| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Durham VA Medical Center and Duke University Medical Center, Durham, North Carolina
Correspondence and requests for reprints should be addressed to David S. Pisetsky, M.D., Ph.D., Durham VA Medical Center, Box 151G, 508 Fulton St., Durham, NC 27705. E-mail: dpiset{at}acpub.duke.edu
ABSTRACT
Nuclear macromolecules, in addition to their intracellular role in regulating cell function, can translocate into the extracellular space where they can activate innate immunity. This translocation can occur in various settings and reflects the dynamic nature of nuclear structure. Of nuclear molecules, DNA and the DNA-binding protein, HMGB1, display distinct patterns of immune activity. For DNA, immune activity depends on sequence, base methylation, and context. While bacterial DNA is an immune activator, mammalian DNA is either inert or inhibitory when free. In contrast, mammalian DNA in the form of immune complexes can trigger immune cell activation. As shown in in vivo and in vitro studies, DNA can exit cells during apoptotic as well as necrotic cell death in a process that may depend on the presence of macrophages. Like DNA, HMGB1 can exit cells and acquire immune properties. For HMGB1, the translocation occurs in macrophages that have been stimulated by Toll-like receptor (TLR) ligands as well as cytokines; HMGB1 release can also occur with apoptotic as well as necrotic death. While HMGB1 alone can display cytokine activity, it may also activate cells in conjunction with other immune stimulators such as TLR ligands. For both DNA and HMGB1, the immune properties may therefore reflect the array of other endogenous as well as exogenous molecules present.
Key Words: innate immunity • DNA • HMGB1 • apoptosis • necrosis
The nucleus is the central organelle in the cell because it is the location for replication, transcription, and the regulation of gene expression. While often viewed as uniform in composition, the nucleus is a dynamic structure with a regional anatomy that can vary during the cell cycle. Furthermore, the nucleus can serve as a repository for molecules that transit into and out of the cytoplasm and even exit the cell entirely. Remarkably, once in the extracellular space, some of these nuclear molecules can serve as immune mediators and trigger the innate immune system in settings of injury and death. As such, the translocation of nuclear macromolecules from the inside to the outside of the cell is a central event in innate immunity.
Among endogenous molecules stimulating innate immunity, two nuclear molecules have attracted the most attention. Thus, both DNA and the DNA-binding protein, HMGB1, can exit cells during death processes and, in the extracellular space, stimulate responses via a variety of receptors to signal danger. While both DNA and HMGB1 reside primarily in the nucleus and, indeed, may interact, they differ in intranuclear and intracellular mobility as well as physical state during death processes. This review will therefore consider the roles of DNA and HMGB1 in innate immunity and propose an integrated picture for their movement and function during death processes.
THE IMMUNE PROPERTIES OF DNA IN SLE
The recognition of DNA's immune properties dates back half a century to its discovery as a target antigen in systemic lupus erythematosus (SLE). This prototypic autoimmune disease is characterized by the production of antinuclear antibodies, with antibodies to DNA (anti-DNA) serving as markers of diagnostic and prognostic significance. As shown in studies on patients with SLE as well as murine models, these antibodies can also mediate tissue injury, with DNA–anti-DNA immune complexes important inducers of glomerulonephritis (1–3).
As an antigen in SLE, DNA most likely functions in the form of the nucleosome, a highly organized structure comprised of DNA wrapped around a histone core. Consistent with the role of nucleosomes as the driving antigen for autoantibody induction, sera from patients with lupus contain antibodies to nucleosomes, the DNA and histone components, and other structures comprised of DNA and histones. In this conceptualization, DNA behaves as an epitope or surface of a larger structure that impacts on the immune system; the use of the term DNA does not imply that this molecule is free in solution and devoid of other nucleosomal components (1, 3).
Because of the importance of DNA as antigen in SLE, efforts to replicate this disease in induced animal models focused primarily on immunization with DNA. With a few possible exceptions, these experiments failed, with DNA, even if attached to a carrier and presented in adjuvant, unable to induce an appreciable autoantibody response. In its immunologic inertness, DNA appeared fundamentally different from proteins and carbohydrates, the other major classes of macromolecule antigens (3).
