Death, danger, and immunity: an infernal trio

The question as to whether cell death is immunogenic, tolerogenic, or silent with regard to immune recognition of dying cells is (one of) the central problem(s) of immunology. Similarly, it is of paramount importance to decipher the exogenous and endogenous 'danger signals' that are able to trigger an immune response. In this context, we are confronted with a triangular, enigmatic, and asymmetric relationship between death, danger, and immunity that points to several yet-to-be elucidated conundrums. Under which particular circumstances is cell death perceived as a danger? Is it the presence or the type of stress that a cell undergoes before it dies that determines whether it becomes 'visible' to the immune system? Or is it the modality of cell death, the subroutine through which the cells ultimately disintegrate, that determines whether cellular demise is immunogenic? What are the molecular characteristics of danger signals linked to cell death? What are the principal differences between alien and endogenous danger signals? What are the sensors or receptors on innate or cognate immune effector cells that perceive danger? Is there a particular combination and spatiotemporal sequence of danger signals or cell death-associated events that is decoded by the immune system to trigger an adequate immune response? And, last but not least, does the immune system itself emit danger signals when it attacks and eventually kills virus-infected target cells or tumor cells? What, on the contrary, are the brakes that are designed to avoid an overreaction of the immune system? A cornucopia of physiological and pathological agents including anti-cancer chemotherapy and viruses can induce cell death. Under conditions of homeostasis in the healthy human adult, more than one million cells die per second; without that, this hecatomb would induce and manifest autoimmune disease. Thus, it appears plausible that 'normal' cell death is non-immunogenic (or even tolerogenic), perhaps because dying cells are eliminated in an efficient fashion, by non-professional and professional phagocytes, before end-stage apoptosis or secondary necrosis can develop. In contrast, after viral infection, cells whose transcription and translation apparatuses have been hijacked by the virus eventually die, often in a way that is immunogenic. Indeed, if the immune system was unable to distinguish between physiological and pathological cell death, all of us would inexorably succumb either to severe systemic autoimmune disease or to uncontrolled viral infection. As a result, it is still an enigma under which circumstances cellular demise can induce an innate or cognate immune response. It is a matter of debate whether apoptotic, necrotic, or senescent cells are intrinsically immunogenic or tolerogenic. Moreover, the exact molecular mechanisms making the difference between immunogenic and tolerogenic cell death are not yet elucidated. The putative immunogenicity of cell death as well as its therapeutic exploitation may have far-reaching implications for the pathophysiology and for the clinical management of cancer, viral infections, and autoimmune disease. This is the topic of the present compendium of articles, which deals with the following specific aspects of the 'infernal trio' death/danger/immunity: the molecular mechanisms determining the switch between immunogenic and non-immunogenic death; the phagocytic recognition of dying cells in the context of the immune response; the effects of dying cells on dendritic cell maturation and antigen presentation; the links between cell death, danger signals, and tumor immunology; the links between danger signals and antiviral immune responses; and finally, the relationship among cell death, danger signals, and autoimmune disease. As to the switch between immunogenic and non-immunogenic cell death, there is increasing evidence that modalities of cell death other than apoptosis are tightly regulated in molecular terms. Thus, the RIP1 kinase has emerged as a key regulator of necrotic cell death, in that its activation is often required for this type of lethal event to occur (1). It can be surmized that the activation of RIP1 kinase also impacts on the 'decision' between silent, non-inflammatory, and inflammatory cell death. Although apoptosis appears to be morphologically homogenous, recent evidence suggests that the acquisition of preapoptotic stress markers, such as the exposure of calreticulin on the cell surface, may have a profound impact on how the cells are handled by the immune system (2). Thus, biochemical heterogeneities in death-inducing signal transduction and damage pathways may impact on the recognition of dying cells by the immune system, presumably by influencing the phagocytosis of dying cells. In addition, the passive or active (regulated) release of HMGB1 from necrotic or apoptotic cells, respectively, may lay the grounds for the activation of antigen-presenting cells (3) and serve as a ligand of Toll-like receptor 4 (TLR4) to facilitate optimal antigen presentation by dendritic cells (4). Thus, it is possible that a combination of preapoptotic and cell disintegration-associated signals elicits productive immune responses against antigens expressed by dying cells. In this sense, damage-associated molecular patterns (DAMPs) may play a major role in eliciting immune responses (5). Cancer cells may or may not elicit an immune response early during oncogenesis. Although it is a matter of debate whether immunosurveillance impacts on the frequency of cancer development (6, 7), it appears that tumors can elicit an immune response, at least in some instances, and that the immune system 'edits' tumors in a way that the characteristics of tumors from immunocompetent animals and immunodeficient hosts are different. Tumor cell death, as it occurs spontaneously or as it is induced by anti-cancer chemotherapy, can be immunogenic, depending on the peculiar death-inducing agent (6). Moreover, it is possible (yet remains to be demonstrated) that somatic mutations linked to the notorious genomic instability of cancer cells yield novel ligands of pattern recognition receptors (such as TLRs), which would then induce a local sterile inflammation that can be either immunogenic or can favor tumor development in the context of so-called 'inflammatory cancers' (8). Productive anti-tumor immune responses may be elicited by manipulating different sets of immune effectors, in particular dendritic (9), T (10), NK (11), or NKT cells (12), all of which are prospective targets of novel immunopharmacological strategies for the stimulation of anti-tumor immune responses. Similarly, regulatory T cells may constitute a target for de-inhibiting immune responses (13). At a conceptual level, it may be particularly interesting to combine chemotherapies that induce immunogenic cell death with immunotherapies sensu stricto to achieve an optimal curative success. Antiviral immune responses involve the recognition of viral structures such as virus-specific nucleic acids or proteins, by a plethora of cell surface and intracellular receptors that, once activated, facilitate the immune responses including the production of interferons that interfere with viral replication. The specific pattern recognition receptors and the signal transduction pathways elicited by viral products in distinct immune effectors are being dissected in the mouse system to an ever-increasing level of refinement (14). The characterization of inherited defects in TLR3 or post-TLR signal transducers affecting humans has also yielded profound insights into the redundancy of the TLRs in antiviral responses as well as into the specific link between specific TLR defects and the genetic susceptibility to just one (or few) viral disease(s) (15). Autoimmune diseases can be triggered by the failure of macrophages to phagocytose apoptotic cells or apoptotic debris (16), presumably because the decomposing cellular material activates antigen-presenting cells other than macrophages such as dendritic cells. Another hypothetical mechanism accounting for sterile inflammatory disease (such as rheumatoid arthritis, systemic lupus erythematosus, psoriasis, etc.) may be the formation of neoligands of innate immune receptors as a result of germline mutations (8). Indeed, recent studies have revealed an inappropriate activation of TLR7, TLR8, and TLR9 in systemic lupus erythematosus and other autoimmune diseases, suggesting that the development of therapeutic antagonists of TLRs may provide a new strategy for inhibiting autoimmune diseases (17). Nonetheless, other strategies targeting inflammatory cytokines or enhancing the function of immunosuppressive T cells also warrant further exploration (13). It appears more than plausible that the present and future accumulation of knowledge will generate novel therapeutic options to activate or inactivate the trio death/danger/immunity at will. We anticipate that this kind of manipulation will unravel new opportunities to stimulate vital antiviral and anticancer immune responses and to block unwarranted inflammatory and autoimmune reactions.