Leukemia (2007) 21, 1387–1394 & 2007 Nature Publishing Group All rights reserved 0887-6924/07 $30.00 www.nature.com/leu REVIEW

Mechanisms of action of antithymocyte globulin: T-cell depletion and beyond

M Mohty1,2,3

1Unite´ de Transplantation et de The´rapie Cellulaire (UTTC), Institut Paoli-Calmettes, Marseille, France; 2Universite´ de la Me´diterrane´e, and INSERM UMR 599, Marseille, France and 3Reduced Intensity Conditioning (RIC) Subcommittee of the Acute Leukemia Working Party of EBMT

The success of allogeneic stem cell transplantation and solid- prevention or rescue treatment of acute rejection in SOT, organ transplantation owes much to improvements in the conditioning for allo-SCT, GVHD, treatment of severe aplastic immunosuppressive regimens that prevent graft-versus-host anemia and various autoimmune diseases. An ever-growing disease (GVHD) or suppress allograft rejection. A better under- standing of the immune mechanisms underlying induction of body of recent data suggested that ATG affect not only T immunological tolerance is the key to successful transplanta- , but also other key cells involved in the immune tion. Polyclonal such as antithymocyte globulins reaction. Indeed, the immunosuppressive activity of ATG has (ATG) have been used for decades. The common belief is that been thought to result primarily from the depletion of peripheral ATG efficacy relies on its capacity to deplete T lymphocytes. T lymphocytes from the circulating pool through complement- The aim of this review is to offer an overview of the recent dependent lysis or activation-associated .1,2 Other findings that have been demonstrated in ATG’s immunomodu- potential mechanisms of action include modulation of surface latory activity. The polyclonal nature of ATG is reflected in its 3 diverse effects on the : (1) T-cell depletion in adhesion molecules or receptor expression. In and peripheral lymphoid tissues through complement- addition to their T-cell-depleting properties, ATGs can modulate dependent lysis and T-cell activation and apoptosis; (2) the by affecting or interfering with the modulation of key cell surface molecules that mediate leuko- function of different immune effectors such as B lymphocytes, cyte/endothelium interactions; (3) induction of apoptosis in B- regulatory T lymphocytes (Treg), natural killer (NK)-T lympho- cell lineages; (4) interference with functional properties; and (5) induction of regulatory T and natural killer cytes and dendritic cells (DC). The aim of this review is to T cells. As a consequence, ATG provides multifaceted synthesize the recent biological effects that have been involved immunomodulation paving the way for future applications in ATGs immunomodulatory activity, and to suggest prospective and suggesting that the use of ATG should be included in future applications. the immunosuppression therapeutic armamentarium to help reduce the incidence of organ rejection and GVHD. Leukemia (2007) 21, 1387–1394; doi:10.1038/sj.leu.2404683; published online 5 April 2007 Generation of ATG, current available preparations, Keywords: ; T-cell depletion; regulatory specificities and pharmacokinetics T cells; dendritic cells; transplantation; immunosuppression Polyclonal ATG is the purified IgG fraction of sera from rabbits, horses or, more rarely, goats that are immunized with Introduction or T-cell lines. The most widely used preparations include horse ATG (Lymphoglobulin, Genzyme) and rabbit ATG The success of allogeneic stem cell transplantation (allo-SCT) (Thymoglobulin, Genzyme). ATG Fresenius (Fresenius-Biotech- and solid-organ transplantation (SOT) owes much to improve- GMBH, Gra¨felfing, Germany) includes only rabbit antibodies, ments in the immunosuppressive regimens that prevent graft- and is produced against the Jurkat cell line, which resembles versus-host disease (GVHD) and/or suppress allograft rejection. activated T cells. ATG Pharmacia (Atgam, Pharmacia Upjohn) is For the past 20 years, the mainstay of immunosuppression has derived from the horses immunized with thymocytes. Despite been the calcineurin antagonists cyclosporine A (CSA) and sharing some common properties, these different products are . However, their immunosuppressive effects can be strictly different drugs, and have never been compared thus far unpredictable. Thus, a better understanding of the mechanisms in a controlled setting. Thus, one should be very cautious when of immunological tolerance to alloantigens, and development of deriving conclusions from the medical literature using these new immunosuppressive strategies are the keys to successful drugs. The rabbit preparation Thymoglobulin is probably the allograft function. The so-called ‘T-cell-depleting’ antibodies most extensively studied polyclonal antibody ATG preparation, such as antilymphocyte or antithymocyte globulins (ATG) have and most of the available in vitro mechanistic data are derived been used for decades. The general common belief is that ATG from experiments performed using Thymoglobulin. Therefore, efficacy relies on its potent capacity to deplete T lymphocytes of many data described in this review are from experiments the graft (as in allo-SCT) or of the host (as in SOT). ATG is used performed with this product. for treatment of various clinical conditions, including the Thymoglobulin is produced by immunizing pathogen-free New Zealand rabbits with fresh human thymocytes. Correspondence: Dr M Mohty, Unite´ de Transplantation et de fragments are removed from donors undergoing cardiac surgery. The´rapie Cellulaire (UTTC), Institut Paoli-Calmettes (IPC), 232 Bd. Virologic screens are run on every thymus fragment prior to their Sainte Marguerite, BP 156, Sainte Marguerite, F-13273 Marseille use in the manufacturing process. The viral safety of this starting Cedex 09, France. E-mail: [email protected] material is comparable to that of human blood derivatives used Received 28 January 2007; revised 3 March 2007; accepted 5 March in transfusions. In addition, Thymoglobulin is a pasteurized 2007; published online 5 April 2007 product. ATG and immunosuppression M Mohty 1388 Table 1 Summary of known target for antithymocyte conditioning regimen and the total dose of ATG appear to be globulin important factors on the overall clinical impact. However, persistence of significant long-term deficits in T-cell reconstitu- Immune Adhesion and cell trafficking Heterogeneous tion was linked to impaired