Immune privilege: a recurrent theme in immunoregulation?

All bond and privilege of nature, break!Let it be virtuous to be obstinate. William Shakespeare, Coriolanus Act V, Scene III The word 'privilege' implies a special status. The concept of immune privilege was coined in the mid-20th century to describe observations that immunogenic tissues and tumors placed at certain anatomical locations survived longer than anticipated or were accepted indefinitely. The implication was that these anatomical sites were in some way specialized because otherwise potent antigenic stimuli failed to elicit effective immune responses. Historically, passive mechanisms, notably physical and physiologic barriers that maintained local segregation between tissues and immune cells and antibodies, were thought to offer the most likely explanations for localized immune privilege. However, the recognition that multiple mechanisms exist to suppress and subvert adaptive immunity and promote tolerance raises the prospect that active suppressive mechanisms help to maintain immune privilege. This volume of Immunological Reviews was inspired by the original concept of immune privilege and by recent rapid progress in identifying immunoregulatory processes that suppress immunity and promote immune tolerance. Our goal in this issue is to assess contemporary thinking on classical immune privilege and to determine if the concept of immune privilege might have wider implications and value for understanding immunoregulation in general. To this end, we solicited reviews from experts working on traditional immune-privileged tissues and on immunoregulatory processes. The classical definition of immune privilege and the consequences of active immunoregulation overlap in the sense that these processes prevent elaboration of effective immunity following challenges with otherwise potent antigenic stimuli. In each case, antigenic challenge provokes weak and ineffective responses, and in some cases, this challenge promotes long-term tolerance to subsequent antigen encounters, even when delivered via non-privileged sites. Presumably, the main purpose of immunoregulation is to prevent undesirable immune responses to peripheral self-antigens. However, self-tolerance mechanisms can also contribute to disease processes by allowing life-threatening disease agents, such as tumor cells and cells infected with certain pathogens, to persist; in effect, tolerance mechanisms become the 'Achilles' heel' of the adaptive immune system in these settings. Thus, the precise context in which initial antigen encounters take place has considerable consequences for eventual immune outcomes. Weakly immunogenic or tolerogenic outcomes following antigenic challenge are not consistent with descriptions of enhanced and accelerated secondary (memory or recall) responses provided as examples of immune system adaptation to primary antigenic challenge in most immunology textbooks. The self-/non-self hypothesis and the clonal selection hypothesis provide useful conceptual frameworks guiding research on lymphocyte selection during development driven by self-antigens and selection of lymphocytes that participate in immune responses driven by non-self antigens. However, the failure of fetal tissues, tumor cells, certain chronically infected cells, and innocuous microorganisms and allergens to provoke effective immune responses, despite their demonstrable antigenicity, belies the ability of the self-/non-self hypothesis to explain all immune system specificity. Otherwise, the mammalian adaptive immune system would mount effective (but undesirable) immune responses to gestating fetuses, would constantly attack commensal bacteria and allergens presented constitutively at mucosal surfaces, and, conversely, would mount effective (and desirable) immunity directed specifically against neoantigens expressed by tumor cells and against foreign antigens expressed by cells infected with pathogens that cause chronic infectious diseases, such as AIDS, tuberculosis, and Leishmaniasis. The implicit reason why the outcomes listed above occur very rarely (or not at all) is that natural immunoregulatory processes actively suppress effective immune responses to antigenic stimuli in each of these situations. This notion drives a considerable body of recent immunological research justified by the need to develop new and better immunotherapies to treat a wide range of diseases in which too much, or too little, tolerance plays a pathogenic role. Hence, the primary purpose in soliciting the reviews contained in this volume is to survey the current status of research on natural immunoregulatory processes and the prospects for developing innovative immunotherapies based on this knowledge. For example, weak or ineffective immune responses against tumor-specific antigens and pathogen-specific antigens allow tumors and chronic infections to persist in immunocompetent hosts. It seems reasonable that similar (if not identical) processes that regulate responses to these antigenic agents of disease may also play