The failure of DNA to induce a specific autoimmune response contrasts sharply with the situation of other autoimmune diseases, such as rheumatoid arthritis or multiple sclerosis, where animal disease models can be created in animals by immunization with self antigens such as collagen or myelin basic protein (4). Together, these considerations suggested that SLE reflects a fundamental disturbance in the immune system that allows a response to an essentially immunologic blank molecule. The alternative explanation for the anti-DNA responses in SLE is the existence of a more immunogenic form of DNA that could induce responses under conditions in which experimental DNA–protein complexes fail.
THE INDUCTION OF IMMUNE RESPONSES TO DNA
The mystery of DNA's immune activity in SLE is now clearing because of studies in two seemingly disparate areas that have converged to provide a picture of the triggering of innate immunity by nuclear molecules. The first area of this research concerns the immune properties of bacterial DNA or, as it is often called, CpG DNA. As shown in studies conducted over the last 20 years, bacterial DNA, unlike mammalian DNA, is immunologically active and can induce cytokine production and B cell mitogenesis. As demonstrated in elegant molecular studies, these responses result from stimulation of the TLR9 receptor which, in contrast to some other Toll-like receptors (TLRs), resides on the inside of cells as opposed to the membrane (5, 6).
The induction of immune responses by bacterial DNA reflects structural microheterogeneity and its content of sequence motifs that center on unmethylated CpG dinucleotides. These sequences occur much more commonly in bacterial than mammalian DNA because of differences in the patterns of base methylation as well as a phenomenon known as CpG suppression. Because of the differential display of the CpG motifs, bacterial DNA, like LPS, can act as a pathogen-associated molecular pattern (PAMP). Importantly, in the studies on the structure–function relationships of immune stimulation by DNA, intact mammalian DNA has been consistently inactive in in vitro or in vivo systems. These observations point to the exquisite specificity for base recognition in the triggering of innate immunity via TLR9.
The demonstration of the intrinsic immune activity of bacterial DNA is notable since it joins DNA with proteins and carbohydrates in the family of immune active macromolecules. Studies on the antibody response to mammalian and bacterial DNA in humans confirmed and extended this role. As this serologic analysis showed, the sera of normal human subjects (NHS) have significant levels of antibody directed to bacterial DNA. These antibodies bind with high avidity and specificity to some, but not all, bacterial DNA antigens. While the antibodies in NHS differ in isotype from those in SLE, their presence nevertheless suggests that, during the ordinary encounter with bacteria, bacterial DNA triggers responses and can drive responses to sequential DNA epitopes (7–9).
Together, these observations raise the possibility that bacterial DNA can initiate or sustain anti-DNA production in SLE because of its intrinsic immunologic activity as well as display of the double-stranded B DNA antigen conformation. According to this model, the lesion in SLE would reflect antigen recognition, not responsiveness, with abnormalities in the immune repertoire in SLE providing an array of B cell precursors that can be triggered by bacterial DNA antigen to produce autoantibodies (10). This possibility is supported by studies showing that bacterial DNA can induce an autoantibody response by immunization of pre-autoimmune NZB/NZW mice under conditions in which mammalian DNA is inactive (11). While supporting the ability of bacterial DNA to stimulate responses, these studies attest to the paucity of immunologic activity in mammalian DNA.
THE ACTIVITY OF IMMUNE COMPLEXES
The second line of research on the immune activity DNA developed originally in efforts to characterize a factor in the sera of patients with lupus that could stimulate the in vitro production of interferon-
(IFN-). As shown in seminal experiments, this factor is an immune complex comprised of DNA and anti-DNA. While either component alone is devoid of activity, the complex can potently stimulate IFN- production by plasmacytoid dendritic cells. Furthermore, this stimulatory activity can be mimicked by mixing antibody preparations with supernatants of apoptotic cells. Studies on the mechanisms of B cell activation in a rheumatoid factor transgenic system in the mouse also showed that immune complexes with DNA have activity not present with DNA alone (12–14).
Subsequent studies on the stimulation by plasmacytoid DCs by DNA immune complexes established a role for both Fc receptors and TLR9 receptor; non–TLR receptor mechanisms may also operate, however. Since TLR9 resides on the inside of the cells in an endosomal compartment, immune complexes may allow DNA internalization, where interaction with TLR9 may occur. The interaction with the Fc receptor could lead to additional signaling or provide an alternative mechanism for DNA entry into the cell (15, 16). While the identity of the DNA in immune complexes from sera has not been characterized extensively, endogenous human DNA is the likely source. In this regard, immune complexes with ribonucleoproteins can also stimulate IFN-, perhaps related to their content of RNA molecules which can trigger TLR3 or TLR7 (14).
