ARE Th2 HELPER T LYMPHOCYTES BENEFICIAL, DELETERIOUS, OR IRRELEVANT IN PROMOTING ALLOGRAFT SURVIVAL?1

It is widely accepted that Th1 helper T lymphocytes (HTL)* initiate allograft rejection by promoting the development of alloantigen-specific cytotoxic T lymphocytes (CTL) and delayed-type hypersensitivity (DTH) responses. Since Th2 HTL are antagonistic to Th1 development and function, a popular hypothesis states that preferential induction of allograft-specific Th2 would down-regulate Th1-driven rejection responses, thereby promoting allograft acceptance. This possibility would have direct application to clinical transplantation, in that preferential induction of Th2 may be accomplished by cytokine manipulation both in vitro and in vivo. While attractive, this hypothesis has not been rigorously tested, and the majority of evidence that supports or refutes its premises is associative in nature. In this Overview, studies that associate allograft-specific Th2 with both beneficial and deleterious responses are summarized, and recent studies aimed at preferential induction of allograft-specific Th2 through cytokine manipulation are discussed. THE Th1/Th2 PARADIGM HTL regulate the nature of an immune response by the profile of the cytokines they secrete. In both mouse and man, polarized HTL phenotypes have been described that produce discrete profiles of cytokines, and production of selected cytokine profiles has been associated with a variety of disease states (reviewed in 1). Th1 HTL produce interleukin (IL) 2, interferon (IFN)-γ, and lymphotoxin, which promote CTL development, DTH responses, and IgG2a antibody isotype switching, which facilitates antibody-dependent cellular cytotoxicity (ADCC). Therefore, Th1 promote cellular immune responses that are believed to serve as terminal effector mechanisms in allograft rejection (reviewed in 2). Th2 HTL produce IL-4, IL-5, IL-6, IL-9, IL-10, and IL-13, thereby promoting humoral, allergic, and mucosal responses (1). As discussed below, Th1 and Th2 perpetuate their own development and are antagonistic to one another. Since Th2 are inhibitory to the development and function of Th1, it has been proposed that preferential induction of allograft-specific Th2 would inhibit Th1-regulated rejection responses, allowing a reduction or elimination of conventional immunosuppression. Further, an early Th1 → Th2 shift would be expected to decrease the number and severity of acute rejection episodes that are believed to contribute to the development of chronic rejection. Hence, cytokine manipulation represents an attractive, albeit speculative, therapeutic approach aimed at favoring the development of tolerogenic Th2. However, several aspects of this hypothesis must hold true if cytokine-based therapies are to have utility in transplantation: Are Th1-regulated effector mechanisms (CTL, DTH, ADCC) the only culprits in graft rejection? Are Th2 also capable of promoting CTL-, DTH-, and ADCC-mediated rejection? Do Th2 promote the development of alternate effector mechanisms that may surface in the absence of classical Th1-regulated responses? Are behaviors of Th1 and Th2 that have been defined primarily in in vitro systems operative during in vivo alloimmune responses? The answers to these questions are emerging, and the picture appears more complex than we might like. CYTOKINE MANIPULATION AND THE INDUCTION OF Th1 AND Th2 Th1 and Th2 subsets are thought to derive from a third phenotype, Th0, which produces a combination of cytokines characteristic of both Th1 and Th2 (1). The pattern of cytokines present during the initial stimulation of precursor cells dictates whether Th0 will preferentially develop into polarized Th1 or Th2. For example, Gajewski et al. (3) demonstrated that the presence of IFNγ during in vitro stimulation results in the development of T cells that produce IFNγ, but not IL-4, upon restimulation. Likewise, Swain et al. (4) found that the presence of IL-4 causes precursor cells to give rise to HTL that produce IL-4 and IL-5, but little or no IL-2 or IFNγ. Finkelman et al. (5) reported that in vivo neutralization of IFNγ promotes an antibody isotype shift from IgG2a (IFNγ regulated) to IgG1 (IL-4 regulated) when mice are challenged with the Th1-inducing pathogen Brucella abortus. Likewise, in vivo treatment with anti-IL-4 monoclonal antibodies (mAb) ablates the Th2-regulated IgE response to Nippostrongylus brasiliensis(6). Hence, cytokine profiles may be manipulated to influence Th1/Th2 balance, both in vitro and in vivo. IL-12 represents a key regulatory