a normal physiologic role in modulating the immunogenicity of allogeneic tissue transplants placed at immune-privileged sites. Consequently, we juxtaposed reviews focused on classical immune-privileged sites with reviews on the interactions between tumors or pathogens and the immune system, as well as reviews on particular immunoregulatory mechanisms that might contribute to these processes. In the following chapter, Elizabeth Simpson 1 provides an illuminating historical perspective on immune privilege. She reminds us that allograft survival at immune-privileged sites, though prolonged, was rarely permanent. Moreover, technical issues created some ambiguities about how to interpret these early observations. Nevertheless, with the benefit of hindsight, Dr Simpson points out that this research provided a number of important clues to immunoregulatory mechanisms that became (and still are) the focus of attention in immunological research. In chapters following, Rachel Caspi 2, Joan Hunt 3, Monica Carson 4, and Andreas Meinhardt 5 and their colleagues review the current status of immune privilege as applied to the eye, maternal–fetal interface, central nervous system (CNS), and gonads, respectively. With the once popular hamster cheek pouch, which has now fallen into neglect, these anatomical sites were all identified as sites of immune privilege in the original pioneering studies discussed by Dr Simpson. Reviews provided by Nick Crispe 6, Cathryn Nagler 7, and Georg Kraal 8 and their colleagues focus on the liver and mucosal surfaces of the gastrointestinal (GI) tract and their associated lymphoid tissues. From these reviews, it is clear that these tissues also display certain characteristics of immune-privileged organs, even though they were not included on the original list of immune-privileged sites. Indeed, the remarkable tendency of liver allografts to protect themselves from host adaptive immunity even affords protection to other organ transplants in clinical situations. Thus, the liver appears capable of establishing systemic immune privilege that extends to other transplanted tissues. A similar trait emerges from studies on the mucosal surfaces of the GI tract and its associated lymphoid tissues. Under normal circumstances, innocuous airborne and food antigenic agents, as well as commensal microorganisms resident in the GI tract, do not elicit immune responses that cause extensive pathology to the mucosal surfaces, despite the fact that a plethora of these 'foreign' antigenic substances are encountered by naive immune cells. Indeed, many allergic diseases and inflammatory disorders of the GI tract may be consequences of chronic immune sensitization and loss of normal regulation of immunity directed against innocuous agents encountered at these mucosal surfaces. The mucosal surfaces of the lung airways and the GI tract exhibit immunoregulatory qualities reminiscent of the immune-privileged status exhibited by the uterus (another mucosal surface) during pregnancy, although the underlying mechanisms that control mucosal immunity and maternal immunity directed against paternally inherited fetal alloantigens may differ. The tendency to provoke tolerogenic rather than immunogenic responses to antigenic stimuli delivered at mucosal surfaces challenges the fundamental tenet of self- versus non-self discrimination in the adaptive immune system as the basis for responses leading to tolerance or immunity, respectively. Clearly, the anatomic context in which foreign antigens are encountered for the first time is a critical factor in shaping the immune responses that are elicited. Until recently, the notion that tumors and pathogens might manipulate adaptive immunity by enlisting immunoregulatory mechanisms was not considered a critical factor in the etiology of cancer and persistent chronic infections, even though these agents of disease express foreign and neoantigens yet still manage to persist for long periods in immunocompetent hosts. Again, these situations are paradoxical and inconsistent with a simplistic interpretation of the self-/non-self hypothesis. Reviews by Thomas Gajewski 9 and David Munn 10 and their colleagues on tumors, and David Sacks and 11 Yuri Persidsky 12 and their colleagues on infectious pathogens, which cause chronic (and persistent) infections in immunocompetent hosts, summarize evidence that tumors and pathogens actively suppress host immunity and promote tolerance. In cancer patients, as in autochthonous tumor models in mice, successful primary tumor development and later metastases to secondary sites depend not only on genetic modifications that release cells from the constraints of normal growth control but also on negotiations between tumor cells expressing neoantigens and the host immune system. The key problem is that these events occur in the preclinical phase of primary tumor growth, and immunotherapies designed to provoke anti-tumor immunity may be ineffective (or fail completely) if innate tolerance established in the early phase of tumor growth is potent enough to