Biology, role and therapeutic potential of circulating histones in acute inflammatory disorders

Abstract Histones are positively charged nuclear proteins that facilitate packaging of DNA into nucleosomes common to all eukaryotic cells. Upon cell injury or cell signalling processes, histones are released passively through cell necrosis or actively from immune cells as part of extracellular traps. Extracellular histones function as microbicidal proteins and are pro‐thrombotic, limiting spread of infection or isolating areas of injury to allow for immune cell infiltration, clearance of infection and initiation of tissue regeneration and repair. Histone toxicity, however, is not specific to microbes and contributes to tissue and end‐organ injury, which in cases of systemic inflammation may lead to organ failure and death. This review details the processes of histones release in acute inflammation, the mechanisms of histone‐related tissue toxicity and current and future strategies for therapy targeting histones in acute inflammatory diseases.

DAMPs. Histones released passively from necrotic cells (or actively by other modes of cell death including NETosis) act on adjacent cells and circulating immune cells via pattern recognition receptors to effect specific biological activity. In in vivo systems, these effects can be difficult to study, as histones are co-released with nuclear DNA and other nuclear DAMPs such as HMGB1 (high mobility group box protein 1), each with their individual activities. Indeed, the mechanism of cell necrosis has significant impact on the kinetics of nuclear DAMP release, 11 and nuclear DAMPs acting as complexes have been reported to exert different activities compared to protein isolates. 12 Furthermore, where purified histones injected into experimental animals are lethal within minutes, 13 necrotic cell death releases nucleosomes (ie: histone-DNA complexes) which overall appear to be less toxic. 14 Indeed a study injecting similar doses of nucleosomes in mice makes no mention of toxicity, 15 and others have demonstrated cofactors such as HMGB1 responsible for the immune-stimulatory effects of nucleosomes. 16 Only through the interplay of plasma proteases and nucleases including DNAse1 and factor VII activating protease does nucleosome decondensation occur 17 ; however, this also degrades the histone component and limits cytotoxicity. 18 These effects may have significant implications for in vitro signalling studies using recombinant proteins, as effects of isolated nucleosome components may not become apparent in this setup.
Fragments of cell membrane and nuclear proteins also interact with complement proteins and complement cascade regulators to facilitate cell turnover and clearance. 19 An important regulator of nucleosome toxicity appears to be factor H of the family of complement regulator proteins. Factor H is actively internalized by apoptotic cells, where it leads to C3 complement activation and cell surface expression, as well as enhanced nucleosome clearance and phagocyte cytokine-release response to nucleosomes. 20 Cells undergoing secondary necrosis can thereby elicit a targeted pro-inflammatory response. 21 Once released from the nucleosome, extracellular histones exert their injurious effects in three ways summarized in Figure 1 and (c) through direct cytotoxicity.

| Chemoattractant effects
Histones are both directly chemoattractant, induce release of chemokines from adjacent cells and induce activation of vascular endothelium to promote adhesion and trans-migration of leukocytes.
Direct chemoattractant effects have been demonstrated in vitro using hepatocellular carcinoma cells. Histone-induced migration of these cells is dependent on the activation of the MAPK/ERK/NF-kB pathway via TLR4 (toll-like receptor 4). 22 Similarly, histones induce secretion of chemokines CXCL9 and CXCL10 from human monocytes and CXCL10 co-localizes with extracellular histone H4 in necrotic (but not healthy) tissue. 23 Extracellular histones H3 and H4 F I G U R E 1 Immunostimulatory effects of passively released histones acting as damage-associated molecular patterns (but not H1 or H2A/H2B) also activate vascular endothelial cells to increase cell surface expression of E-selectin, ICAM-1 and VCAM-1, thereby increasing leukocyte adhesion, rolling and transmigration in a TLR9-dependent manner. 24 In higher concentrations, these histones are toxic to the endothelium and represent a putative mechanism for pulmonary haemorrhage and ARDS in sepsis 13 or pancreatitis. 25 Histones acting on endothelial cells via TLRs 2 and 4 also activate NF-kB and AP-1 pathways to induce tissue factor expression, 26 thereby creating a pro-thrombotic milieu contributing to the microvascular thrombosis seen in many acute inflammatory diseases. Together, these effects describe the positive feedback loop that can lead to necroinflammation-where the death of relatively few cells induces further injury through inflammatory cell recruitment leading to organ failure, especially within the liver and/or kidney. 27