thymopoiesis rather than to response pathways persistent active ATG that declined to sub-therapeutic levels antigens by 1 month after allo-SCT.18 CD1a CD11a/CD18 (LFA-1) CD2 CD3/TCR CD44 CD5 CD4 CD49/CD29 (VLA-4) CD6 ATG and T-cell depletion CD6 CD50 (ICAM-3) CD11b CD7 CD51/61 CD29 A well-documented effect of treatment with ATG is T-cell CD8 CD54 (ICAM-1) CD38 depletion. When looking at the composition of the ATG CD16 CD56 CD40 CD19 CD58 (LFA-3) CD45 preparation (Table 1), T-cell target antigens such as CD3/TCR CD20 LPAM-1(a4b7) CD95 and CD152 are T-lineage specific. Molecules such as CD2, CD25 CD102 (ICAM-2) CD126 CD4, CD5, CD6, CD8, CD25 and CD28 are expressed CD28 CD195 (CCR5) CD138 predominantly, although not exclusively, on T cells. HLA class CD30 CD197 (CCR7) I molecules can be found on all nucleated cells. CD2 is also CD32 CD184 (CXCR4) present on NK cells, and CD4 can be found on , DCs CD40 CD80 and . CD5 is found on a subset of B cells, and the CD86 activation marker CD25 is also present on B cells and CD152 (CTLA-4) monocytes. Both CD28 and CD30 are also present on activated HLA class I B cells. It is well established that low T-cell counts are usually HLA class II achieved in peripheral blood after ATG treatment, but the extent b2-M of T-cell depletion in lymphoid tissues has been characterized in cynomolgus monkeys using Thymoglobulin at low, high and very high doses. Four protocols of administration of Thymoglo- The immune cell types present in the normal human thymus, bulin were studied. The low dose (equivalent to 0.15–0.20 mg/ which serves as the source of thymocytes used to inoculate kg/day in humans) is not used clinically; however, it resulted in rabbits in the production of Thymoglobulin, have been serum concentrations equivalent to those observed in SOT. High extensively investigated. The distribution of cell types in the dose (equivalent to 0.8–1.0 mg/kg/day in humans) corresponds thymus includes T lymphocytes (72%), B lymphocytes (6%), and to the clinical dosage used in SOT. Very high dose (equivalent to -presenting cells (APC) and stromal cells (22%).4 Table 1 3.5 mg/kg/day in humans) corresponds to that used clinically for summarizes the multitude of antibody specificities detected in allo-SCT.10 In this animal model, there was a dose-dependent Thymoglobulin.2,5–13 An ever-expanding array of target antigens depletion of T cells as indicated by CD3 þ cell decrease. is being characterized in Thymoglobulin. These antigens can Moreover, both T-cell subsets, CD4 þ and CD8 þ cells, were broadly be categorized as immune response antigens, adhesion also depleted in a dose-dependent manner. CD20 þ cells (B and cell-trafficking molecules and molecules involved in cells) and CD16 þ /56 þ cells (NK cells) were also found to heterogeneous pathways. Antigens that have been shown to decrease. However, it is important to bear in mind that be involved in immune responses are found on T cells, B cells, circulating lymphocytes in the periphery at any given time NK cells, macrophages and DCs. Adhesion and cell-trafficking represent only a small percentage of total lymphocytes. Thus, antigens are also found abundantly. Other antigens involved examining depletion in the secondary lymphoid in heterogeneous pathways (inflammation, apoptosis, proli- organs would provide a more accurate picture of T-cell feration,yetc) have also been characterized (Table 1).2,4,7,8,10–15 depletion. Indeed, depletion of T cells was observed in the ATG pharmacokinetics depends on the dose and schedule of , and in the lymph nodes, where depletion of both CD4 þ administration as well as the number of ‘targeted’ immune and CD8 þ T cells was clearly dependent on the dose of effectors. The pharmacokinetics of ATG was studied initially in Thymoglobulin. B cells and NK cells were affected only at the the SOT setting, and clearance of ATG was found to be very high dose. These data extend the observation that consistent with single exponential decay and elimination half- Thymoglobulin provides T-cell depletion beyond peripheral life of 30 days demonstrating that rabbit ATG has a long half-life lymphocytes to depletion within secondary lymphoid tissues of in human plasma and an apparent volume of distribution of spleen and lymph nodes. The underlying mechanisms of T-cell about twice plasma volume.16 Studies of the clearance of ATG depletion by ATG include complement-dependent lysis, pre- given as part of the preparative regimen prior to allo-SCT dominantly in the blood compartment. Indeed, complement- indicated that the levels of rabbit ATG gradually declined, but dependent lysis is likely to contribute to the depletion of T cells, could still be detected up to 5 weeks after allo-SCT.17 especially in the intravascular compartment, where the con- Interestingly, the latter study found a correlation between the centration of complement is maximal. Concentrations grade of acute GVHD and the concentration of ATG in serum that are attained in the serum during treatment with ATG before transplantation. Patients with high serum levels of ATG efficiently activate human complement in clinically relevant before allo-SCT had a low risk of developing severe acute situations.1,19 In peripheral lymphoid tissues, apoptosis with GVHD, as compared to those with lower levels, suggesting that subsequent phagocytosis by macrophages appears to be the the measurement of ATG levels may be of value.17 However, primary mechanism. In the cynomolgus monkey model, T-cell these studies and most studies of ATG pharmacokinetics have depletion in the lymph nodes was associated with apoptosis. measured ATG in blood samples using an ELISA assay that may Apoptosis was documented by externalization of membrane overestimate the levels of biologically active antibody mole- phosphatidyl-serine measured by the binding of annexin V in cules. Waller et al. addressed this issue and could show that in FACS analysis. TUNEL analysis of sections of lymph nodes also all, the type of ATG, the timing of ATG administration during the revealed DNA breaks, indicative of apoptosis. With both

