Release mechanisms of major DAMPs

Damage-associated molecular patterns (DAMPs) are endogenous molecules which foment inflammation and are associated with disorders in sepsis and cancer. Thus, therapeutically targeting DAMPs has potential to provide novel and effective treatments. When establishing anti-DAMP strategies, it is important not only to focus on the DAMPs as inflammatory mediators but also to take into account the underlying mechanisms of their release from cells and tissues. DAMPs can be released passively by membrane rupture due to necrosis/necroptosis, although the mechanisms of release appear to differ between the DAMPs. Other types of cell death, such as apoptosis, pyroptosis, ferroptosis and NETosis, can also contribute to DAMP release. In addition, some DAMPs can be exported actively from live cells by exocytosis of secretory lysosomes or exosomes, ectosomes, and activation of cell membrane channel pores. Here we review the shared and DAMP-specific mechanisms reported in the literature for high mobility group box 1, ATP, extracellular cold-inducible RNA-binding protein, histones, heat shock proteins, extracellular RNAs and cell-free DNA.


Introduction
Damage-associated molecular patterns (DAMPs) are endogenous molecules which serve as potent activators of the immune system [1]. Examples of DAMPs include nuclear and mitochondrial DNA, RNA, nucleotides and nucleosides, DNA-binding molecules, temperature-shock proteins, and uric acid [1][2][3]. DAMPs normally reside inside the cell playing diverse roles in homeostasis, but are released to the extracellular space when cells are exposed to stress [1]. Cellular stressors that can lead to DAMP release include a wide array of physical (trauma, radiation), chemical (toxins, osmolarity), metabolic (ischemia/reperfusion), and infectious (viruses, bacteria, protozoa) factors [1,4,5]. Once outside the cell, DAMPs are recognized by other cells via their interaction with cellular receptors such as pattern recognition receptors (PRRs), which then upregulate stress response mechanisms that often converge to form a positive feedback loop of tissue injury and inflammation [1]. Indeed, the relevance of DAMPs to various diseases is supported by a number of studies. In sepsis, circulating DAMPs correlate with disease severity and their inhibition has been shown to improve the outcomes in experimental models of sepsis [1]. During cancer different types of DAMPs can promote tumor establishment and progression as well as metastasis [6]. DAMPs are upregulated systemically and locally in patients with autoimmune diseases such as rheumatoid arthritis and their neutralization has shown to prevent the disease progression in animal models [7]. Therefore, DAMP-inhibiting molecules have the potential to significantly attenuate inflammation and in the future may yield a novel class of anti-inflammatory drugs able to finally treat trauma, ischemia/reperfusion injury, sepsis, neuroinflammation, and other pathophysiological conditions irresponsive Max Brenner and Ping Wang have contributed equally to this work.
to existing immune modulatory drugs. Indeed, molecules targeting a number of DAMPs are already in the process of being developed as potential therapeutic agents [8]. At the moment, the predominant strategies to decrease the effects of DAMPs consist of antibody neutralization, competitive antagonism, and enzymatic inactivation [1,8]. An alternative approach at a more fundamental level is to actively suppress the cellular release of DAMPs. While thorough understanding of the mechanisms of DAMP release may lead to new treatments to attenuate the proinflammatory activity of DAMPs, it might also help to overcome the difficulties we are facing with other forms of immunotherapy such as cytokine removal [9].
Here we review the mechanisms of DAMP release that have been reported. We first summarize universal mechanisms affecting most if not all DAMPs (i.e., different types of cell death, lysosomal-and exosomal-exocytosis) (Fig. 1). We then outline the mechanisms that have been described for selected DAMPs individually, including high mobility group box 1 (HMGB1), ATP, extracellular cold-inducible RNA-binding protein (eCIRP), histones, heat shock proteins (HSPs), extracellular RNAs (exRNAs) and cell-free DNA (cfDNA), as they have been well-studied among other DAMPs (Table 1).

Universal mechanisms of DAMP release
Different DAMPs share common mechanisms for their release. The release mechanisms can be largely divided into two categories: passive release mainly due to cell death and active release from live cells represented by exocytosis. To be precise, with the exception of necrosis, the other forms of cell death are not entirely passive but rather a regulated process, and some of the mechanisms described under cell death do not always lead to cell death (e.g., NETosis can be suicidal and vital). Thus, it has to be noted that the following categorization is not completely clear-cut and somewhat vague with overlap.

