHD yet preserves graft-versus-leukemia activity (^{151, 152}). Administration of bortezomib to mice results in a more immature DC phenotype (¹¹³). Bortezomib reduces the phagocytic capacity of human monocyte-derived DC, skews their phenotypic maturation, and reduces their cytokine production and immunostimulatory capacity. It also reduces their chemokine secretion and migration (¹⁵³) while promoting their apoptosis and reducing the yield of viable DC (¹⁵³), preferentially targeting immature DC (¹⁵⁰).

There are also anti-inflammatory agents that modulate RMC function. NF-κB inhibitors block DC maturation and can induce tolerance in murine cardiac transplantation (^153–156). Interestingly, NF-κB is implicated as a critical regulator of MDSC suppressive function (¹⁵⁷). Furthermore, cyclooxygenase-2 inhibitors prevent production of PGE2 and reduce numbers of MDSC (¹⁵⁸) and can prolong murine cardiac allograft survival (¹⁵⁹). There is also evidence that a PGE2 receptor (EP4) agonist suppresses the activation of macrophages and prolongs mouse cardiac allograft survival (¹¹⁶).

Thus, various experimental IS agents currently under investigation are capable of modifying RMC function. Typically, they reduce DC maturation but appear to have varying effects on MDSC expansion and function.

In Vivo RMC Targeting with Abs and Other Novel Approaches

T cell–depleting Abs also target RMC. Thus, polyclonal antithymocyte globulin (ATG) inhibits human DC Ag uptake and maturation, induces complement-mediated lysis of DC, and decreases the capacity of DC to stimulate allogeneic T cells in vitro (¹⁶⁰). Additionally, ATG polarizes DC toward expression of indoleamine dioxygenase (¹⁶¹) that inhibits T cell proliferation. Anti-CD52 mAb (alemtuzumab; Campath-1H) depletes peripheral blood DC, but not tissue DC, due to differential expression of CD52 on DC in these sites (^{162, 163}). It causes a sustained reduction of total peripheral DC in kidney transplant recipients (¹¹⁸). In addition to T cells, human DC express CD25 after stimulation (^{164, 165}), making them a potential target for anti-CD25 (IL-2 receptor α subunit) mAb. Furthermore, anti-CD25 mAb treatment diminishes the ability of human DC to stimulate T helper cells (¹⁶⁴) but does not affect HLA-DR or costimulatory molecule expression by the DC after LPS stimulation (¹⁶⁵). Recent work using daclizumab (humanized anti-CD25 mAb) has shown that it potently inhibits Ag-specific T cell activation by human mature DC in vitro (¹⁶⁶). Interestingly, anti-CD25 mAb combined with IL-12 depletes MDSC in a mouse model of colon carcinoma (¹⁶⁷).

Costimulation blockade is an emerging strategy to promote graft survival by interfering with T cell activation, in which APC play an important role. Development of costimulation blockers has focused mainly on targeting T cell surface costimulatory molecules, although some also target APC (^{168, 169}). Notably, anti-CD28 mAb induces tolerance to rat kidney allografts in association with accumulation of circulating and graft-infiltrating MDSC that suppress effector T cell expansion (⁸⁵). Belatacept (CTLA4-Ig) is the first costimulation blocker approved for renal transplantation. There is evidence that CTLA4-Ig binding to CD80/CD86 molecules provides a reverse signal to DC that results in the induction of indoleamine dioxygenase (¹⁷⁰) and that enhanced secretion of inhibitory products by CTLA4-Ig–exposed DC promotes alloantigen-specific transplant tolerance (¹¹⁹). However, it has been reported recently that CTLA4-Ig IS activity may not depend on a DCreg phenotype but on its presence during DC/T cell interaction (¹⁷¹). Interestingly, Ab blockade of CTLA4 reduces the suppressive potential of MDSC in tumor-bearing mice (¹⁷²). Anti-CD40 mAbs prolong renal and islet allograft survival in NHP (^{173, 174}), whereas mouse mDC under CD40 blockade have a tolerogenic profile in vivo (¹²⁰) and are responsible for inducing peripheral Treg and delaying cardiac allograft rejection (¹⁷⁵).

Gene silencing of TLR adaptors, namely myeloid differentiation primary response gene 88 and TIR domain-containing adapter-inducing IFN-β, using small interfering RNA (siRNA) reduces DC maturation and prolongs murine cardiac allograft survival (¹²¹). Administration of recombinant G-CSF (Neupogen) prolongs skin transplant survival in mice and induces MDSC in peripheral lymphoid compartments (¹²²). Suppressive granulocytic and monocytic MDSC are expanded in human stem cell donors during G-CSF mobilization protocols for allogeneic hematopoietic stem cell transplantation (¹⁷⁶). Furthermore, human inhibitory receptor ILT2, expressed on activated T cells and engaged by HLA-G on DC, has been shown to amplify MDSC and to promote long-term allograft survival (⁴¹).

Thus, although previously thought to act primarily on T cells, T cell–depleting inhibitory Abs also profoundly affect DC function, and novel approaches using costimulation blockade, siRNA, or recombinant growth factors can promote MDSC.

CONCLUSION

RMC constitute an important, heterogeneous innate immune cell population with considerable promise for cell therapy. The influence of IS agents on these cells is becoming increasingly apparent. Although the use of RMC as cellular therapeutics is beginning to advance from preclinical models to patients with inflammatory diseases, further insights into the differentiation and function of Mreg, DCreg, and MDSC are required to maximize the utility of these cells. In addition to conventional IS drugs, novel therapeutic agents can promote the regulatory function of RMC while preventing their immunostimulatory potential. These agents are likely to prove of considerable importance in exploiting the properties of RMC to promote transplant tolerance.

ACKNOWLEDGMENT

The authors thank Ms. Miriam Freeman for excellent administrative support.