cytokine that promotes Th1 and inhibits Th2 differentiation (reviewed in 7). Indeed, neutralizing endogenous IL-12 converts Th1-dominated responses to Th2, while administration of exogenous IL-12 converts Th2 responses to Th1 in vivo (8). IL-12 is produced by stimulated antigen-presenting cells (APC), and its production is up-regulated by IFNγ and granulocyte-macrophage colony-stimulating factor. IL-12 up-regulates IFNγ production by T cells and natural killer cells which, in turn, promotes the development of polarized Th1 and the subsequent expression of cellular immunity. IFNγ is antagonistic to many of the activities of IL-4 (1), and inhibits the proliferation of Th2, but not Th1 (3). In addition to this indirect influence of IL-12 on Th1/Th2 balance, IL-12 has been shown to directly promote Th1 and inhibit Th2 development independent of IFNγ in some experimental systems (9, 10). It follows that neutralizing IL-12 activity would mute Th1-regulated rejection responses, and favor the development of graft-specific Th2. Hence, IL-12 antagonists may be useful as an inductive immunosuppressive therapy aimed at promoting an early posttransplant Th1 → Th2 shift. Experiments that test this hypothesis are discussed below. IL-4 is believed to be central to the development of Th2 (reviewed in 1, 7). Mast cells (11) and/or a rare population of NK1.1+ CD4+ T cells (12) are believed to be the initial source of IL-4 required for Th0 differentiation into Th2. IL-4 subsequently produced by differentiated Th2 serves as an autocrine growth factor, perpetuating clonal expansion of Th2. IL-4 indirectly inhibits Th1 development and function by decreasing IL-2 receptor expression (13) and by antagonizing many of the activities mediated by IFNγ (1). In addition, Th2-derived IL-10 inhibits cytokine production by established Th1, but not Th2, apparently by altering APC function (reviewed in 14). Specifically, IL-10 inhibits IL-12 production (7), providing another level at which Th2 may down-regulate Th1 responses. IL-10 has been reported to suppress expression of the costimulatory molecule B7 (15), which may be more requisite for Th1 than Th2 responses (16). Hence, IL-10 supplementation has been explored as an inductive strategy aimed at promoting a dominant posttransplantation Th2 response (see below). GUILT BY ASSOCIATION Much of the evidence that relates Th2 cytokine production to either graft survival or rejection is based on assessing intragraft cytokine gene expression by reverse transcriptase-polymerase chain reaction (PCR) or identifying graft infiltrating cells (GIC) that stain for cytokine protein by immunohistochemistry within allografts. Several issues must be considered when interpreting these data. First, these assays do not take into account the antigen specificity of the cells producing cytokines. Limiting dilution analysis (LDA) studies have revealed that the frequency of donor-specific cells infiltrating allografts is fairly low (17), indicating that the majority of GIC are irrelevant to the rejection process. Hence, RT-PCR may detect cytokine mRNAs produced by cells of irrelevant specificity that are trafficking through the graft during routine immune surveillance. It should also be noted that the presence of cytokine mRNA may not always reflect secretion of cytokine protein. For example, Dallman et al. (18) reported that while IL-2 mRNA was readily detectable in renal allografts of rats made tolerant by donor-specific transfusion, GIC failed to make IL-2 upon restimulation. Finally, the number of GIC that stain positive for cytokines by immunohistochemistry is considerably higher than the number of donor-specific cells quantified by LDA (i.e., 16). This discrepancy is not readily explained, but may reflect cytokine production by irrelevant cells that would not be detected by antigen-specific LDA. It is also possible that immunohistochemistry detects cytokine that is bound to receptor-positive cells in addition to cells that are actively producing cytokines. Additional concerns stem from the reported variability of intragraft cytokine gene expression, which is associated with graft rejection. For example, Strom and co-workers reported that intragraft IL-2 gene expression correlated with mouse pancreatic islet allograft rejection (19), and that IL-4 mRNA was rare or not detectable. In syngeneic islet transplants, IL-2 and IL-4 mRNAs were not detectable. In a subsequent study, these investigators found that intragraft IL-2 expression did not correlate with human renal allograft rejection (20). In this