Introduction: HMGB1 in inflammation and innate immunity

Authors


Ulf Andersson, Department of Women’s and Children’s Health, Karolinska Institutet, Karolinska University Hospital, S-171 76 Stockholm, Sweden.
(fax: +46851775562; e-mail: Ulf.Andersson@ki.se).

Introduction

The high-mobility group box-1 (HMGB1, also termed HMG1 and amphoterin) protein is the most abundant member of the HMG family of DNA-binding proteins. In addition to its nuclear expression, HMGB1 is released to the extracellular space upon cell damage and also as a result of active secretion. Active secretion of HMGB1 has been demonstrated in many cell types, such as monocytes, macrophages, dendritic cells, hepatocytes, endothelial cells, glial cells and neurons. In general, stimulation of cells for example by cytokines, lipopolysaccharide (LPS), hypoxia or even by cell–extracellular matrix contact enables HMGB1 export from the nucleus to the cytoplasm and the extracellular space. Receptor for advanced glycation end products (RAGE) and certain Toll-like receptors (TLRs) act as the major signalling membrane-bound receptors of HMGB1 at the cell surface and within the cell (reviewed in [1]).

High-mobility group box-1 has been identified as a key molecule linking tissue damage and stress to activation of innate immune mechanisms. It is included in the group of molecules known as damage-associated molecular patterns (DAMPs) or alarmins. HMGB1 is involved in activation of innate immune mechanisms in the context of viral and bacterial infections and in several diseases in which acute or chronic inflammation plays a role, such as stroke, epilepsy, atherosclerosis, cancer and autoimmune diseases (reviewed in [2]).

The 4th International HMGB1 symposium ‘Signals of Tissue Damage’, held in Helsinki, Finland, in June 2010, provided the opportunity for leading groups in the field to discuss the major advances in the understanding of the roles of HMGB1 in physiological and pathophysiological processes. The main themes included the functions of HMGB1 within and outside the cell and its membrane receptors, as well as its involvement in inflammatory diseases and in interactions between the nervous and immune systems. The current volume of the Journal of Internal Medicine includes three reviews by distinguished contributors to the meeting [3–5], highlighting recent advances in diseases in which innate immunity and HMGB1 play a key role. Preclinical trials targeting HMGB1 have been successful in several animal models of inflammatory diseases, and we expect that these studies will advance to clinical trials within the next few years.

HMGB proteins regulate cytosolic nucleic acid-mediated interferon (IFN) responses

Molecules of the HMGB family, including HMGB1, HMGB2 and HMGB3, have been studied for almost 40 years as nuclear components important for transcription and the chromatin structure. In addition, the extracellular role of HMGB1 as a mediator of inflammation and tissue regeneration has been extensively explored during the past decade. However, HMGB proteins also exist as cytosolic molecules, the functions of which remain unclear. In the current volume of the Journal of Internal Medicine, Tadatsugu Taniguchi, a pioneer in the type-I IFN field, and his colleagues review their novel work indicating that HMGB proteins act as universal intracellular sensors of cytosolic nucleic acids [3]. The absence of cytosolic HMGB proteins severely impairs type-I IFN and cytokine induction by DNA or RNA targeted to activate the cytosolic (DAI, RIG-I, MDA5, AIM2 and IFI204) and endosome-based (TLR3, TLR7 and TLR9) nucleic acid-sensing receptors [3, 6, 7]. Cytosolic DNA stimulation also induces the formation of the inflammasome in macrophages via activation of AIM2, triggering the secretion of interleukin (IL)-1β. The use of HMGB gene-deficient mice or pan-HMGB siRNA, quenching the expression of HMGB1, HMGB2 and HMGB3 in responder cells, selectively prevents nucleic acid-induced innate immunity. Accordingly, virus replication is increased in HMGB-deficient cells emphasizing that HMGB sensing of virus-derived nucleic acids is critical for antiviral innate immunity. The overall message is that immunogenic nucleic acids bind HMGBs (promiscuous sensing), and this is required for subsequent recognition by specific pattern-recognition receptors (discriminative sensing) to activate the innate immune responses.

By contrast, the induction of cytokines remains unaffected in control cultures with HMGB siRNA-treated cells stimulated by LPS. mRNA induction of the genes activated by a variety of cytokines also occurs normally in these cells, indicating that gene transcription in general is not affected by the absence of HMGBs. Future studies will be needed to clarify the mechanisms by which HMGB–nucleic acid complexes bind and activate their intracellular receptor signalling cascades. It appears that the HMGBs may function upstream of the cytosolic receptor signalling pathways after nucleic acid activation. Similar studies are currently being conducted by several groups to determine how extracellular complexes of HMGB1 and certain proinflammatory partner molecules including IL-1β or LPS may act in synergy.