Leukemia ATG and immunosuppression M Mohty 1389 methods, the percent of apoptotic cells was related to the dose importance. LFA-1, the leukocyte function-associated antigen, of Thymoglobulin administered.10 A third mechanism is is an intercellular adhesion molecule selectively expressed on opsonization by immunoglobulin antibodies and complement leukocytes. The three ligands of LFA-1 are the leukocyte followed by phagocytosis of opsonized lymphocytes. Never- adhesion receptors ICAM-1, ICAM-2 and ICAM-3. LPAM-1 theless, the latter mechanism is still hypothetical.1,10 and VLA-4 are integrins, a large family of glycoproteins that With respect to activated T cells, depletion of pre-activated T attach cells to ligands on other cells. LPAM-1 is the ligand for cells with ATG may occur through two different pathways, the MAdCAM-1, and VLA-4 is the ligand for VCAM-1. CXCR4, antibody-dependent cell-mediated cytotoxicity or Fas-depen- CCR7 and CCR7 are chemokine receptors important for dent apoptosis. The concentrations of Thymoglobulin necessary leukocyte/endothelium interactions and leukocyte traffick- to trigger either cytotoxic mechanism are lower (10 mg/ml) than ing.2,3,7,8,10,12 Surface expression of LFA-1 has been determined those required to destroy resting T cells (100 mg/ml).1,20 Because at various times in the presence of increasing concentrations of of the high affinity of Thymoglobulin antibodies for the human Thymoglobulin. The results show a down-modulation of LFA-1 Fc receptor, antibody-dependent cell-mediated cytotoxicity can on lymphocytes, monocytes and . This downmodu- be achieved at very low antibody density on the membrane of lation occurs even at low drug concentrations. However, there is targeted T cells. This activity is independent of antibody a clear dose response to increasing drug concentrations. The specificity, and it occurs primarily with activated T cells as down-modulation of LFA-1 makes it no longer available for cell– targets. Resting lymphocytes are less susceptible to this cytotoxic cell interactions, such as those in leukocyte/endothelium pathway.1 The translation of these findings into clinical interactions. Thymoglobulin also contains antibodies to ICAM- treatment regimens requires some caution. The central mechan- 1 (CD54), the main ligand of LFA-1, ICAM-2 (CD102) and isms of depletion (i.e. complement-dependent lysis and activa- ICAM-3 (CD50). The ability to down-modulate LFA-1 occurs at tion-associated apoptosis) are dose-dependent and are not drug concentrations that are substantially below concentrations influenced by corticosteroids, CSA or rapamycin. T-cell deple- achieved in clinical use (80–200 mg/ml).1,2,19 A similar down- tion in both the intravascular and lymphoid compartment modulation of CD49d, the b chain of the leukocyte integrin requires supramitogenic concentrations of Thymoglobulin that VLA-4, is observed on lymphocytes and monocytes when can be achieved with a dose of 2–3 mg/kg for 2–4 days. This is incubated with increasing drug concentrations. The leukocyte the reason why low-dose regimens are probably not sufficient integrin b-4-beta-7 (LPAM-1) has also been shown to be for a substantial depletion in the secondary lymphoid organs but downmodulated. Thymoglobulin also contains antibodies to will achieve blood levels high enough for antibody-dependent CD49e and CD49f (VLA-5 and VLA-6).2 cell cytotoxicity (0.1–1 mg/ml) and activation-induced cell death With respect to chemokine receptors, human lymphocytes or (10–100 mg/ml). phytohemagglutinin-activated blasts mixed in vitro with Thymo- globulin shows moderate, but significant down-modulation of cell-surface expression of CCR7.2 In addition to down- Modulation of adhesion and cell-trafficking molecules modulation of the CCR7 , the presence of by ATG antibodies to CXCR4 and CCR5 was shown by the decreased Adhesion molecules and leukocyte infiltration play a key role in binding of specific monoclonal antibodies to lymphocytes and SOT (i.e. ischemia-reperfusion injury), GVHD and rejection.21–24 monocytes.2 Lymphocyte/endothelium interactions are a complex series of The range of adhesion molecules and chemokine receptors events initiated by lymphocytes sticking then rolling along the down-modulated may contribute to the therapeutic effect of surface of the endothelium. Inflammatory mediators such as ATG. The latter was demonstrated in a primate model of are released and the endothelium changes from an ischemia-reperfusion. In this model, ABO-compatible human anti-adhesive to a pro-adhesive status. Selectins and integrins on blood was pre-incubated with saline, and a CD25 monoclonal the surfaces of lymphocytes and the endothelium play a critical antibody or Thymoglobulin. The treated blood was infused into role in this process. These events are important in the initiation the femoral artery of a cynomolgus monkey after 1 h of warm of allograft rejection and ischemia reperfusion injury. After ischemia. The microcirculation was observed through an intra- engagement, an activation step occurs next with fixation of vital microscope and CDD camera. The results demonstrated b- on endothelial cells through CXCR4, CCR5 rolling and massive sticking of leukocytes with control blood and CCR7. Leukocyte integrin b-4-beta-7 (LPAM-1), LFA-1 and treated with saline or with a CD25 . By VLA-4 are activated. Leukocyte arrest is achieved through the comparison, there was no clumping or agglutination with blood interactions of integrins and immunoglobulin superfamily treated with Thymoglobulin.27 adhesion receptors. Thymoglobulin contains antibodies specific While it is now clearly established that ATG induces the for the integrins LPAM-1 and VLA-4, the chemokine receptors modulation of key functional molecules that mediate leukocyte/ CXCR4, CCR5 and CCR7, and the leukocyte adhesion receptors endothelium interactions, and interferes with leukocyte adhesion ICAM-1, -2 and -3 (Table 1).2,25 In general, modulation (i.e. the to the endothelium in vivo, some crucial factors should be taken internalization of the antigen–antibody complexes after binding) into consideration in optimizing clinical application of these is a major functional mechanism of action of ATG. When the findings. Timing of drug administration appears to be a key factor surface antigen is no longer expressed, the related pathway is in both the SOT3,28–30 and allo-SCT settings,31 but further studies inhibited as long as the antibody is present. Such down- are still needed to evaluate the differential impact of ATG on the regulation of surface antigens is not restricted to the blood increasing number of adhesion molecules and their tissue compartment but occurs in secondary lymphoid organs as location, recycling patterns and role in different clinical settings. well.10 Down-modulation is not inhibited by corticosteroids, calcineurin inhibitors or rapamycin.2,10,26 Such antibody- induced modulation modifies the interactions between leuko- Effects of ATG on B lymphocytes cytes and endothelium. Some of the target antigens in Thymoglobulin, involved in the modulation of cell surface B cells are precursors to plasma cells, which are involved in adhesion proteins and chemokine receptors, are of great humoral rejection. Interestingly, it has been reported previously