setting, expression of IL-15, which shares biologic activities with IL-2, appeared to be a better indicator of acute rejection (20). In contrast, Martinez and Krams reported that intragraft expression of IL-5 correlated with human renal (21) and liver (22) allograft rejection. IL-2 and IFNγ mRNAs were rarely detected (21), and IL-4 mRNA was present in the majority of liver allografts regardless of clinical status (22). Interestingly, Thai et al. (23) found up-regulation of both Th1 (IL-2, IFNγ) and Th2 (IL-4, IL-10) cytokine mRNAs in mouse liver allografts that were spontaneously accepted without immunosuppression. While these discrepancies may be explained in part by differences in species and/or transplanted tissues, variability in intragraft cytokine profiles have been reported by laboratories using the same experimental system. In the immediately vascularized mouse cardiac allograft model, Dallman et al. (24) reported that IFNγ and IL-5 mRNAs were detectable in normal heart tissue and were up-regulated in both syngeneic and allogeneic transplants. In this report (24), IL-2, IL-4, and lymphotoxin were expressed only in allografts. In contrast, Morgan et al. (25) found that IFNγ mRNA was detectable only in rejecting allografts, as was the case for IL-2 and IL-4. However, detection of IL-4 mRNA required the use of nested PCR techniques, which markedly enhances PCR sensitivity. Finally, we reported that unmodified cardiac allograft rejection was characterized by intragraft IL-2 and IFNγ expression, and that Th2 cytokine mRNAs were not readily detectable (26). In our hands, intragraft IL-4, IL-5, and IL-10 mRNAs were prevalent only when recipient mice were depleted of CD8+ T cells (26), or are treated with IL-12-neutralizing agents (27, see below). This discussion is not meant to detract from studies that evaluate intragraft cytokine profiles as a measure of Th1/Th2 dominance, but to emphasize that extrapolations between experimental systems (and even between different laboratories working in similar systems) should be viewed with some degree of skepticism. Hence, this line of investigation is likely most informative when shifts in cytokine profiles are related to permutations within a given set of interrelated experiments. Th2 CYTOKINES ARE ASSOCIATED WITH PROLONGED GRAFT SURVIVAL Streilein and co-workers initially established a relationship between Th2 and neonatal tolerance induced by injecting newborn mice with allogeneic cells (28, 29). Compared with normal animals, splenocytes from tolerant adult mice produced less IL-2 and more IL-4 upon in vitro stimulation with tolerizing alloantigens. Altered cytokine production was associated with an increase in the frequency of IL-4-producing cells and a corresponding decrease in the number of IL-2-producing cells (29). Donckier et al. (30) found that treatment of neonatal mice with anti-IL-4 mAb during the inductive phase reversed tolerance, pointing to a critical role for IL-4 in this process. Neutralizing IL-4 during the inductive phase partially restored the ability of splenocytes to produce IL-2 and IFNγ upon restimulation, but had no effect on CTL hyporesponsiveness, which is an additional feature of this system. Similar observations were reported by Gao et al. (31) using anti-IL-4 and anti-IL-10 mAb. These authors further characterized CTL hyporesponsiveness, and found that when splenocytes from tolerant animals were cultured with normal lymphocytes, CTL responses were not suppressed. Additional mixing experiments revealed that CD4+ T cells from tolerant mice were capable of providing help for isolated CD8+ CTL derived from naive animals. Further, the addition of normal CD4+ cells to cultures of tolerant CD8+ cells did not reverse CTL hyporesponsiveness (31). Collectively, these studies suggest that neonatal tolerance is associated not only with an emergence of Th2, but also with deletion and/or anergy of CD8+ CTL. Several immunosuppressive regimens that promote adult allograft survival appear to inhibit Th1 yet spare Th2 cytokine production (i.e., 16, 32-34). This observation suggests that “graft accommodation,” which occurs in some individuals with time after transplantation, results in part from an immunosuppressive drug-induced shift from a Th1 → Th2 dominated response. In support of this possibility, Gajewski et al. (35) reported that Th1 clones were more susceptible to CsA than were Th2 clones, and Ramirez et al. (36) found that the presence of glucocorticoids during the primary stimulation of CD4+ T cells promoted expression of Th2, but