suppress attempts to clinically induce immune responses to tumor antigens. Tumors that succeed in evading innate immune surveillance mechanisms are likely to do so by exploiting natural processes of peripheral self-tolerance induction. Emerging evidence suggests that analogous considerations may also apply to pathogens that cause chronic infections when the immune system fails to sterilize the infected immunocompetent host, leading to persistence of pathogens. Like tumors, T cells that recognize antigens expressed by chronic pathogenic organisms, fail to differentiate into competent effector T cells that protect the host and instead become anergic or die, while the targets they could destroy persist and cause life-threatening disease. This implicit immunological paradox features in an increasing body of contemporary immunological research, with good reason, since it is becoming increasingly clear that effective immunotherapies in the clinic will require improved understanding of these inhibitory pathways that allow tumors and chronic infections to persist. In effect, the task is to break the state of pathologic tolerance or immune privilege established by these disorders. As is apparent from Dr Simpson's review onward, informative parallels emerge when the concept of classical immune privilege is linked with research on immunoregulation. Contemporary understanding of molecular and cellular mechanisms underlying immune privilege embraces many cutting-edge developments driving current immunological research. To this end, we invited Wolfgang Weninger 13, Sharon Wahl and their colleagues14, and Thomas Ferguson 15 and their colleagues to review immune cell trafficking, transforming growth factor-β (TGF-β), and Fas/Fas ligand, respectively, as key immunoregulatory mechanisms that may help confer privilege locally. In the final chapter of this volume, we return to allografts, but with a therapeutic twist. Stephen Cobbold, Herman Waldman 16 and their colleagues review prospects for protecting tissue allografts by manipulating host immune responses using immunomodulatory reagents. In clinical transplantation, the goal is to create immune privilege via immune intervention by uncoupling the immunogenicity of transplanted donor tissues from their antigenicity. Preferably, this objective should be realized with minimal recourse to globally immunosuppressive drugs, i.e. it should be local to the transplant as well as specific for donor alloantigens. Hence, in the spirit of this issue of Immunological Reviews, clinical efforts to protect tissue transplants from rejection might be viewed as attempts to induce immune privilege by artificial immunotherapeutic interventions. As guest editors, we thank our colleagues and their coauthors who accepted our challenge to think 'outside the box' by relating their individual insights and perspectives to the concept of immune privilege. In the remainder of this opening chapter, we discuss seven themes that recur in the reviews in this volume. Identifying these recurrent themes may help elucidate our understanding of how local immune privilege works, i.e. how complex immunoregulatory processes determine the appropriate local immune response to antigenic stimuli and how these responses can become dysregulated or subverted in the course of disease. Extending the classical concept of immune privilege beyond the traditional areas of tissue allografts has prompted some interesting ideas from the contributors to this issue. Indeed, one contributor concluded that it might be advisable to reserve the notion of immune privilege exclusively for transplanted tissues, on the grounds that a descriptive term applied too widely may lose its specific meaning and impact. While this point is well taken, we think that revisiting the concept of immune privilege–or whatever term is chosen–has value because it can reshape the way we think about certain disease processes, particularly those in which the immune system is actively suppressed. According to a strict interpretation of the self-/non-self discrimination hypothesis, antigenic tumors and certain chronic infections are not playing by the rules and demand further explanation. The paradoxical absence of effective immune responses in these settings emphasizes the need to focus on the context in which antigens are first encountered by the peripheral immune system. These observations parallel the consequences of introducing antigens via classical immune-privileged sites and resemble peripheral tolerance mechanisms, which ensure peripheral self-antigens not expressed in primary lymphoid tissues do not provoke autoimmunity under normal (homeostatic) conditions. Immune privilege in the intact eye arises in part due to physiologic barriers that segregate eye antigens from immune surveillance by inhibiting cellular trafficking needed to elaborate the afferent and efferent arms of cellular immune responses 2, 15. However, eye immune privilege is maintained if these barriers are breached via processes that deliver antigens from the anterior chamber of the