Passive release (cell death)
DAMPs are well known to be released during different types of cell death. Necrosis is the most common cell death to cause passive release of DAMPs, whereas necroptosis, apoptosis, pyroptosis, ferroptosis and extracellular traps can also contribute to DAMP release. Theoretically, necrosis can cause the release of mixed DAMPs since the cell boundary is lost due to membrane rupture and any cellular components would be released. On the other hand, different forms of cell death can be rather specific to the types of DAMPs as to their release according to the mechanisms. For example, apoptosis leads to the release of nuclear DAMPs following chromatin condensation and DNA fragmentation. Extracellular traps released during NETosis contain DAMPs which are mainly nuclear molecules and antimicrobial enzymes. Conversely, DAMPs can give clues to by which cell death they were released. The fragment length of cfDNA is different according to the type of cell death; cfDNA originating from apoptotic cells is ~ 180 bp due to the fragmentation, while cfDNA released from necrotic cells can be as long as > 10,000 bp [10]. HMGB1 released by pyroptosis via  inflammasome pathway is hyperacetylated, which is not seen when it is released from necrotic or apoptotic cells. In addition, the redox state of HMGB1 is in the disulfide form after pyroptosis, in the fully reduced or disulfide form after necrosis, and in the fully oxidized form (sulfonyl HMGB1) after apoptosis [11]. Different DAMPs can be released at the different stages even within the same type of cell death. During apoptosis, ATP is released at the pre-apoptotic stage while HMGB1 is released at the late stage [12,13]. Besides the preceding DAMPs, it still largely remains elusive and further studies are awaited as to the types and stages of the cell death of their origin.

Necrosis and necroptosis
Passive release of DAMPs is typically described to occur as a result of necrosis. Necrosis is commonly caused by extreme chemical or physical insults such as the presence of toxins or trauma, and is characterized by cell swelling and plasma membrane rupture [14]. Tissue ischemia and hypoxia also trigger necrosis by depleting intracellular ATP to unbalance the pump-leak mechanism leading to an influx of Na + and water, which causes cell swelling. Reperfusion can further damage the cells by inducing the generation of multiple oxidants and free radicals [15]. A number of DAMPs have been found to be released by necrosis including, but not limited to, HMGB1, ATP, histones, HSPs, exRNAs, cfDNA, and possibly eCIRP [1,3,5,10,[16][17][18][19][20][21][22].
Although cytoplasmic membrane rupture can be uncontrolled accidental (mechanical or chemical) events, it can also be a regulated process governed by specific caspases and kinases. For instance, necroptosis occurs as a result of receptor-interacting serine/threonine-protein kinase 1 (RIPK1) activation followed by RIPK3-dependent phosphorylation of mixed lineage kinase domain like pseudokinase (MLKL) to induce MLKL oligomerization, which results in plasma membrane rupture [14]. DAMP release has been significantly less studied in the context of necroptosis than of necrosis. However, as membrane integrity is lost in necroptosis in a fashion similar to necrosis, theoretically necroptosis also results in the release of DAMPs and other cellular components to the extracellular space [23].

Pyroptosis
Another form of caspase-dependent cell death is pyroptosis, which is induced via the activation of caspase-1 following that of inflammasomes such as NLRP3 or via caspase-4/5/11 activation typically initiated by intracellular LPS [24,31]. The inflammatory caspase (−1, −4, −5, −11) activation further induces the cleavage of gasdermin D (GSDMD) to promote pore formation in the membrane, allowing the release of intracellular molecules [24,31]. A recent study has revealed that intracellular protein DDX3X promotes NLRP3 inflammasome activation, which can be inhibited by the induction of stress granules causing the sequestration of DDX3X, thus acts as a live-or-die checkpoint in stressed cells [32]. Pyroptosis is typically known for the release of IL-1ß, though some other DAMPs such as HMGB1, ATP and cfDNA can be released by cells undergoing this type of cell death [1,[33][34][35][36][37]. Even sharing the same signaling pathway, IL-1ß is well known to be released through GSDMD pore whereas a study has shown HMGB1 was released as a result of cell lysis during pyroptosis [38]. In addition, post-translational modifications are required prior to pyroptosis for the release of HMGB1 as described later [39].