not Th1 mRNAs upon restimulation of these cells. Further, Sayegh et al. (16) reported that blocking the CD28-B7 T cell costimulatory pathway promoted long-term acceptance of rat renal allografts, which was associated with inhibition of intragraft IL-2 and IFNγ, but not IL-4 and IL-10 as assessed by immunohistochemistry. In contrast, Larsen et al. (37) found that blocking either the CD28-B7 or CD40 ligand (CD40L)-CD40 costimulatory pathways lead to prolonged survival of vascularized mouse cardiac allografts, but had little effect on intragraft expression of IL-2, IFNγ, and IL-10 mRNAs. Unlike the findings of Sayegh et al., these investigators reported decreased intragraft expression of IL-4 after CD28-B7 blockade (37). Still, although blockade of either the CD28-B7 or CD40L-CD40 costimulatory pathways resulted in prolonged graft survival, manifestations of chronic rejection developed. Interestingly, simultaneous blockade of both the CD28-B7 and CD40L-CD40 pathways inhibited expression of both Th1 and Th2 cytokines and ablated chronic rejection (37). These observations indicate that Th2 cytokines may not be necessary for tolerance induction, and that long-term graft survival free of chronic rejection may require inhibiting both Th1 and Th2 function. This idea has been more fully developed in a recent review by Strom et al. (38), who point out that allograft tolerance may be induced in IL-4 knockout mice. DO Th2 CELLS PLAY A DOMINANT PROTECTIVE ROLE IN TRANSPLANTATION? In general, the hypothesis that preferential induction of Th2 would promote graft survival is based on the ability of Th2 cytokines to down-regulate Th1 development and function (1, 7). However, several reports indicate that Th2 do not completely inhibit Th1 responses after transplantation, and may in fact promote alternate forms of rejection. Lang et al. (39) reported that during acute human liver allograft rejection, bile levels of IL-4 and IL-10 were approximately 10-fold higher than those observed for IL-2 and IFNγ, indicating that overexpression of Th2 cytokines is not protective. We have induced intragraft Th2 cytokine gene expression in the mouse vascularized cardiac transplant model by either depleting recipients of CD8+ T cells (26) or by treatment with IL-12 antagonists (27). In both settings, intragraft IFNγ expression was not inhibited and grafts were rejected, indicating that Th2 cytokines did not prevent in vivo Th1 development. Further, Chen et al. (40) reported that induction of neonatal tolerance was dependent upon ablating Th1 development; treatment with exogenous IFNγ prevented tolerance despite the fact that IL-4-producing Th2 were still present. Others have evaluated the ability of exogenous IL-10, which inhibits Th1 cytokine production in vitro, to promote graft survival. Contrary to what might be expected, Qian et al. (41) found that posttransplant treatment with mouse IL-10 accelerated cardiac allograft rejection. Similarly, Zheng et al. (42) treated islet cell allograft recipients with a noncytolytic IL-10-immunoglobulin fusion protein, which has an enhanced in vivo half-life relative to native IL-10. Treatment with this IL-10 fusion protein tended to accelerate, rather than prolong, graft survival and increased expression of the activated CTL product, granzyme B, in the regional lymph nodes. In contrast, Qin et al. (43) reported that retroviral-mediated gene transfer of viral (v) IL-10, which lacks several of the proinflammatory activities of cellular IL-10 (14), prolonged graft survival in the nonvascularized neonatal cardiac transplant model. This approach has been extended to the vascularized cardiac transplant model, in which a plasmid encoding vIL-10 is introduced by in situ perfusion of the graft with DNA:liposome complexes (L.A. DeBruyne, D.K. Bishop, J.S. Bromberg, unpublished). DNA:liposome-mediated vIL-10 gene transfer significantly prolonged vascularized graft survival, which was associated with decreased HTL, CTL, and alloantibody responses. However, histologic evaluation revealed extensive deposition of collagen and fibrous material within the graft, suggesting that vIL-10 gene transfer may result in an accelerated pathology similar to chronic rejection. This observation supports the idea that Th2 cytokines may contribute, in part, to the development of chronic rejection (37, 38). Th2 CYTOKINES ARE ASSOCIATED WITH GRAFT REJECTION While the above studies suggest that Th2 do not play a dominant protective role and may in fact be irrelevant in alloimmunity, additional lines of evidence indicate that