eye to the spleen in murine models (anterior-chamber-induced immune deviation). Hence, even though immune privilege manifests locally and may involve physical and biochemical barriers in specific organs and tissues, maintenance of immune privilege also depends on processes that act systemically. Similarly, the blood–brain barrier and absence of lymphatic drainage from the CNS has been construed as evidence of CNS isolation from the immune system. Nevertheless, the CNS and the immune system have evolved subtle and complex ways to interact that drive systemic immune awareness of CNS antigens, while actively suppressing effective cellular immune responses to CNS antigens under normal circumstances 4, 12. Likewise, tumors appear to create systemic tolerance to antigens presented in the tumor microenvironment and in upstream draining lymph nodes 9, 10. Pregnancy in female mice also induces a transient state of partial systemic tolerance to paternally inherited fetal alloantigens 17, and testicular and ovarian self-antigens continuously drive tolerogenic processes involving regulatory T cells (Tregs) specific for self-antigens expressed exclusively in immune-privileged gonadal tissues 18. During fetal gestation, systemic immunosuppression may result from localized induction of specific molecular and cellular mechanisms, such as human leukocyte antigen HLA-G 3, indoleamine 2,3 dioxygenase (IDO) expression 19, negative costimulatory molecules 20, or Tregs 21, 22, at the maternal–fetal interface or in regional lymphoid tissues. These observations reveal that immune privilege often involves active transfer of self-antigens from local (privileged) sites to secondary lymphoid tissues and that antigen transfer continuously moderates immune responses to antigen, even if the same antigen is delivered subsequently via other (non-privileged) sites. Analogous considerations apply to oral tolerance induction, when primary antigen encounter in the GI tract leads to systemic antigen-specific (oral) tolerance via processes that take place in regional lymphoid tissues 7, 8. The original notion of physical barriers that segregate immunogenic tissues from immune effector cells is no longer considered sufficient to explain an immune-privileged state, even when combined with the more modern notion of 'passive ignorance', in which segregated circulations prevent encounters that could activate antigen-specific T cells. The contemporary paradigm has moved decisively toward active immunoregulatory mechanisms, usually involving several functioning in parallel, as the explanation for weak immune responses elicited by antigenic stimuli. In summary, no single mechanism accounts for immune privilege. This conclusion has potentially important implications for clinical immunotherapy. Strategies to break tolerance will likely have to inhibit more than one mechanism, and strategies to induce tolerance will need to enlist mutually reinforcing pathways. Raising a pertinent example from our own research, acutely depriving pregnant mice of the immunoregulatory enzyme IDO by pharmacologic inhibition was sufficient to allow maternal T cells to reject allogeneic fetuses; yet, chronic absence of this pathway (by ablating the IDO gene) had no effect on pregnancy, although IDO inhibitor no longer possessed abortifacient activity in IDO-ablated pregnancies 23-25. These observations suggest that subordinate T-cell regulatory mechanisms, redundant in normal mice, could readily substitute for lack of IDO in pregnancy, even though these mechanisms only moved to the fore when mice were genetically deficient for IDO expression. Conversely, some molecules, such as TGF-β, cytotoxic T-lymphocyte antigen-4 (CTLA-4), and forkhead box protein 3 (Fox P3), appear indispensable for normal T-cell regulation, since mice with ablations in these genes die in utero, or soon after birth due to spontaneous autoimmunity 26-28. However, these molecules may have multiple effects on distinct subsets of immune cells. This possibility raises concerns that clinical interventions targeting these indispensable mechanisms might create undesirable side effects, such as autoimmunity; indeed, anti-CTLA-4 antibody therapy to induce anti-tumor activity in cancer patients sometimes results in autoimmunity 27. In part, the paradigm shift toward active suppression can be traced to recent recognition of Tregs as potent and obligatory enforcers of tolerance. Not surprisingly, Tregs are a recurring theme in the reviews in this volume. Indeed, now that investigators know how to detect Tregs, they are turning up in many settings where immunoregulation manifests, either at classically defined immune-privileged sites or in other (non-transplant) settings where tolerance is induced, such as mucosal surfaces, pregnancy, tumors, and some chronic infections. The activity of Tregs at sites of immune privilege may obviate the need for physical segregation of antigen and responder cells. However, demonstrating active suppression by Tregs in vivo remains a formidable challenge. Tregs are now implicated