Leukemia ATG and immunosuppression M Mohty 1390 that polyclonal rabbit ATG can be used with plasmapheresis to activation, whereas immature DCs induce T-cell tolerance. treat antibody-mediated renal allograft rejection.32 ATG pre- A working paradigm for DC function holds on the fact that parations, especially Thymoglobulin, have been shown to immature DCs are distributed in peripheral tissues, and that they contain antibodies against a number of B-cell antigens, are specialized for uptake of pathogen-derived antigens.35,36 including B-cell-specific and non-specific surface proteins Pathogen-derived products can activate DCs via TLRs to migrate CD19, CD20, CD40, CD80, CD30, CD38 and CD95. Anti- to secondary lymphoid organs and to present antigens rather bodies against the -specific marker CD138 were also than antigen uptake.37 Mature DCs have increased expression of characterized.4,7,8,11,12,14 With this background, Zand et al. MHC and costimulatory molecules, and cytokines that promote tested the complement-independent, apoptosis-inducing prop- adaptive . erties of Thymoglobulin on naı¨ve B cells, CD40 ligand-activated In transplantation, DCs are likely to determine the balance B cells and plasma cells. Potential surface targets were assayed between rejection and tolerance. In SOT, implantation of the by competitive inhibition of monoclonal antibody binding. graft represents the ‘danger’ signal activating the immune Thymoglobulin was found strongly to induce apoptosis in vitro system. One response that takes place is the release of against naive, activated B cells and resident chemokines and cytokines from immature donor DCs within plasma cells at clinically relevant concentrations (1–100 ng/ml). the graft as well as the maturation of these DCs. These mature Such activity was described against numerous B-cell surface donor DCs enter the circulation via high endothelial venules proteins and the authors observed that cross-linking of CD30, and traffic to lymph nodes where they encounter recipient CD38, CD95, CD80 and HLA-DR likely accounts for this T cells (Figure 1). The interactions between activated DCs and activity. F(ab’)2 fragments of rATG showed 90% of the activity T cells would trigger rejection or tolerance. The maturation of the intact molecule, suggesting participation of the Fc status of DCs has also been proposed to play a role in the fragment. Inhibition of caspase- and cathepsin-dependent induction and maintenance of , with apoptotic pathways partially inhibits Thymoglobulin-induced immature DCs being considered as ‘tolerogenic’.38 Thus, one B-cell apoptosis.4 can imagine donor immune effectors entering secondary The impact of ATG on normal B cells was extended to lymphoid tissues in which they will encounter a mix of mature myeloma cells. Both ATG-Fresenius33 and Thymoglobulin4 and immature DCs, but also different DC subsets. In humans, at induced significant levels of apoptosis in myeloma cells as least two subsets of peripheral blood circulating DCs have been assayed by caspase induction, Annexin-V binding, subdiploid characterized thus far: the myeloid (MDC) and plasmacytoid DNA fragmentation, plasma membrane permeability, and loss (PDC) subsets. MDCs demonstrate remarkable plasticity and, of mitochondrial membrane potential. Addition of complement augmented the myeloma cell death greatly. Binding of Thymoglobulin to individual myeloma cell-surface proteins – Peripheral tissue primarily CD38, CD52, CD126 and CD138 – was demonstrated by competitive inhibition experiments with targeted monoclonal antibodies. Three pathways of cell death were identified involving caspase activation, cathepsin D and the genistein- «Alloantigens» CD8 sensitive tyrosine kinase pathway.4,33 Similarly, Timm et al.34 demonstrated potent anti-myeloma activity of Thymoglobulin in CD4 vitro and in vivo in an animal model of myeloma. Thymoglo- bulin was able to induce dose- and time-dependent apoptosis of 4 several myeloma cell lines, including those resistant to Immature DC (antigen conventional anti-myeloma agents. Importantly, the anti-mye- capture and DC loma activity was preserved even when myeloma cells were processing) grown with different cytokines demonstrating the ability to 1 overcome microenvironment-mediated resistance. Thymoglo- 4 bulin induced the apoptosis of freshly isolated primary myeloma CD8 cells from patients, and in a mouse model of CD4 myeloma, Thymoglobulin delayed the tumor growth in a dose- Mature DC B B dependent manner providing evidence for prospective evalua- (migration DC 2 B NK-T CD8 tion of ATG in patients with myeloma, either alone or in and ) NK-T DDC CD8 combination with other agents. CD8 4 NK-T 1 3 Treg CD8 5 Effects of ATG on DCs Treg Treg CD4 Treg CD4 Being the most potent APCs both in vitro and in vivo, DCs play a CD4 CDCD4 key role in the initiation of the immune response and are 1 considered promising targets for immunotherapy. In the transplantation field, both donor and recipient DCs play a major role, and depending on the environment in which the Figure 1 Proposed mechanisms through which ATG can interfere immune interactions between DCs and effector T cells take with the immune response. (1) T-cell depletion in blood and place, an inflammatory or a tolerogenic effect can occur. DCs peripheral lymphoid tissues through complement-dependent lysis and T-cell activation and apoptosis; (2) induction of B-cell apoptosis; can be activated by inflammatory cytokines, or microbial (3) modulation of key cell-surface molecules (adhesion and chemo- products such as lipopolysaccharide (LPS) and toll-like receptor kine receptors) that mediate leukocyte/endothelium interactions; (TLR) ligands. These effects are extremely important in produ- (4) Interference with DC functional properties (maturation and cing the ‘danger signals’ that mature DCs and lead to T-cell migration); and (5) Induction of Treg and NK-T cells.

