HMGB1 release by inflammasomes

High-mobility group box 1 (HMGB1, also known as amphoterin) was identified originally as a highly conserved non-histone DNA-binding factor that is expressed by virtually all nucleated eukaryotic cells.¹ Structurally, HMGB1 is a 30 kDa protein that is composed of two homologous DNA-binding motifs (called ‘A’ and ‘B’ boxes) followed by a negatively charged acidic tail in its carboxyl-terminus (Fig. 1A).^2,3 Although the nuclear role of HMGB1 is incompletely understood, the protein has been implicated in bending DNA to facilitate gene transcription, DNA replication and repair.^4–7 The importance of HMGB1 in regulating these nuclear processes is emphasized by the phenotype of HMGB1-deficient mice, which die shortly after birth because of severe energy deficits and hypoglycemia, possibly as a result of impaired glucocorticoid receptor signaling.⁸

In addition to its roles in regulating nucleosome function and transcription, HMGB1 more recently emerged as an extracellularly released mediator of inflammation and repair responses in lipopolysaccharide (LPS)-induced endotoxemia and sepsis.⁹ Passive immunization with HMGB1-neutralizing antibodies prevented organ damage in animal endotoxemia and sepsis models.^9–12 Moreover, increased HMGB1 serum levels were confirmed in septic patients.¹³ Subsequent studies implicated HMGB1 in the induction of systemic inflammation following trauma and ischemia/reperfusion injury in the liver, heart, kidney and brain.^14–17 Elevated HMGB1 levels in circulation were also associated with lung injury, myocardial infarction and rheumatoid arthritis.^18–21 HMGB1 induces the recruitment of inflammatory cells, contributes to dendritic cell maturation and to proliferation of activated T cells.^22,23 Thus, extracellular HMGB1 appears to be associated with inflammatory and repair responses during both infectious and (sterile) trauma and autoimmune disorders. Therefore, HMGB1 is widely considered to be a danger-associated molecular pattern (DAMP) that—similar to interleukin-33 and in agreement with Matzinger's danger model—engages immune receptors upon its release in the extracellular milieu.^24,25 Indeed, recombinant HMGB1 produced in Escherichia coli is a potent inducer of pro-inflammatory cytokines when administered to cells or injected in mice.^9,26–30 HMGB1-induced activation of immune receptors may involve the assembly of immunostimulatory complexes with endogenous or microbial co-factors. Indeed, the receptor for advanced glycation endproducts (RAGE) was potently activated by HMGB1/CpG DNA complexes, whereas highly purified HMGB1 has only weak pro-inflammatory activity.^30,31 In addition to RAGE, HMGB1 was suggested to induce secretion of pro-inflammatory cytokines and chemokines by inducing NF-κB signaling downstream of Toll-like receptor (TLR)2 and TLR4 activation.^32–34 Thus, extracellular HMGB1 might bind microbial components and/or endogenous substances that are released from injured tissue to induce or augment the production of inflammatory mediators during infections and in response to (sterile) trauma.