The purpose of the HMGB-orchestrated surveillance system of cytosolic nucleic acids is discussed in the review from an evolutionary perspective. Eukaryotic cells must maintain their chromosomal integrity and sense invading foreign nucleic acids that are not degraded by nucleases. Cytosolic HMGBs bind nonself nucleic acids but may lack the capacity to eliminate nonself nucleic acids; hence later signalling mechanisms might have evolved to activate genes for cytokines including IFNs to protect the surrounding cells from invasion. During this evolutionary process, HMGBs have continued to serve as sentinels of cytosolic nucleic acids to activate nucleic acid-induced signalling pathways.

HMGB1 in the pathogenesis of systemic lupus erythematosus (SLE)

Systemic lupus erythematosus is a chronic autoimmune disease that involves systemic inflammation and organ damage preferentially affecting the skin and the renal, nervous, cardiovascular and haematopoietic systems. Both genetic and environmental components contribute to the development of SLE, and important pathogenetic factors include: increased cell death combined with impaired clearance of apoptotic cells; a failure to maintain immune tolerance to nuclear antigens enabling autoantibody formation against nuclear antigens with anti-double-stranded DNA (dsDNA) antibodies, which represent the serological hallmark of SLE; generation of immune complexes formed by nuclear antigens and autoantibodies that activate macrophages, dendritic cells and complement, causing tissue damage; and a dysregulated type-I IFN system mediating excessive IFN-α production and contributing to activation of adaptive immune reactions.

Recent research indicates that HMGB1 is involved in the pathogenesis of SLE and highlights the potential of this molecule both as a biomarker and a novel target of therapy [4, 8–14]. Flares of SLE are characterized by strong immune activation and extensive cell death, which both provoke extracellular HMGB1 release. The extracellular HMGB1 generates increased systemic levels that have been demonstrated to correlate with SLE disease activity. However, conventional HMGB1 assessment based on enzyme-linked immunosorbent assay (ELISA) provides false-negative results in SLE patients, because of excessive HMGB1 autoantibody formation in this disease [15, 16]. HMGB1 autoantibodies prevent the ELISA antibodies from recognizing serum/plasma HMGB1. However, these inherent problems are overcome in western blot analysis performed under reducing conditions. Whether the HMGB1 autoantibodies act in a protective or harmful way in SLE patients remains unresolved. Cells dying from primary necrosis or apoptotic cells undergoing secondary necrosis will release nuclear HMGB1. Inadequate removal and degradation of apoptotic cells are well-known features of SLE. HMGB1 accounts for reduced efferocytosis (i.e. reduced phagocytosis of apoptotic cells) in SLE in two different ways. First, extracellular HMGB1 decreases phagocytosis by binding to the ‘eat-me’ signal phoshatidylserine exposed on the surface of apoptotic bodies. Second, it also inhibits macrophage activity by blocking αVβ3-integrin-dependent recognition of apoptotic material [17, 18].

Urbonaviciute and Voll [4] describe in their review in the current volume of the Journal of Internal Medicine how by injecting HMGB1–nucleosome complexes they have been able to demonstrate that these complexes may inhibit immune tolerance to nuclear components including dsDNA even in non-SLE-prone mice [9]. The nucleosome needs to be complexed with HMGB1, because as individual components neither the nucleosome nor HMGB1 can inhibit immune tolerance. HMGB1 binds with low avidity to the chromatin in living cells but binds very strongly during apoptosis. Poorly degraded apoptotic nuclear material will thus provide a source of nucleosome–HMGB1 complexes in vivo. These complexes signal via TLR2, but which part of the complex is involved in receptor binding remains unknown.

Many patients with SLE display a continuous production of IFN-α by plasmacytoid dendritic cells that become activated by immune complexes containing nucleic acids and autoantibodies. Type-I IFNs induce maturation of monocytes to dendritic cells needed for antigen presentation and activate B lymphocytes to produce antibodies including autoantibodies. These observations support the view that IFN-α is a central cytokine in the pathogenesis of SLE. As HMGB proteins are directly involved in the regulation of IFN type-I production, as outlined by Yanai et al. [3], it is likely that intracellular HMGB1 is responsible, at least in part, for the increased IFN-α production in SLE patients. However, to date, only extracellular HMGB1 but not cytoplasmic HMGB expression has been studied in SLE patients.