Leukemia ATG and immunosuppression M Mohty 1391 depending on environment, can differentiate into either best-described population consists of cells expressing CD4, macrophages (with M-CSF) or distinct subsets of tissue-localized CD25 and the forkhead/winged helix transcription factor Foxp3. DC, that is, epithelial Langerhans cells (with IL-15 or TGF-b)or CD4 þ CD25 þ Tregs develop in the thymus.43 interstitial DCs (with IL-4). Upon activation, MDCs secrete IL-12 CD4 þ CD25highFoxp3 þ Treg cells are anergic, and can and mature into APCs able to prime T cells. On the other hand, inhibit CD4 þ CD25À T cells, CD8 þ T cells, DCs, NK cells PDC, also known as the type I IFN-producing cells, are a major and B cells in a cell-to-cell contact and dose-dependent member of the innate immunity effectors, specialized in rapidly manner.44–49 The IL-10 and TGF-beta cytokines are likely to secreting massive amounts of type I IFN following viral have a major functional importance for Treg cells in vivo, stimulation. PDCs can promote the function of NK cells, B cells, particularly in the context of disease.50,51 An ever-growing body T cells and MDCs through type 1 IFN during an antiviral immune of evidence supports important immunoregulatory functions of response (for a detailed review, see Liu39). The role of DCs is now Treg in maintaining both self-tolerance and tolerance toward well demonstrated as a factor of major importance in the autoantigens52 and alloantigens.53 Thus, efforts are currently hierarchy of the induction of immune reactions. Therefore, ongoing to optimize in vitro methods for Treg generation,54,55 GVHD or solid-organ-graft rejection pathophysiology is likely and clinical trials were already proposed to use ex vivo to involve complex interactions where DCs and effectors from expanded Treg.56,57 Therefore, understanding the conditions both the adaptive and innate immune system play a major role that are required for the generation of Treg would allow the before the establishment of the allogeneic immune response. development of novel strategies for inducing immunologic With this background, the impact of ATG has been investigated tolerance. With this background, Lopez et al.58 assayed the on the phenotype and function of DCs generated in vitro. First, ability of Thymoglobulin to impact the generation of Treg in Monti et al. attempted to determine whether ATG is able to bind vitro. They reported for the first time that Thymoglobulin, but not and interfere with human DC function. ATG bound to human the anti-CD52 mAb () or the IL-2R antagonists DCs at least in part by recognizing CD1a, MHC I, MHC II, causes rapid and sustained expansion of CD4 þ CD25 þ T cells CD11a, CD86, CD32, CD4, CD11b, CD29 and CD51/61. This when cultured with human peripheral blood lymphocytes. The binding, which was more relevant in mature DCs, induced expanded cells displayed enhanced expression of GITR, CTLA- complement-mediated lysis. ATG did not influence either the 4, Foxp3 and efficiently suppressed a direct alloimmune macropynocytosis properties of DCs or the receptor-mediated response of the original responder lymphocytes. It is interesting endocytosis of FITC-Dextran. In MLR assay, ATG was able to to note that these cells did not suppress memory responses. Ex significantly inhibit T-cell proliferation by binding on but vivo expansion of Treg was due mainly to conversion of not on DCs, suggesting that ATG is able to interfere in the CD4 þ CD25– into CD4 þ CD25 þ T cells and to a lesser degree activation of T cells by DCs in two different ways: by inhibiting to proliferation of natural CD4 þ CD25 þ T cells. The induction the capacity of lymphocyte to proliferate after the DC stimulation of Treg depended on production of Th2 cytokines in the and by inducing a complement-mediated lysis of DC.12 More generating cultures. Although the in vivo effects of ATG on the recently, Gillet-Hladky et al.40 investigated ATG effects on in induction, expansion and function of Treg in allograft recipients vitro differentiation and maturation of -derived DCs remain to be characterized fully, these novel data suggest that (Mo-DCs). ATG was found to inhibit the maturation of immature ATG not only may promote expansion/generation of Treg cells, Mo-DCs and allowed the generation of DCs expressing ILT-3, but may also be useful in future ex vivo expansion of these cells CD123, CCR6 but not CCR7 and producing indoleamine 2,3- for cellular therapy in and transplantation. dioxygenase mRNA, a phenotype compatible with tolerogenic In comparison to the naturally occurring Treg lymphocytes, DCs. In the allo-SCT setting, Fang et al. found that after NK-T cells are a subset of T cells that recognize CD1d, a incubation of peripheral blood mononuclear cells from healthy nonpolymorphic, class I, MHC-like antigen-presenting mole- donors with ATG-Fresenius and complement, there was a cule. Recognition is mediated by the means of a T-cell receptor significant reduction of both MDCs and PDCs. The mean relative that includes an invariant TCR b-chain with rearranged Va14 reduction of PDC and MDC numbers after ATG and complement and Ja281 gene segments in mice. NK-T cells are predominantly treatment was approximately 80 and 85%, respectively. The CD4À/CD8À, and secrete large amounts of IL-4 and IFN- authors further confirmed these findings in vivo in blood samples gamma after engagement of the CD3/TCR without previous from patients treated with or without ATG, collected immediately exposure to antigens.59 In a mouse model assessing the impact after completing cytotoxic therapy or after ATG treatment. of antithymocyte serum (the mouse equivalent to rabbit ATG), Despite a marked decrease of circulating MDCs and PDCs after Lan et al. showed that NK-T cells become predominant after both ATG and cytotoxic therapy, complete depletion of both total lymphoid irradiation (TLI) and rabbit antithymocyte serum typesofDCswasobservedinpatientsonlyaftertheATG- (ATS) treatment of C57BL/6 spleen cells. Interestingly, although containing conditioning.41 However, the actual impact of ATG the overall population of TCRab decreased from 33.2 to 3.6%, on tissue DCs such as Langerhans’ cells or dermal intestinal DCs NK-T cells represented 92.4% of these cells after completion of is still unknown. The latter might be of great clinical relevance, the regimen.60 Host mice conditioned with TLI and ATS have because these DC subsets are located in the organs mainly improved survival compared with those having total-body affected by GVHD. However, the above results support the irradiation (TBI) and ATS after allogeneic bone marrow hypothesis that reduction of GVHD or solid-organ-graft rejection transplantation (BMT). In addition, injection of ATS during the after ATG pretreatment may be caused not only by the substantial second week of TLI protected the hosts further such that the in vivo T-cell depletion, but also by a significant effect on DCs combined infusion of marrow and PBMC failed to kill any hosts that are likely oriented towards a tolerogenic phenotype. during a 120-day observation period.60 Most importantly, to determine whether protection against GVHD after TLI and ATS conditioning required the presence of host NK-T cells, the ability Effects of ATG on regulatory and NK-T cells of bone marrow and PBMC mixtures from MHC mismatched donors to induce GVHD in either wild-type or NK-T-cell Dominant tolerance is now known to be assured by multiple deficient CD1À/À hosts was determined. Differential survival of subtypes of regulatory and suppressor T lymphocytes.42 The wild-type and NK-T-cell deficient CD1À/À hosts after BMT from

