Extracellular DNA NET-works with dire consequences for health

Abstract

Neutrophils play a central role in innate immune defense. Advances in neutrophil biology have brought to light the capacity of neutrophils to release their decondensed chromatin and form large extracellular DNA networks called neutrophil extracellular traps (NETs). NETs are produced in response to many infectious and non-infectious stimuli and, together with fibrin, block the invasion of pathogens. However, their formation in inflamed blood vessels produces a scaffold that supports thrombosis, generates neo-antigens favoring autoimmunity, and aggravates damage in ischemia/reperfusion injury. NET formation can also be induced by cancer, and promotes tumor progression. Formation of NETs within organs can be immediately detrimental, such as in lung alveoli, where they affect respiration, or be harmful over longer periods of time. For example, NETs initiate excessive deposition of collagen, resulting in fibrosis, thus likely contributing to heart failure.

Here, we summarize the latest knowledge on NET generation and discuss how excessive NET formation mediates propagation of thrombosis and inflammation and, thereby contributes to various diseases. There are many ways in which NET formation could be averted or NETs neutralized to prevent their detrimental consequences, and we will provide an overview of these possibilities.

Introduction

Neutrophils, white blood cells with multi-lobular nuclei, are the first line of defense during infection and inflammation.¹ In fact, through their phagocytic and antimicrobial activity, they are responsible for protecting and maintaining health and promoting tissue repair.² In response to stimuli, a subset of neutrophils can release chromatin structures called neutrophil extracellular traps (NETs) that are decorated with histones and cytoplasmic and granular proteins.³ Studies have shown that low density granulocytes (LDGs) are more prone to generate NETs.⁴ However, although several molecular markers have been proposed to identify the neutrophil population capable in generating NETs,⁵ they have not been confirmed yet. In fact, whether NET producing neutrophils belong to a different subpopulation of neutrophils or are simply older more mature and primed cells still needs to be determined.

NET formation, is another mechanism of host defense. Initially recognized as a strategy to capture and kill bacteria, and thus protect the host from microbial invasion,³ it is now well established that NET release has more complex consequences. Excessive NET formation and/or a decrease of NET removal may promote inflammation, autoimmunity, vascular disease, thrombosis, and ischemia-reperfusion injury and can contribute to cancer growth and progression.^{1, 6} Targeting NET formation, NETs, and their components could therefore be a promising novel strategy for therapeutic interventions.

Mechanism of NET formation

NET structures were initially described in 2004 by Brinkmann and colleagues³ as a consequence of stimulation of neutrophils with phorbol myristate acetate (PMA), lipopolysaccharide (LPS), and interleukin (IL)-8. Since then, many more factors have been shown to induce NETs, including extracellular pathogens (Candida albicans,⁷ Streptococcus aureus),⁸ viruses,^{9, 10} other cytokines, cholesterol and urate crystals,^{11, 12} lipids,¹³ high glucose,¹⁴ activated platelets,^{15, 16} complement (C5a),¹⁷ hypoxia,¹⁸ and estrogen receptor modulators (e.g tamoxifen).¹⁹ The mechanism by which neutrophils form NETs depends on the nature of the stimulus.^{1, 20} However, they all lead to a common outcome: chromatin decondensation, nuclear rupture, and finally NET release (Figure 1). NET release is preceded by plasma membrane rupture, a process that has been referred to as lytic NETosis. Of note, the term “NETosis” was originally coined to distinguish NET formation as a distinct form of cell death from apoptosis and necrosis,²¹ but it is now clear that this is a misnomer as not all pathways of NET formation lead to cell lysis and cell death.²² We and others use the suffix –osis paired with NET to describe the biological process of NET formation. To avoid controversy, we avoid using the term NETosis in this review. PMA, LPS, and various types of bacteria, like Pseudomonas aerugionosa, activate the Raf-MEK-ERK pathway that stimulates ROS production by NADPH oxidase 2 (NOX2) complex (NOX2-dependent NET formation).²³ Neutrophils with an increased NOX2 activity, as observed in myeloproliferative disorder patients with a gain of function mutation in Janus Kinase 2 (JAK2) protein, have excessive cytoplasmic ROS and are more prone to form NETs upon stimulation.²⁴ On the other hand, stimuli such as calcium ionophores (A23187, ionomycin), nigericin, certain microbes, UV light, and some crystals, require mitochondrial reactive oxygen species (mROS) for NET formation (NOX2-independent NET formation).²⁵ ROS appear to be essential for NET formation, whether production is mediated by NOX or mitochondria.