The activity of DNA in immune complexes contrasts with the activity of free mammalian DNA which is either inert or inhibitory. As shown in vitro studies, free mammalian DNA can inhibit the activity of bacterial DNA, raising the possibility that high concentrations of extracellular DNA can attenuate stimulation by bacterial DNA and possibly down-regulate innate immunity. This inhibitory activity can be termed "safety" in contrast to "danger" elicited by TLR ligands (17). Thus, as these experiments indicate, the formation of immune complexes can radically transform the activity of mammalian DNA to allow stimulation, rather than inhibition, of innate immunity.
Since even high concentrations of mammalian DNA are inactive, it appears unlikely that the role of the complex is solely to deliver DNA to TLR9. Even if mammalian DNA contains some CpG motifs, the presence of nonstimulatory or inhibitory motifs appears predominant. An interaction of active motifs within mammalian DNA may nevertheless cause activation of TLR9, although the triggering of other internal receptors appears more plausible. The operation of such receptors can be inferred from the activity of mammalian DNA in complexes with transfection reagents known as cytofectins (18). These reagents promote DNA internalization and, while the mode of action of cytofectins and immune complexes may differ, the potency of DNA in cytofectin complexes provides further evidence that the context of DNA determines its activity on innate immunity (18–20).
The immunostimulatory activity of DNA-containing immune complexes has attracted great interest as a mechanism for promoting nonspecific immune activation in the pathogenesis of SLE. While such complexes can sustain or intensify immune activation, they do not explain the initial generation of the anti-DNA antibody which is critical to the activation by DNA. Furthermore, the necessity for complex formation suggests that for endogenous DNA to activate innate immunity, it must exist in association with another moiety or structure to allow access to internal DNA receptors. These considerations (Table 1) thus focus attention on the mechanisms by which DNA is released from cells and the molecules to which it is attached during this process.
|
As shown using a variety of assays, DNA appears at high levels in the blood in many conditions including SLE, pulmonary infarction, malignancy, and trauma among others. These conditions are all associated with cell activation or cell death, although death has been considered the most likely source of blood DNA (21, 22). As now conceptualized, cell death can be divided into mechanistically distinct processes, with apoptosis and necrosis representing the prototype pathways. Apoptosis, or programmed cell death, is a highly regulated process characterized by cellular collapse and cleavage of nuclear molecules. In contrast, necrosis is a random process, provoked by physical or chemical trauma that culminates in cell lysis (23). According to current paradigms, necrosis is pro-inflammatory while apoptosis is anti-inflammatory, perhaps related to release or spillage of internal molecules.
To investigate nuclear molecule dynamics during death, our laboratory has used in vitro and in vivo system to assess the amount and properties of extracellular DNA released from dying cells. In addition, we have investigated the influence of macrophages on the release of DNA, since macrophages can scavenge dead cells by phagocytosis and therefore potentially modify the amount of DNA arising from such cells. In these experiments, we have measured DNA by both an ELISA for nucleosomes as well as a direct chemical measurement using the dye PicoGreen. This dye binds specifically to double-stranded DNA and allows sensitive detection by a fluorometric assay.
Results of these studies have provided a novel perspective on DNA release during cell death as well as highlighted differences in the in vitro and in vivo settings. Thus, during in vitro culture, Jurkat leukemia T cells made apoptotic by chemical agents release DNA in a time-dependent process. This extracellular DNA shows laddering by gel electrophoresis, indicating cleavage into nucleosomal fragments. In contrast to cells undergoing apoptosis, cells made necrotic by treatment with heat or ethanol fail to release DNA even after prolonged culture and despite changes in cellular permeability, as demonstrated by propidium iodide staining (24).
As our experiments showed, in the in vitro setting, macrophages can display important but divergent effects on DNA release. With apoptotic cells, the presence of macrophages (either the RAW264.7 cell line or murine bone marrow–derived macrophages) can reduce the amount of DNA released, whereas macrophages can increase the amount of DNA released by necrotic cells. For necrotic cells co-cultured with macrophages, the DNA present in the medium showed laddering suggestive of nuclease digestion. While indicating the importance of macrophages in mediating DNA release, these findings also showed that the size of the DNA in the extracellular space is not itself a measure of whether death occurs by apoptosis or necrosis (25).