Leukemia ATG and immunosuppression M Mohty 1392 MHC mismatched donors demonstrated the requirement for NK- clinically significant cellular rejection in SOT and GVHD in T cells in the suppression of GVHD.61 NK-T-cell suppression of allo-SCT. GVHD required IL-4,62 with the resultant Th2-polarized donor T cells as weaker inducers of GVHD compared with Th1- polarized donor T cells.63 The finding that conditioning with Acknowledgements TLI and ATS protects mice against acute GVHD after allo-SCT prompted the Stanford group to test this strategy in humans. Of We thank D Blaise (Institut Paoli-Calmettes, Marseille, France) 37 allo-SCT recipients, only two had acute GVHD after allo- and B Gaugler (INSERM UMR 599, Marseille, France) for helpful SCT. In the transplant recipients, the fraction of donor CD4 þ T discussions. We also thank the physicians of the Hematology and cells that produced IL-4 after in vitro stimulation increased by a Medical Oncology Departments at the Institut Paoli-Calmettes for factor of five, and the proliferative response to alloantigens in their important study contributions. We apologize for the vitro was reduced, as compared with normal control subjects colleagues whose work could not be cited for the lack of space. and control subjects who underwent conditioning with a single- We also thank the ‘Association pour la Recherche sur le low-dose TBI (200 cGy), strongly confirming the ability of a (ARC; ARECA Program)’, the ‘Ligue Nationale contre le Cancer’, regimen of TLI and Thymoglobulin to decrease the incidence of the ‘Fondation de France’, the ‘Fondation contre la Leuce´mie’, the acute GVHD, while allowing graft antitumor activity.64 Simi- ‘Agence de Biome´decine’, the ‘Association Cent pour Sang la Vie’ larly, in the SOT setting, Brennan et al. prospectively compared and the ‘Association Laurette Fuguain’, for their generous and short courses of ATG (Thymoglobulin) and basiliximab in continuous support for our clinical and basic research work. Our patients at high risk for acute rejection or delayed graft function group is supported by several grants from the French ministry of who received a renal transplant from a deceased donor. At 12 health as part of the ‘Programme Hospitalier de Recherche months, the ATG group, as compared with the basiliximab Clinique (PHRC)’. group, had significantly lower incidences of acute rejection and of acute rejection that required treatment with antibody, References suggesting that ATG-based induction therapy, as compared with basiliximab, reduced the incidence and severity of acute rejection in the renal SOT setting.65 1 Genestier L, Fournel S, Flacher M, Assossou O, Revillard JP, Bonnefoy-Berard N. Induction of Fas (Apo-1, CD95)-mediated apoptosis of activated lymphocytes by polyclonal antithymocyte globulins. Blood 1998; 91: 2360–2368. 2 Michallet MC, Preville X, Flacher M, Fournel S, Genestier L, Conclusions and future perspectives Revillard JP. Functional antibodies to leukocyte adhesion mole- cules in antithymocyte globulins. Transplantation 2003; 75: The polyclonal nature of ATG is reflected in its diverse effects on 657–662. 3 Brennan DC. Faith supported by reason: mechanistic support for the immune system: (1) T-cell depletion in blood and peripheral the use of polyclonal antibodies in transplantation. Transplantation lymphoid tissues through complement-dependent lysis and 2003; 75: 577–578. T-cell activation and apoptosis; (2) modulation of key cell- 4 Zand MS, Vo T, Huggins J, Felgar R, Liesveld J, Pellegrin T et al. surface molecules that mediate leukocyte/endothelium interac- Polyclonal rabbit antithymocyte globulin triggers B-cell and tions; (3) induction of apoptosis in B-cell lineages and myeloma plasma cell apoptosis by multiple pathways. Transplantation cell lines; (4) interference with DC functional properties; and (5) 2005; 79: 1507–1515. 5 Raefsky EL, Gascon P, Gratwohl A, Speck B, Young NS. Biological induction of Treg and NK-T cells. As a consequence, ATG and immunological characterization of ATG and ALG. Blood provides multifaceted immunomodulation (Figure 1), which is 1986; 68: 712–719. worth continuous investigation in different clinical settings. The 6 Bonnefoy-Berard N, Vincent C, Revillard JP. Antibodies against current research evidence suggests prospective future functional leukocyte surface molecules in polyclonal antilympho- applications for ATG based on the recently characterized cyte and antithymocyte globulins. Transplantation 1991; 51: immunomodulatory effects. Historically, however, ATG-based 669–673. 7 Rebellato LM, Gross U, Verbanac KM, Thomas JM. A comprehen- immunosuppressive therapy – both in SOT and allo-SCT – has sive definition of the major antibody specificities in polyclonal gone in and out of favor. In the early era of SOT and allo-SCT, rabbit antithymocyte globulin. Transplantation 1994; 57: 685–694. immunosuppressive options were limited to the use of corticos- 8 Bourdage JS, Hamlin DM. Comparative polyclonal antithymocyte teroids, and methotrexate. ATG was used clinically globulin and antilymphocyte/antilymphoblast globulin anti-CD in the 1970s, but there was controversy over its efficacy, antigen analysis by flow cytometry. Transplantation 1995; 59: production and especially production standardization. The 1194–1200. 9 Preville X, Nicolas L, Flacher M, Revillard J. A quantitative flow introduction of CSA in the early 1980s was a major advance. cytometry assay for the preclinical testing and pharmacological SOT and allo-SCT results were significantly improved and the monitoring of rabbit antilymphocyte globulins (rATG). J Immunol enthusiasm for ATG in general declined. The advent of reduced- Methods 2000; 245: 45–54. intensity conditioning regimens in allo-SCT, the increasing use 10 Preville X, Flacher M, LeMauff B, Beauchard S, Davelu P, Tiollier J of organs with a high risk of delayed function in the SOT field, et al. Mechanisms involved in antithymocyte globulin immuno- and the discovery of different new mechanisms of action for suppressive activity in a nonhuman primate model. Transplanta- tion 2001; 71: 460–468. ATG with a better knowledge of dosages, safety and tolerance 11 Pistillo MP, Tazzari PL, Bonifazi F, Bandini G, Kato T, Matsui T renewed the interest for the use of polyclonal antibodies. The et al. Detection of a novel specificity (CTLA-4) in ATG/TMG SOT and allo-SCT communities are now faced with the globulins and sera from ATG-treated leukemic patients. Trans- challenge on how to define new immunosuppressive strategies, plantation 2002; 73: 1295–1302. including ATG, which will optimize the outcome of transplanta- 12 Monti P, Allavena P, Di Carlo V, Piemonti L. Effects of anti- tion. Prospective and randomized studies assessing ATG benefits lymphocytes and anti-thymocytes globulin on human dendritic cells. Int Immunopharmacol 2003; 3: 189–196. are needed because a strong body of evidence suggests that the 13 Tsuge I, Kojima S. Comparison of antibody specificities of four use of ATG as part of the immunosuppression therapeutic anti-/anti-lymphocyte globulin products. Curr Ther Res armamentarium may help reduce the incidence and severity of 1995; 56: 671–677.