An intriguing question that has attracted considerable attention is how HMGB1 travels from the nucleus to reach the extracellular milieu given that it lacks a classical secretion signal. Such signal peptides are usually required for transporting secretory proteins through the endoplasmic reticulum (ER) lumen to the extracellular space in Golgi-derived secretory vesicles.^35,36 However, certain cytoplasmic and nuclear proteins that lack classical secretion signal peptides can reach the extracellular milieu through ER- and Golgi-independent pathways.³⁷ Indeed, HMGB1 was shown to be released from LPS-activated monocytes and macrophage cell lines consistent with its role as a prototypical alarmin.³⁸ Another protein thought to undergo unconventional protein secretion is the pro-inflammatory cytokine IL-1β.³⁹ IL-1β and the related cytokine IL-18 are both produced as inactive cytosolic precursor proteins that are only released following their maturation by the cysteine protease caspase-1. Caspase-1 itself requires activation by large protein complexes termed ‘inflammasomes,’ which are assembled on NOD-like receptor (NLR) and HIN-200 protein scaffolds.⁴⁰ NLRs contain a centrally located NACHT domain, whereas HIN-200 proteins such as the interferon-inducible member absent in melanoma 2 (AIM2) are characterized by the presence of a carboxy-terminal DNA-binding HIN domain.^41,42 Inflammasomes activate signaling pathways that contribute to immune responses and host defense mechanisms when they (directly or indirectly) sense the presence of particular microbial pathogens and DAMPs in the host cytosol or in intracellular compartments.⁴³ For instance, the NLRC4 inflammasome is responsible for caspase-1 activation in macrophages infected with the bacterial pathogens Salmonella typhimurium, Pseudomonas aeruginosa, Legionella pneumophila and Shigella flexneri.⁴³ On the other hand, the NLRP3 inflammasome induces caspase-1 activation upon infection with Staphylococcus aureus and Klebsiella pneumonia.^44–46 The NLRP3 inflammasome also induces secretion of IL-1β and IL-18 from LPS-primed macrophages that are subsequently treated with endogenous alarmins such as ATP and uric acid, or exposed to microbial toxins such as nigericin.⁴³ Intriguingly, NLRP3 inflammasome activation was recently shown to be essential not only for IL-1b secretion, but also for HMGB1 release from LPS-primed macrophages that are treated with ATP or exposed to nigericin.¹⁰ Similarly, extracellular release of HMGB1 from S. typhimurium-infected macrophages relied on activation of caspase-1 by the NLRC4 inflammasome.¹⁰ Unlike IL-1β and IL-18, HMGB1 did not undergo caspase-1-mediated processing. Nevertheless, the enzymatic activity of caspase-1 is required for secretion of IL-1β, IL-18, HMGB1 and growth factors such as fibroblast growth factor 2.^10,47 The observation that HMGB1 release proceeded unabated in macrophages lacking known caspase-1-substrates (IL-1β, IL-18 and caspase-7) suggests that caspase-1 may induce non-canonical secretion of HMGB1 and other proteins through proteolytic activation of an unknown secretion apparatus (Fig. 1B). In this context, the small GTPase Rab39a was recently characterized as a caspase-1 substrate that has been implicated in secretion of IL-1β from LPS-activated THP-1 cells.⁴⁸ Moreover, several mutually non-exclusive models have been proposed to explain the release of IL-1β, IL-18 and DAMPS such as HMGB1 in shedded microvesicles, secretory lysosomes or exosomes.³⁷

In addition to mediating the non-conventional secretion of pro-inflammatory cytokines, DAMPs and growth factors, inflammasome activation in activated immune cells triggers the induction of a specialized caspase-1-mediated cell death program called pyroptosis.^49,50 Pyroptosis is generally regarded as a pro-inflammatory cell death mode because it is associated with osmotic lysis and the release of the intracellular content into the extracellular milieu (Fig. 1B). Concurrently, reduced serum HMGB1 levels in caspase-1-deficient mice were shown to correlate with their resistance to LPS, and HMGB1 neutralizing antibodies protected IL-1β/IL-18-deficient mice from LPS-induced endotoxemia.¹⁰ In this regard, pyroptosis may resemble necrosis, an inherently proinflammatory and pathological cell death mode that results in extensive tissue damage.⁴⁹ Indeed, HMGB1 has been identified as a critical danger signal contributing to necrosis-associated inflammation.³⁸ This is illustrated by the observation that necrotic cell debris from HMGB1-deficient cells was significantly impaired in inducing pro-inflammatory cytokines.^38,51 It is thus tempting to speculate that cellular injury in response to trauma and microbial infections may trigger pyroptosis and necrotic cell death in order to induce inflammatory and repair responses through the extracellular release of DAMPs such as HMGB1 and S100 proteins (Fig. 1B).⁴⁹

In conclusion, inflammasomes may release HMGB1 in circulation through unconventional protein secretion, or as a consequence of cell lysis during pyroptosis. Depending on the pathological context, these mechanisms might contribute differentially to the extracellular release of HMGB1 and the induction of HMGB1-mediated immune responses. Further characterization of the molecular mechanisms of pyroptosis and unconventional protein secretion may shed light on the intricacies of inflammasome-mediated release of danger signals and inflammatory cytokines.