Ferroptosis
Ferroptosis is a programmed cell death accompanied by iron accumulation and lipid peroxidation. Its morphological features include a loss of membrane integrity, cytoplasmic swelling, swelling of cytoplasmic organelles and moderate chromatin condensation [40]. Intracellular iron accumulation, which can be induced by ferroptosis activators such as erastin and RSL3, causes oxidative stress directly by generating reactive oxygen species (ROS) via the Fenton reaction as well as activating the enzymes responsible for lipid peroxidation and oxygen homeostasis. Glutathione peroxidase 4 (GPX4) is an anti-oxidative enzyme which plays a major role in regulating ferroptosis by preventing lipid peroxidation. Thus, inhibition of GPX4 is a well know mechanism for inducing ferroptosis [40]. Although it is a relatively new concept and less has been elucidated yet, HMGB1 and cfDNA have been regarded to be released by ferroptosis [40,41]. A study has shown that HMGB1 release during ferroptosis was due to HMGB1 acetylation induced by histone deacetylase (HDAC) inhibition mediated by autophagy [41].

Extracellular traps
Neutrophil extracellular traps (NETs) are web-like chromatin-based structures released by neutrophils primarily for pathogen clearance via a regulated process, called NETosis [1]. It is typically described that in the neutrophils undergoing NETosis activated peptidylarginine deiminase 4 (PAD4) citrullinates histones leading to chromatin decondensation accompanied by the dissolution of nuclear and granule membranes. DNA and histones mix with granule-derived antimicrobial peptides in the cytoplasm and are extruded into the extracellular space [1]. Conventionally NETosis was regarded to be a suicidal process leading to cell death (i.e., suicidal NETosis), however later it was found that NETs can also be released from live cells (i.e., vital NETosis) [1]. NETs can be detrimental as they contain not only typical DAMPs, such as histones, cfDNA and eCIRP, but also antimicrobial enzymes, such as neutrophil elastase (NE) and myeloperoxidase (MPO), both of which can directly cause tissue damage [1,42]

HMGB1
HMGB1 is a nuclear protein capable of binding chromosomal DNA to fulfill its nuclear functions in stabilizing nucleosomal structure and stability and regulating gene expression. As a prototypical DAMP, HMGB1 can be passively released by somatic cells undergoing cytoplasmic membrane destruction due to accidental necrosis or regulated cell death processes such as necroptosis, pyroptosis ferroptosis, or apoptosis [16, 23, 27,

eCIRP
CIRP is an RNA chaperone protein which plays a role in the regulation of a variety of cellular stress responses. Extracellular CIRP (eCIRP) is a DAMP that perpetuates inflammation and contributes to various diseases [5]. An in vitro study showed that passive release by necrosis might not be a major source of eCIRP, although passive release is a likely important source of CIRP in conditions such as trauma, ischemiareperfusion injury, and sepsis [5,51]. Extracellular traps have also shown to be sources of eCIRP [42]. Like HMGB1, CIRP does not contain a signal peptide, thus its release is unlikely to be mediated by ER-Golgi dependent classical pathway. The nuclear to cytoplasm translocation of CIRP requires post-translational modifications such as methylation and phosphorylation [75,76]. The phosphorylation of CIRP was shown to be mediated by GSK3β and casein kinase II (CK2) [76]. CIRP migrates from the nucleus to cytoplasmic stress granules under certain conditions such as oxidative stress, ER stress, hyperosmotic, and heat shock [75]. Stress granules are RNA-protein complexes that assemble when cells undergo polysome disassembly effectively interrupt protein translation in response to the conditions such as those above. When the offending condition is controlled, stress granules disassemble and translation resumes [77][78][79]. Stress granules interact with the inflammasome pathway as they inhibit NLRP3 inflammasome assembly by sequestering DDX3X protein [32]. Studies have shown stress granule proteins can be released via extracellular vesicles such as exosomes, though it's still not clear whether CIRP can be released through the same pathway [80][81][82]. Exocytosis of secretory lysosomes is also likely to contribute to CIRP release as a study showed CIRP was enriched at the lysosomal compartment of macrophages subjected to hypoxia [51].

Histones
Histones are components of chromatin in the nucleus together with DNA, but can act as DAMPs by binding to PRRs once they are released to the extracellular space [1]. Histones can be released passively by necrosis like other DAMPs [1,18]. Upon apoptotic signaling, core histones (H2A, H2B, H3, and H4) and a link histone (H1) undergo post-translational modifications (e.g., H2B phosphorylation at serine 14, H2B acetylation at lysine 15, etc.), which have been reviewed in depth by Füllgrabe et al. [83]. During apoptosis, the modified histones separate from the genomic DNA and translocate to the cytoplasm. The histones protrude from the plasma membrane to be exposed at the cell surface and released to the extracellular space through a mechanism that has not yet been elucidated [26]. One of the main components of NETs, citrullinated histones, mediated by PAD4, are released by the NETosis mechanisms described earlier [1]. Histones can also be secreted actively from living cells via exosomal exocytosis. A study has shown that histones were present on the outer surface of exosomes released from LPS-challenged macrophages and interacted with TLR4 directly [52].