Leukemia ATG and immunosuppression M Mohty 1393 14 Ankersmit HJ, Roth GA, Moser B, Zuckermann A, Brunner M, has in vitro and in vivo activity in . Leukemia Rosin C et al. CD32-mediated platelet aggregation in vitro by anti- 2006; 20: 1863–1869. thymocyte globulin: implication of therapy-induced in vivo 35 Banchereau J, Steinman RM. Dendritic cells and the control of thrombocytopenia. Am J Transplant 2003; 3: 754–759. immunity. Nature 1998; 392: 245–252. 15 Zand MS, Vo T, Pellegrin T, Felgar R, Liesveld JL, Ifthikharuddin JJ 36 van Kooyk Y, Geijtenbeek TB. DC-SIGN: escape mechanism for et al. Apoptosis and complement-mediated lysis of myeloma cells pathogens. Nat Rev Immunol 2003; 3: 697–709. by polyclonal rabbit antithymocyte globulin. Blood 2006; 107: 37 Sallusto F, Lanzavecchia A. Efficient presentation of soluble 2895–2903. antigen by cultured human dendritic cells is maintained by 16 Bunn D, Lea C, Bevan D, Higgins R, Hendry B. The pharmaco- / colony-stimulating factor plus kinetics of anti-thymocyte globulin (ATG) following intravenous 4 and downregulated by alpha. J Exp Med infusion in man. Clin Nephrol 1996; 45: 29–32. 1994; 179: 1109–1118. 17 Remberger M, Sundberg B. Rabbit- levels in 38 Steinman RM, Hawiger D, Nussenzweig MC. Tolerogenic patients receiving thymoglobulin as part of conditioning before dendritic cells. Annu Rev Immunol 2003; 21: 685–711. unrelated donor stem cell transplantation. Haematologica 2005; 39 Liu YJ. IPC: professional type 1 interferon-producing cells and 90: 931–938. plasmacytoid dendritic cell precursors. Annu Rev Immunol 2005; 18 Waller EK, Langston AA, Lonial S, Cherry J, Somani J, Allen AJ 23: 275–306. et al. Pharmacokinetics and pharmacodynamics of anti-thymocyte 40 Gillet-Hladky S, de Carvalho CM, Bernaud J, Bendahou C, Bloy C, globulin in recipients of partially HLA-matched blood hemato- Rigal D. Rabbit antithymocyte globulin inhibits monocyte- poietic progenitor cell transplantation. Biol Blood Marrow derived dendritic cells maturation in vitro and polarizes mono- Transplant 2003; 9: 460–471. cyte-derived dendritic cells towards tolerogenic dendritic cells 19 Bonnefoy-Berard N, Revillard JP. Mechanisms of immunosuppres- expressing indoleamine 2,3-dioxygenase. Transplantation 2006; sion induced by antithymocyte globulins and OKT3. J Heart Lung 82: 965–974. Transplant 1996; 15: 435–442. 41 Fang L, Fehse B, Engel M, Zander A, Kroger N. Antithymocyte 20 Bonnefoy-Berard N, Genestier L, Preville X, Revillard JP. TNF globulin induces ex vivo and in vivo depletion of myeloid alpha and CD95-L contribute to apoptosis of activated lympho- and plasmacytoid dendritic cells. Transplantation 2005; 79: cytes triggered by ATGs. Transplant Proc 1999; 31: 775–777. 369–371. 21 Buysmann S, Bemelman FJ, Schellekens PT, van Kooyk Y, Figdor 42 Sakaguchi S. Naturally arising CD4+ regulatory t cells for CG, ten Berge IJ. Activation and increased expression of adhesion immunologic self-tolerance and negative control of immune molecules on peripheral blood lymphocytes is a mechanism for responses. Annu Rev Immunol 2004; 22: 531–562. the immediate lymphocytopenia after administration of OKT3. 43 Bensinger SJ, Bandeira A, Jordan MS, Caton AJ, Laufer TM. Major Blood 1996; 87: 404–411. histocompatibility complex class II-positive cortical epithelium 22 Muller TF, Kraus M, Neumann C, Lange H. Detection of renal mediates the selection of CD4(+)25(+) immunoregulatory T cells. allograft rejection by complement components C5A and TCC in J Exp Med 2001; 194: 427–438. plasma and urine. J Lab Clin Med 1997; 129: 62–71. 44 Lim HW, Hillsamer P, Banham AH, Kim CH. Cutting edge: direct 23 Fuggle SV, Koo DD. Cell adhesion molecules in clinical renal suppression of B cells by CD4+ CD25+ regulatory T cells. transplantation. Transplantation 1998; 65: 763–769. J Immunol 2005; 175: 4180–4183. 24 Burne-Taney MJ, Rabb H. The role of adhesion molecules and T 45 Azuma T, Takahashi T, Kunisato A, Kitamura T, Hirai H. Human cells in ischemic renal injury. Curr Opin Nephrol Hypertens 2003; CD4+ CD25+ regulatory T cells suppress NKT cell functions. 12: 85–90. Cancer Res 2003; 63: 4516–4520. 25 von Andrian UH, Mackay CR. T-cell function and migration. Two 46 Chen W. Dendritic cells and (CD4+)CD25+ T regulatory cells: sides of the same coin. N Engl J Med 2000; 343: 1020–1034. crosstalk between two professionals in immunity versus tolerance. 26 Merion RM, Howell T, Bromberg JS. Partial T-cell activation and Front Biosci 2006; 11: 1360–1370. anergy induction by polyclonal antithymocyte globulin. Trans- 47 Romagnani C, Della Chiesa M, Kohler S, Moewes B, Radbruch A, plantation 1998; 65: 1481–1489. Moretta L et al. Activation of human NK cells by plasmacytoid 27 Chappell D, Beiras-Fernandez A, Hammer C, Thein E. In vivo dendritic cells and its modulation by CD4+ T helper cells visualization of the effect of polyclonal antithymocyte globulins on and CD4+ CD25hi T regulatory cells. Eur J Immunol 2005; 35: the microcirculation after ischemia/reperfusion in a primate 2452–2458. model. Transplantation 2006; 81: 552–558. 