In vivo systems can also help dissect the DNA release process. Thus, in normal mice, the intraperitoneal administration of apoptotic or necrotic Jurkat cells leads to a prompt rise in the amount of DNA in the blood that returns to baseline after 24 h. With both types of dying cells, DNA shows laddering with a similar size distribution by gel electrophoresis. This result is similar to that obtained in vitro using mixed cultures of Jurkat cells and macrophages. Other studies with this model showed that the extent of DNA release from transferred dead cells can be modified by treatment of recipient mice with clodronate (which eliminates macrophages), dexamethasone, or the induction of peritoneal exudates (26–28).
Together, these studies provide evidence that macromolecule release from dead cells is a modifiable process that may depend on the presence of more than one cell population, especially macrophages (Table 2). Furthermore, they suggest that, to the extent that extracellular DNA can be a player in the innate immune system, the distinction between apoptosis and necrosis may not be as stark as sometimes portrayed. Thus, in the intact animal or in mixed cell cultures, both apoptotic and necrotic cells can release DNA that is similar in properties as measured by size distribution. These considerations do not imply that the extracellular DNA from apoptotic and necrotic cells are equivalent immunologically, since the immune activity of DNA depends on context and the presence of other molecules that may affect its intracellular trafficking. It is important therefore to consider other molecules that may be released in death settings.
|
HMGB1 is a nonhistone nuclear protein that has also been implicated in the stimulation of innate immunity during death processes. Structurally, HMGB1 is 214 amino acids long and can be divided into an A box, B box, and a C-terminal tail domain, with the A and B boxes responsible for DNA interaction. This protein can bind DNA in a non–sequence-specific manner to bend DNA, although it binds preferentially to distorted DNA structures such as strand junctions. In addition to binding DNA, HMGB1 can interact with nuclear proteins. While the precise function of HMGB1 is not known, this protein can regulate transcription, perhaps by altering the architecture of DNA to promote interactions with other factors. Depending on cell type, HMGB1 can also appear in the cytoplasm (29, 30).
Like DNA, HMGB1 can leave the cell where, in the extracellular space, it can serve as an alarmin and promote immune activation. An alarmin is an intracellular molecule that, when released from cells during death processes, can stimulate innate immune processes (31). In another terminology, an alarmin is a DAMP or damage (or death)-associated molecular pattern. In the case of HMGB1, the release processes was first detected in the setting of macrophage activation rather than death. In these experiments, supernatants of cultures stimulated with LPS were characterized to identify other molecules that could mediate septic shock and therefore represent novel targets of therapy. As these experiments showed, HMGB1 appears in high concentrations in the medium of macrophages activated by LPS as well as pro-inflammatory mediators such as TNF-, IFN-, and nitric oxide. Furthermore, HMGB1 levels are high in the serum of patients with shock and antibodies to HMGB1 can block shock in mice (32). A series of insightful experiments established a broad range of immunostimulatory activities of HMGB1, consistent with its role as a late mediator of LPS and its function as a cytokine. This stimulation appears dependent on the RAGE receptor, although activation of TLR2 and TLR4 may also contribute to stimulation (33–37).
As shown subsequently, death processes, in addition to macrophage activation, can lead to HMGB1 release, with this protein posited as an important mediator of inflammation induced by necrotic cells. As shown in vitro, cells induced to undergo necrosis by freeze-thawing release large concentrations of HMGB1. This release process differs from that occurring during activation, where acetylation and phosphorylation alter the charge of HMGB1. These modifications affect the intracellular trafficking of HMGB1, with the modified molecule transiting to the cytoplasm into vesicles for secretion (38, 39). In contrast, for HMGB1 release during necrosis, the process appears passive and results from the diffusion of this protein away from chromatin and out of the cell. Reflecting its nuclear function, HMGB1 is only weakly adherent to chromatin, differing markedly from histones in the strength of its DNA interaction. Thus, unanchored to the nucleus, HMGB1 can leave readily when the permeability barriers break down during death (38).