HSPs
HSPs are a family of molecular chaperones maintaining cellular homeostasis [84]. Extracellular HSPs act as DAMPs and correlate with the severity of several disorders such as sepsis and trauma [85,86]. HSPs can be released passively by necrosis [19,53]. HSPs lack a secretory signal, thus are unlikely to be released via ER-Golgi transportation. Proposed mechanisms for active HSP release include secretion via ATP-binding cassette (ABC) transporter in the lysosomal pathway and via secretory granules [53, 87,88]. Above all, the most accepted mechanism of HSP release is via extracellular vesicles. HSPs are found in two types of extracellular vesicles; exosomes and ectosomes [53]. While exosomes are released by the mechanism described in the earlier section, ectosomes are formed by the outward budding of plasma membrane in response to the increase of cytosolic free Ca 2+ and are released directly to the extracellular space. Ectosomes are also called microvesicles or microparticles [89]. Different types of HSP family, such as HSP27, HSP60, HSP70 and HSP90, have been found to be released via extracellular vesicles as reviewed by Maio et al.
[53]. While HSPs might be located within the vesicle lumen, multiple studies showed HSPs were at least in part present on the surface of the extracellular vesicles, allowing them to interact with surface receptors directly [90][91][92][93][94].

exRNAs
While some exRNAs are known to have beneficial effect, other act as DAMPs to aggravate inflammation [95]. Though exRNAs can be released by necrosis and apoptosis, they are regarded to be more stable when encapsulated within extracellular vesicles, typically in exosomes, to avoid degradation by RNases in biological fluids such as saliva, breast milk, blood, cerebrospinal fluid, follicular fluid, and urine [22,30,60]. The import of exRNAs, specifically miRNAs, to extracellular vesicles is mediated by the association with argonaute 2 (Ago2) regulated by KRAS-MEK-ERK signaling [96]. Other RNA binding proteins (RBPs), such as Y-box protein 1 (YBX1), hnRNPA2B1 and SYNCRIP (hnRNPQ), have also been reported to play a role in the sorting of miR-NAs [97][98][99][100]. RNAs loaded to the extracellular vesicles are released via exocytosis to become exRNAs. RBPs not only mediate the import but also act as carriers of exRNA in the circulation even without the encapsulation by extracellular vesicles [101,102]. It is still unknown whether RBP/ exRNA complexes in the blood are predominantly released passively by cell death or exported actively via an independent pathway.
cfDNA DNA in the extracellular space can serve as a DAMP. Various types of cell death, such as necrosis, apoptosis, pyroptosis ferroptosis, and NETosis are known to release DNA [1,10,20,26,30,33,40]. Its mechanisms of release differ according to the pathogenic condition exemplified by NETosis in sepsis patients and necrotic cells in trauma patients [20]. A recent study showed that cellular senescence was a major determinant of cfDNA kinetics by negatively regulating its release. The elimination of senescent cells through apoptosis recovered cfDNA release [103]. In addition to the passive release by cell death, DNA can be released actively via extracellular vesicles, including exosomes and ectosomes [104]. A study suggests that majority of cfDNA in the blood is present in the form of exosomes, thus avoiding its degradation by nucleases [59]. Besides nuclear DNA, cell-free mitochondrial DNA (cf-mtDNA) also acts as a DAMP [105]. Cf-mtDNA was found in platelet-driven ectosomes along with phospholipase A2 (PLA2). PLA2 is a bactericidal enzyme found to digest the cellular membrane leading to the leakage of mtDNA into the extracellular space [106].

Conclusions and future prospects
DAMPs represented here by HMGB1, ATP, eCIRP, histones, HSPs, exRNAs and cfDNA can be released by several different active and passive mechanisms such as exocytosis of lysosomes/exosomes, necrosis/necroptosis, apoptosis, pyroptosis, ferroptosis, and extracellular traps. Some mechanisms such as necrosis are shared with more DAMPs, while others are relatively DAMP-specific, e.g., channel pores for ATP. Despite the detailed mechanisms reviewed here, clearly a lot still remains to be elucidated. A more comprehensive understanding of the mechanisms DAMP release and their regulation will not only enable the design of new investigative tools but may also result in new potential therapeutic approaches to attenuate inflammation and tissue injury and thus to improve the outcomes of pathological conditions associated with excessive DAMP release. Data availability Not applicable.

Conflict of interest
The authors declare no conflicts of interest.

Consent to participate Not applicable.
Consent for publication Not applicable.

Ethical approval Not applicable
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.