48 Trzonkowski P, Szmit E, Mysliwska J, Dobyszuk A, Mysliwski A. 28 Agha IA, Rueda J, Alvarez A, Singer GG, Miller BW, Flavin K et al. CD4+CD25+ T regulatory cells inhibit cytotoxic activity of T Short course induction immunosuppression with thymoglobulin CD8+ and NK lymphocytes in the direct cell-to-cell interaction. for renal transplant recipients. Transplantation 2002; 73: Clin Immunol 2004; 112: 258–267. 473–475. 49 Murakami M, Sakamoto A, Bender J, Kappler J, Marrack P. 29 Goggins WC, Pascual MA, Powelson JA, Magee C, Tolkoff-Rubin CD25+CD4+ T cells contribute to the control of memory CD8+ N, Farrell ML et al. A prospective, randomized, clinical trial T cells. Proc Natl Acad Sci USA 2002; 99: 8832–8837. of intraoperative versus postoperative Thymoglobulin in adult 50 Nakamura K, Kitani A, Strober W. Cell contact-dependent cadaveric renal transplant recipients. Transplantation 2003; 76: immunosuppression by CD4(+)CD25(+) regulatory T cells is 798–802. mediated by cell surface-bound transforming growth factor beta. 30 Kaden J, Strobelt V, May G. Short and long-term results after J Exp Med 2001; 194: 629–644. pretransplant high-dose single ATG-fresenius bolus in cadaveric 51 Hara M, Kingsley CI, Niimi M, Read S, Turvey SE, Bushell AR et al. kidney transplantation. Transplant Proc 1998; 30: 4011–4014. IL-10 is required for regulatory T cells to mediate tolerance to 31 Mohty M, Bay JO, Faucher C, Choufi B, Bilger K, Tournilhac O alloantigens in vivo. J Immunol 2001; 166: 3789–3796. et al. Graft-versus-host disease following allogeneic transplanta- 52 Vigouroux S, Yvon E, Biagi E, Brenner MK. Antigen-induced tion from HLA-identical sibling with antithymocyte globulin- regulatory T cells. Blood 2004; 104: 26–33. based reduced-intensity preparative regimen. Blood 2003; 102: 53 Fontenot JD, Gavin MA, Rudensky AY. Foxp3 programs the 470–476. development and function of CD4+CD25+ regulatory T cells. 32 Shah A, Nadasdy T, Arend L, Brennan J, Leong N, Coppage M et al. Nat Immunol 2003; 4: 330–336. Treatment of C4d-positive acute humoral rejection with plasma- 54 Hoffmann P, Boeld TJ, Eder R, Albrecht J, Doser K, Piseshka B et al. pheresis and rabbit polyclonal antithymocyte globulin. Transplan- Isolation of CD4+CD25+ regulatory T cells for clinical trials. tation 2004; 77: 1399–1405. Biol Blood Marrow Transplant 2006; 12: 267–274. 33 Ayuk FA, Fang L, Fehse B, Zander AR, Kroger N. Antithymocyte 55 Li L, Godfrey WR, Porter SB, Ge Y, June CH, Blazar BR et al. globulin induces complement-dependent cell lysis and caspase- CD4+CD25+ regulatory T-cell lines from human cord blood have dependent apoptosis in myeloma cells. Exp Hematol 2005; 33: functional and molecular properties of T-cell anergy. Blood 2005; 1531–1536. 106: 3068–3073. 34 Timm MM, Kimlinger TK, Haug JL, Kline MP, Greipp PR, Rajkumar 56 Bluestone JA. Regulatory T-cell therapy: is it ready for the clinic? SV et al. Thymoglobulin targets multiple plasma cell antigens and Nat Rev Immunol 2005; 5: 343–349.

Leukemia ATG and immunosuppression M Mohty 1394 57 Gregori S, Bacchetta R, Hauben E, Battaglia M, Roncarolo MG. prevents graft-versus-host disease: the role of CD1-reactive Regulatory T cells: prospective for clinical application in natural killer T cells. Biol Blood Marrow Transplant 2003; 9: transplantation. Curr Opin Hematol 355–363. 2005; 12: 451–456. 62 Zeng D, Lewis D, Dejbakhsh-Jones S, Lan F, Garcia-Ojeda M, 58 Lopez M, Clarkson MR, Albin M, Sayegh MH, Najafian N. A novel Sibley R et al. Bone marrow NK1.1(À) and NK1.1(+) T cells mechanism of action for anti-thymocyte globulin: induction of reciprocally regulate acute graft versus host disease. J Exp Med CD4+CD25+Foxp3+ regulatory T cells. J Am Soc Nephrol 2006; 1999; 189: 1073–1081. 17: 2844–2853. 63 Fowler DH, Kurasawa K, Husebekk A, Cohen PA, Gress RE. Cells 59 Zeng D, Lan F, Hoffmann P, Strober S. Suppression of graft-versus- of Th2 cytokine phenotype prevent LPS-induced lethality host disease by naturally occurring regulatory T cells. Transplanta- during murine graft-versus-host reaction. Regulation of cytokines tion 2004; 77: S9–S11. and CD8+ lymphoid engraftment. J Immunol 1994; 152: 60 Lan F, Zeng D, Higuchi M, Huie P, Higgins JP, Strober S. 1004–1013. Predominance of NK1.1+TCR alpha beta+ or DX5+TCR alpha 64 Lowsky R, Takahashi T, Liu YP, Dejbakhsh-Jones S, Grumet FC, beta+ T cells in mice conditioned with fractionated lymphoid Shizuru JA et al. Protective conditioning for acute graft-versus-host irradiation protects against graft-versus-host disease: ‘natural disease. N Engl J Med 2005; 353: 1321–1331. suppressor’ cells. J Immunol 2001; 167: 2087–2096. 65 Brennan DC, Daller JA, Lake KD, Cibrik D, Del Castillo D. Rabbit 61 Lan F, Zeng D, Higuchi M, Higgins JP, Strober S. Host conditioning antithymocyte globulin versus basiliximab in renal transplantation. with total lymphoid irradiation and antithymocyte globulin N Engl J Med 2006; 355: 1967–1977.

Leukemia