in an array of immunologic systems. Though originally described as cells responsible for enforcing self-tolerance to peripheral autoantigens 29, Tregs can also suppress T-cell responses to tumor antigens 30, fetal antigens during pregnancy 21, mucosal antigens 31, and even infectious pathogens, including HIV and Leishmania32, 33. Hence, Tregs provide a unifying cellular mechanism that can explain weak or absent T-cell responses to antigenic challenges. However, the origin, antigen specificity, homing functions, and suppressor mechanisms of Tregs are still subject to intense research. The contribution of different types of antigen-presenting cells (APCs) to immune privilege is shrouded in even more mystery than the role of Tregs, despite substantive recent progress in defining APC subsets and the cellular and molecular processes they use to acquire, process, and present antigens to effector immune cells 19, 34-37. One prominent model is that mature APCs stimulate effective T-cell responses, while immature APCs create abortive T-cell responses or tolerance. This model is supported by the observation that targeted delivery of antigen to dendritic cells (DCs) in the absence of maturation signals leads to abortive T-cell responses and subsequent unresponsiveness and suppression of autoimmune disease progression, while maturation stimuli delivered simultaneously induced T-cell stimulatory outcomes 38, 39. However, it is not clear how immature DCs might actively create suppression in an antigen-specific manner. Recent evidence suggests that certain APC subsets may actively suppress immune responses. Some macrophage subsets (M2) possess anti-inflammatory functions that promote tissue repair and suppress responses to bacterial stimuli 40, 41. Similarly, specific subsets of plasmacytoid DCs (pDCs) that accumulate in the tumor microenvironment and in tumor-draining lymph nodes (TDLNs) of cancer patients and tumor-bearing mice may suppress tumor-specific T-cell responses 9, 10. Indeed, accumulation of IDO+ pDCs in sentinel lymph nodes of melanoma patients correlated with worse clinical outcomes, consistent with a role for IDO+ pDCs in helping to maintain the privileged status of tumors and associated lymphoid tissues. An analogous role for T-cell regulatory DC subsets emerged from studies on the mechanism of oral tolerance induction 42, 43. Some DC subsets in gut-associated lymphoid tissues appear to stimulate Tregs preferentially. This finding implies that DCs and Tregs may collaborate to amplify immunoregulatory outcomes. This collaboration requires more study, but in our own research, the potent and dominant T-cell suppressive effects of tiny populations of splenic and TDLN pDCs expressing IDO are striking and may well involve collaborations with Tregs 19. Indeed, a recent report shows that specific DC subsets preferentially naive T cells to differentiate into Tregs in an In summary, these observations suggest that it may be critical to which APC subsets present antigens to T cells. The of even a of APCs with potent and dominant immunoregulatory functions may suppress the functions of a of other unresponsiveness to antigenic stimuli manifests in tissues not included in the original list of immune-privileged sites. In the of the its from the original list may a tissue that was As described by Crispe later in this the liver has a remarkable ability to protect from immune responses following transplantation, that it can induce an immune-privileged state following allografts also induce systemic immunoregulatory effects that of other is also an to place tissue allografts, as by the pioneering of in who transplanted allografts into the of patients with 1 using minimal immunosuppression via Tumors and chronic infections create local sites of immune suppression and areas for T cells, in which the for long the that an immune-privileged at when infected by the which from this Thus, local sites of immune privilege can be established even in tissues such as which has been viewed as a immunogenic tissue from a transplant provided that the are created for initial antigen As a final we the therapeutic implications of natural immunoregulatory mechanisms that contribute to immune privilege. is for developing innovative immunotherapies to either or suppress adaptive immunity for clinical targeted have to clinical benefit for patients with autoimmune diseases and allografts In tissue and of autoimmune diseases, the in effect, is to create an artificial immune-privileged state that a organ under or autoimmune Conversely, in cancer a number of to immunity to tumor antigens are under the goal is to break the induced immune-privileged status that tumors to immune effector The same may be to certain chronic infections that display features of immune suppression and privilege. As we have a from the wider perspective by the reviews presented in this issue. that will also these perspectives useful to help stimulate innovative conceptual to immunoregulatory mechanisms and provide insights into ways to clinical outcomes following as