While the release of HMGB1 during necrosis is consistent with its biophysical properties, its behavior during apoptosis has been more uncertain. Studies by Scaffidi and coworkers indicated that apoptotic cells do not release HMGB1, with studies by FLIP (fluorescence loss of photobleaching) in fact demonstrating that its nuclear diffusion decreases dramatically during apoptosis, implying greater adherence to chromatin (40). Since apoptotic cells are generally considered anti-inflammatory, a retention of HMGB1 in the nucleus could limit immune activation during this death process. These results are notable since they suggest that, while apoptotic cells release DNA, they do not release a companion DNA-binding protein (i.e., HMGB1).
To resolve this seeming contradiction, we reevaluated HMGB1 release during death processes, using Jurkat cells as models. In these experiments, apoptosis was induced by chemical agents and HMGB1 release measured by Western blotting. Confocal microscopy with staining with anti-HMGB1 was used to confirm translocation. Together, these experiments showed clearly that, in apoptosis as well as necrosis, translocation of HMGB1 can occur, with the nucleus of cells showing dramatically reduced HMGB1 content in association with nuclear condensation (41).
Since DNA release occurs during apoptosis, HMGB1 release is not unexpected and suggests that permeability changes of nuclear and cytoplasmic membranes allow the diffusion of a variety of macromolecules into the extracellular space. The reasons cells differ in their behavior during apoptosis is unknown, although this process is not uniform among cells and may be influenced by both the cell type as well as inducing agent. Furthermore, comparisons between the amount of HMGB1 released during apoptosis and necrosis can be misleading, since the extent of macromolecule release during necrosis may also vary depending upon the inducing agent or treatment. Thus, in our experience, the extent of HMGB1 released from cells undergoing freezing and thawing is much greater than cells that have been treated with ethanol or heat. The most relevant system for inducing necrosis is speculative, since necrosis is a random and unregulated process that may follow a host of damaging agents. Table 3 summarizes the immune properties of HMGB1.
|
While original studies suggested that HMGB1 has direct cytokine activity, other studies have shown a more complicated situation. Thus, studies with purified mammalian HMGB1 show much lower activity than molecularly cloned material, which may have contamination by endotoxin or DNA. These considerations suggest that HMGB1 may not be fully active as a cytokine but rather may act in concert with other endogenous or exogenous molecules to trigger innate immunity (42). In this conceptualization, HMGB1 may tune the system, serving a thermostatic function to allow other molecules to trigger responses. Furthermore, HMGB1, like antibody or cytofectin, may serve as a carrier for DNA to allow triggering of internal DNA receptors. Indeed, HMGB1 can enhance DNA transfection, suggesting its ability to shuttle nucleic acid into cells (43).
Together, these considerations suggest that immune activity of nuclear molecules may result from the ensemble of species present, their physical interaction, and their ability to trigger more than one receptor type simultaneously. Whether DNA promotes HMGB1 stimulation or vice versa is a matter of semantics, as both components may have to be present simultaneously in a complex to induce cell activation. The rules by which complexes stimulate immunity are not known, although stimulation by organelle fragments or particulates may be more analogous to a cell–cell interaction (where multiple molecular interactions occur) than the stimulation by a cytokine, a hormone, or other small molecular mediator (e.g., prostanoid), which may act in a unitary manner.
While these issues require much further investigation, the studies on DNA and HMGB1 nevertheless indicate the diversity of intracellular molecules with immune activity and the importance of context in the stimulation of innate immunity by nuclear macromolecules. Future studies will track the dynamics of nuclear molecule trafficking during death processes, the role of macrophages in determining the extracellular release, and the signaling pathways stimulated by the complex mixtures of large molecules. Hopefully, this research will elucidate the pathogenesis of SLE and other autoimmune diseases as well as provide new approaches to treat the broad range of immune-mediated diseases.
FOOTNOTES
Supported by VA Medical Research Service and the Lupus Research Institute.
Conflict of Interest Statement: D.S.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript.
(Received in original form January 23, 2007; accepted in final form March 1, 2007)
REFERENCES
inducer in systemic lupus erythematosus. J Immunol 1999;163:6306–6313.
RIIa is expressed on natural IFN--producing cells (plasmacytoid dendritic cells) and is required for the IFN- production induced by apoptotic cells combined with lupus IgG1. J Immunol 2000;165:3519–3526.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