| Pattern recognition receptor responses
The intracellular signalling pathways of extracellular histones as DAMPs acting via TLRs2/4/9, MyD88, NF-kB and the NLRP3 inflammasome have been well documented and recently reviewed. [28][29][30] Functionally, histones injected into the renal artery of rats induced necroinflammation as well as IL-6, TNF-α and iNOS release. 31 These effects were reduced in TLR2/4 knock-out mice and more pronounced following LPS priming, which increased TLR2/4 mRNA transcription. Low doses of histone H3 (10 μg/mL) have been shown to induce release of IL-6 and IL-8 in ARPE-19 cells, as well as lead to the phosphorylation of ERKs, p38 MAPK and JNK and inhibition of these kinases all resulted in reduced cytokine release. 32 Higher doses (50 μg/mL), however, led to cell death in a manner that could not be inhibited using signalling kinase inhibitors. Histones also exacerbate ischaemia/reperfusion injury by a TLR9/MyD88-dependent mechanism and enhance extracellular DNA-mediated activation of TLR9 in immune cells. 33 Further to their effect on TLRs, histones also appear to induce IL-1β secretion and activation via an NLRP3/ASC/caspase1dependent mechanism, leading to neutrophil recruitment to sites of inflammation. 34 Critically, induction of leukocyte cytokine production and release is not dependent on free, circulating histones; nuclear material within blebs from apoptotic cells can induce similar stimulatory effects within resident or infiltrating phagocytes. 35

| Signal recognition
A large number of different signals have been shown to be able to induce NET formation, including bacteria, 36,37 viruses, 38 yeasts, 39,40 parasites, 41 organic crystals, 42 non-organic matter, 43 cytokines 44 and cellular breakdown products including nuclear DAMPs. 45,46 In order to detect such a variety of signals, there is overlap and convergence of receptor pathways. This may explain some variability in early genetic knock out studies when defining which receptor is critical in mediating NET release. It would seem molecular pattern-related NET release is mediated predominantly through TLRs 2, 4 and 9, 31,46,47 immune complex-related NET release is mediated via Fc receptors and MAC-a 48 and larger pathogens or inorganic matter lead to NETosis though size. The inability to phagocytose large particles within a given time appears to drive neutrophils to autodigest and release NETs in a process dependent on dectin-1. 49 While many signals leading to NETosis may make this an unlikely therapeutic target, it suggests that blocking destructive NETosis in sterile inflammation is possible without affecting a potentially beneficial antimicrobial response.   63 Genetic deletion of PAD4 leads to an inability of neutrophils to release NETs in response to calcium ionophore treatment or lipopolysaccharide (LPS), 64 and pharmacological inhibition of PAD4 inhibits NET formation in murine and human neutrophils. 65 Overexpression of PAD4, on the other hand, has been shown to cause histone hypercitrullination, nuclear decondensation and release of NET-like structures in an osteosarcoma cell line. 66 Nuclear translocation of granular proteases is the next step towards NET release. Neutrophil azoruphilic granules contain neutrophil elastase (NE), proteinase 3 (PR3) and cathepsin G (CG); however, only NE is translocated to the nucleus and neither inhibition of PR3 nor CG can prevent this translocation. 7 Furthermore, the process does not appear to be mediated by fusion of granules with the nucleus, but rather NE dissociates from the granular membrane in a ROS-dependent manner, before degrading cytosolic actin, arresting actin dynamics and translocating across the nuclear membrane using specific translocation mechanisms. 67 Binding of nucleic acid by proteases initiates a process of degradation of nuclear binding proteins 68 and controlled integration of MPO into the forming NET.

| Signal transduction
Nuclear NE leads to early degradation of linker histone H1, followed by core histone H4 which coincides with nuclear chromatin decondensation. 7 Histone H3 appears to be resistant to degradation in intact nuclei, but not in purified form, suggesting one of the purposes of post-translational modification is to render histone H3 resistant to NE-related degradation. This offers novel targets for therapy that have not yet been exploited.
The pathway described above is the best described due to the use of PMA as experimental stimulant of NETosis. In this experimental setup, the three steps are sequential; however, there have been recent reports of NET-like structures being released rapidly (minutes), by budding of DNA/histone/protease-containing vesicles from the nucleus followed by active exocytosis of NET-containing vesicles. 69,70 This potentially bypasses most of the mechanisms described above and requires further study.

| Autophagy
Although most studies support the conclusion that autophagy is essential for NETosis, 71,72 inhibition of mammalian target of Rapamycin (mTOR), a regulatory and inhibitory protein complex, has been reported to reduce NETosis stimulated by bacterial LPS. 73 Stimulation of human neutrophils with vasculitis-associated antibodies led to massive vacuolization, increased LC3BI degradation and could be F I G U R E 2 Signalling cascade leading to neutrophil extracellular trap release in murine and human neutrophils reduced with the inhibitors of autophagy 3-methyladenine (3MA) and LY294002. 74 Similarly, LC3B containing vacuoles were observed preceding NETosis in LPS or septic plasma-induced NETosis in human neutrophils which was also effectively inhibited by 3MA and bafilomycin A1. 75 Knock down of phosphatase and tensin homologue deleted on chromosome 10 (PTEN), a potent regulator of autophagy, reduced PMA-induced NETosis in HL-60 cells and overexpression increased it. 76 PKC has been shown to stimulate autophagy which in response to certain stimuli can be independent of mTOR, offering a potential explanation for this discrepancy. 77 Figure 2 demonstrates how different stimuli resulting in NETosis can have differential effects on autophagy.