Future studies of the functional role of HMGB1 in SLE should take into account that the redox status of a single amino acid residue, the cysteine in position 106 (C106), determines whether or not HMGB1 alone will act as a cytokine-inducing molecule [19–21]. HMGB1 functions as a cytokine-promoting molecule only when the C106 resides in its reduced form expresses a thiol group, whereas oxidation will eliminate this proinflammatory activity [19–21]. There is thus a need for assessment of both total HMGB1 levels and redox modifications of the molecule to understand its functional role. However, the redox status of HMGB1 does not seem to influence its proinflammatory role when complexed with partner molecules. The discrepancy between the functional consequences of oxidation of HMGB1 as an individual molecule or in complex with others is probably explained by separate receptor usage. Uncomplexed HMGB1 signals via TLR4 [20], which requires C106 in its reduced form, whereas HMGB1–partner molecule complexes signal via the reciprocal receptors for the HMGB1–partner molecule [22]. Taken together, these findings provide strong evidence that HMGB1 is an important agent in SLE, but future studies are required to delineate the functional consequences of HMGB1 as an individual molecule and in complex with others.

IL-1β and HMGB1 in epilepsy

Epilepsy is a major medical problem, affecting up to 1% of the population. The current treatments for epilepsy are symptomatic, and about 30% of patients with epileptic do not respond to available anti-epileptic drugs. Therefore, there is an unmet medical need to develop treatment for drug-resistant forms of the disease and, importantly, drugs that can modify the disease process.

For about a decade, it has been known that brain inflammation is a contributing factor to different forms of epilepsy, both in experimental and clinical settings. The pioneering work of Vezzani et al. [23, 24] showed that intracerebral application of IL-1β increases seizure activity provoked by chemoconvulsant compounds in rats and mice. Furthermore, IL-1 receptor antagonism and inhibition of IL-1β biosynthesis demonstrated anticonvulsant effects. Based on studies investigating the role of IL-1β signalling in epilepsy, clinical trials have been initiated in the USA in patients with epileptic targeting IL-1β in the brain.

In addition to IL-1β and other classical proinflammatory cytokines, danger signals (DAMPs) released from damaged or stressed tissue could play a role in brain inflammation that predisposes to epilepsy. In particular, HMGB1 is currently recognized as one of the most important of the DAMPs, linking tissue injury or stress to activation of innate immune mechanisms. Maroso et al. have tested the hypothesis that HMGB1 might play a role in brain inflammation that predisposes to epilepsy. They have demonstrated that intracerebral injection of HMGB1 increases seizures caused by chemoconvulsants in a TLR4-dependent manner [25]. HMGB1 not only increases seizure frequency and duration but also accelerates seizure onset that starts within minutes after the proconvulsant challenge. These authors suggest that HMGB1 is essential for precipitation of the first seizure after a challenging event. Epileptic discharge would then promote further release of HMGB1 and activation of TLR4 resulting in recurrence of epileptic activity [25].

The molecular mechanisms underlying the proconvulsant activities of IL-1β and HMGB1 appear to be quite similar to each other. Both IL-1β and HMGB1 enhance phosphorylation of the NR2B subunit of the N-methyl-d-aspartate (NMDA) receptor, which results in enhanced NMDA activity and Ca2+ influx into neurons. Seizure generation and recurrence are dependent on NMDA receptor activation. In addition to these rapid effects, long-term effects of IL-1β and HMGB1 may also share similar mechanisms. It has been suggested that both factors augment the transcriptional activation of endothelial cells and perivascular astrocytes leading to damage of the blood–brain barrier, which predisposes to epileptic activity. In addition to these similarities in acute and long-term effects, IL-1β has been shown to bind to HMGB1 and enhance its inflammatory activity [26, 27], raising the possibility that the two factors might act as a complex to enhance epileptic activity.

As is usually the case, fascinating novel developments raise further intriguing questions. One such question is: how does a previously normal brain become chronically epileptic? Chronic changes in firing properties of the epileptic brain may depend on structural changes, such as the development of abnormal neuronal network properties because of aberrant sprouting of neurites, which is known to accompany different forms of epilepsy. It may be worth noting in this context that HMGB1 was initially isolated from brain extracts as an extracellularly acting factor [28] based on its ability to induce neurite sprouting. Further studies are therefore warranted to understand whether upregulation of HMGB1 expression and its release to the extracellular space in epileptic brain induce aberrant neurite sprouting, which contributes to epileptogenesis. Misexpression of HMGB1 might thus contribute to development of chronically epileptic tissue in a manner that is analogous to its role in brain development.

Conflict of interest statement

No conflict of interest was declared.

Ancillary