Th2 may actually be For example, Martinez and Krams reported an between human renal (21) and liver (22) allograft rejection and intragraft production of Th2 cytokines. In these IL-5, rather than IL-2 or IFNγ, during rejection Interestingly, IL-5 expression was associated with an of in the majority of rejecting liver allografts Similarly, we reported that in cardiac allografts was associated with intragraft Th2 cytokine gene expression after in vivo of CD8+ T cells (26). which are to Th2 cytokines, are in the of and several a role for as effector cells in allograft rejection to be have been shown to of that are to produce IL-4 It should be noted that in mouse cardiac allografts was observed only when recipients were depleted of CD8+ and not when were to Th2 cytokines and D.K. Bishop, unpublished). This may be related to our observation that CD8+ T cells are not readily to a Th2 in and their ability to produce IFNγ that promote Th2 development see below). Hence, CD8+ Th1 appears to be requisite for Th2 of directly test the hypothesis that Th2 are capable of allograft rejection, the polarized Th2 by in vitro stimulation of splenocytes with in the presence of IL-4. these Th2 were into syngeneic mice cardiac allografts, graft rejection with and similar to that observed in recipients with polarized these Th2 their polarized after and were readily detectable within the rejecting allografts. In similar we cardiac mice with splenocytes from mice that had cardiac allografts D.K. Bishop, unpublished). transfer of these splenocytes resulted in accelerated rejection, characterized by of the cardiac Collectively, these observations indicate that Th2 are capable of allograft rejection, and may do in a Th1/Th2 AND SURVIVAL As discussed IL-12 is a key in Th1 Th2 dominance, this cytokine promotes Th1 and inhibits Th2 development in a variety of experimental systems Hence, we evaluated the use of IL-12 antagonists as an inductive therapy to promote Th2 and inhibit Th1 In our initial studies cardiac allograft recipients were treated with either neutralizing antibodies or of the IL-12 which to the of the IL-12 receptor and serve as a of IL-12 IL-12 with IL-12 would a neutralizing in that represents a native protein and should not be In contrast, antibodies are in and may an antibody which would the of IL-12 treatment with antibodies or induced Th2 cytokine expression, graft rejection was rather than inhibited (27). Further, IL-12 failed to inhibit IFNγ gene expression or in vivo Th1 development, indicating that IL-12 is not requisite for studies revealed that while IL-12 was in fact antagonistic to CD4+ Th1 development, blockade with stimulated CD8+ Th1 development this through the of the or of and on CD4+ CD8+ cells to be The recent development of mice should into these It is to that IL-12 is produced in of IL-12 (7), which suggests that may play a regulatory role in responses in vivo. whether produced IL-12 has biologic activity in the of transplantation, we Th1 responses in cardiac allograft recipients that were in either IL-12 or knockout produces IL-12 While IL-12 mice are capable of producing mice a IL-12 knockout in that is not in the absence of Th1 developed in both and cardiac allograft However, in vivo Th1 development was enhanced in recipients with knockout which suggests that endogenous produced in the knockout may This possibility was further by the observation that in vivo treatment of allograft recipients with mAb decreased Th1 function to the level in knockout The above studies that Th1 develop independent of and that IL-12 may have both CD8+ and antagonistic CD4+ whether endogenous IFNγ, which may be produced independent of to Th1 development, IL-12 allograft recipients were treated with neutralizing Interestingly, neutralizing IFNγ had no effect on Th1 development, which suggests a and pathway for Th1 after These observations emphasize the nature of Further, they the for cytokine manipulation aimed at inhibiting graft-specific Th1 while Th2. The most inductive will likely to be those that both Th1 and Th2 responses. In these indicate that Th2 may be and/or irrelevant in the of These observations likely reflect in inductive the of and whether the graft is immediately vascularized or further studies are to better deleterious activities of Th2, and to if is beneficial Th2 responses to allografts. beneficial activities are may these Th2 be from the and promote allograft what we have the to this will likely be that not that The authors for experimental observations their and and for IL-12 and for many