| Effects of concentration and histone type on different cells and/or tissues
A wide variety of organisms actively release histones and histone degradation products as microbicidals (histone-derived antimicrobial peptides; HDAP). Table 1  Cellular injury mediated by extracellular histones has been described experimentally or in human disease of the lung, 94 heart, 95 liver, 96 kidney 31 and vascular endothelium. 62 Table 2 details the effects of extracellular histones observed in specific cell types.
Breakdown of DNA in NETs with DNAse only partially ameliorates NET-related toxicity, as it does not affect the histone component. 125 Concentrations below 10 μg/mL seem to have a signalling function and can induce calcium transients in cells. 126 Concentrations greater than 10 μg/mL (or 20 μg/mL in the presence of serum) induce cell death by an uncertain mechanism, which may involve the formation of non-specific cationic pores in cell membranes. 127,128 Concentrations above 100 μg/mL cause rapid necrosis.
Core histones H3 and H4 are most frequently reported to increase in plasma from sepsis patients as well as experimental sepsis and therapeutic administration of antibodies to these histones improve outcomes in these models. 102,129,130 It is conceivable that histone citrullination as described above renders these less susceptible to degradation and easier to detect, creating a publication bias.
As core histones oligomerize readily with each other in solution 131 and will surely rapidly do so upon histone release from any cell type, it is difficult to dissociate toxicity of individual histones from each other in biological systems. When used in in vitro studies, recombinant histones H2A and H2B were also able to induce cellular currents 126 or activate thrombin. 122

| Calcium/ionic pore effects
The interaction of histones with cell membranes is heavily reliant on charge. Positively charged histones preferentially bind anionic phospholipids such as cardiolipin or phosphatidylserine, but not zwitterionic phospholipids like phosphatidylcholine. 132 Furthermore, adding negative charge (eg, a phosphate head group as in phosphatidylinositol bis-phosphate) increases the binding capacity of histones as measured by calorimetry. 128 Histones have also been shown to expose phosphatidylserine on the surface of red blood cells in a dose-dependent manner 47

| Effects on coagulation
The ability of NETs and histones to influence the coagulation cascade and actually initiate venous thrombosis 8,133,134 is the most recent detail in the emerging field of NETosis research. Clinically, circulating nucleosomes are independent prognostic markers of disseminated intravascular coagulopathy (DIC) 135 and some countries, notably Japan, are actively promoting the use of anticoagulants as histone detoxification agents in DIC. 136 Positive correlations between histone levels and coagulopathy can also be seen in trauma patients 137 and patients with sepsis. 129 Figure 3 summarizes the effect of histones and NETs on the coagulation cascade. Histones act synergistically to produce a profound pro-coagulant drive. Histones are able to induce platelet aggregation and factor V/Va expression and prothrombinase activity, leading to thrombin activation independent of the intrinsic coagulation pathway. 122 Histones also inhibit thrombomodulin and protein C activation, 138 an effect most pronounced with histones H3 and H4, thus reducing a natural thrombin inhibitor system. Furthermore, histone H4 binding promotes thrombin autoactivation, probably by fixing the prothrombin molecule in a conformational state conducive to proteolytic attack. 139 The only exception is linker histone H1, which has been shown to reduce thrombin activation and prolong  140 ; this mechanism is likely insignificant in acute inflammation, as histone H1 is amongst the first nuclear proteins to be degraded in the process of NETosis.
The presence of DNA in NETs also allows activation of the intrinsic coagulation pathway, demonstrated by NET-enhanced thrombin generation in platelet-poor plasma, reduced by factor XII/ factor XI depletion or DNAse treatment. 141 DNAse treatment in platelet-rich plasma further increases thrombin generation, indicating differential effects of histones and NETs in different microenvironments. The addition of histones and DNA also increases fibrin fibre thickness, clot stability and delayed clot lysis 142 as well as reducing anti-thrombin-mediated thrombin inactivation and plasmin activity. 143 In in vivo systems, the interplay between von Willebrand factor (vWF), platelets and neutrophils anchors neutrophils to otherwise healthy vessel walls and permits NETing neutrophils to initiate clot formation, 125,144 with propagation that may occur or be enhanced by the mechanisms described above. Deficiencies in degradation of vWF produce clinical microangiopathies (eg, thrombotic thrombocytopenic purpura), the severity of which is also closely correlated with circulating NET components in humans. 145  NETs must be overcome, which is well illustrated by the finding that

HISTONE DETOXIFICATION IN PATHOLOGY
DNAse therapy in addition to tPA is more effective than either therapy alone. 152 In models of sterile and infective acute inflammatory diseases, administration of histones exacerbates end-organ injury consistently (Table 3). Similarly, damaging effects are at least partly ameliorated by the application of histone-targeted or histone-specific therapies.  Heparin (unfractionated and/or low molecular weight) [166] thrombosis, worsens tissue perfusion and contributes significantly to organ injury. Recognition of this phenomenon may allow targeted therapy, limiting systemic injury and improving survival.

CONF LICTS OF INTEREST
The authors confirm that there are no conflicts of interest.