Figure 1: Process of NET formation.

Different stimuli activate distinctive intracellular pathways that lead to NET formation. Stimuli may result in ROS production through activation of NADPH-oxidase via the Raf-MEK-ERK pathway. Increases in intracellular calcium leads to PAD4 activation. In the nucleus, PAD4 citrullinates histones promoting chromatin decondensation. Cytosolic or mitochondrial ROS production allows the release of neutrophil elastase (NE) and myeloperoxidase (MPO) from neutrophil granules. MPO and NE migrate to the nucleus, also contributing to chromatin decondensation. Several stimuli lead to the NLRP3 inflammasome pathway, potentially causing caspase-1 activation, while intracellular LPS and bacteria activate caspase-11. NE and caspases-1 and −11, cleave/activate gasdermin D (GSDMD) that forms pores in both nuclear and plasma membranes.

Decondensation of chromatin and nuclear swelling, together with GSDMD pores, allow rupture of the nuclear membrane and release of the chromatin in the cytosol, resulting in decoration of chromatin with neutrophil proteins. Subsequently, plasma membrane ruptures at one or more sites and NETs, decorated with neutrophil proteins such as PAD4, MPO, NE, LL37 and Histones, are released.

Cytoplasmic granule dissolution and activation of neutrophil proteases is a second important step during NET formation. Neutrophil elastase (NE) is released from neutrophil granules (azurophilic granules) through a complex pathway, involving ROS and hydrogen peroxide production (NOX2-dependent pathway).²⁶ This pathway activates myeloperoxidase (MPO) facilitating the release of active NE form neutrophil granules. Although MPO is important for NE release and activity, inhibition of MPO only delays the process of NET formation.²⁷ Once in the cytoplasm, NE, and probably other proteases, relocate to the nucleus where they cleave histones, likely facilitating nuclear decondensation.²⁸ MPO also binds chromatin and is released together with NETs. However, NE is not always required for NET formation. For example, in NE knockout mice, NET formation, upon stimulation with PMA or platelet activating factor, was not impaired,²⁹ suggesting that other proteases can compensate for its loss. The increase in NET formation observed in mice deficient in the serine protease inhibitor Serpin B1 further substantiates the role of other proteases in NET formation and chromatin decondensation.^{30, 31}

Chromatin decondensation is also induced by histone post-translational modifications (PTMs). During the early stages of NET formation, the increase in intracellular calcium activates the enzyme peptidylarginine deiminase 4 (PAD4).⁶ In the nucleus, PAD4 citrullinates histones, converting specific positively charged arginine residues to neutral citrulline residues. The overall reduction in the positive charge of histones reduces the forces that hold nucleosomes to negatively charged DNA leading to chromatin decondensation. Histone citrullination by PAD4 is a marker of ongoing NET formation and is considered by many to be important for chromatin decondensation.⁶ However, for stimuli such as PMA, histone citrullination has not always been observed. This could be due to the elevated cleavage of histones by neutrophil proteases.^{32, 33} Just as not all NET formation pathways proceed via NAPDH oxidase or neutrophil elastase, it has been proposed that NET formation may also occur independently of PAD4.³⁴ There are clearly different intracellular pathways that lead to NET release, and a recent study shows that several NET formation pathways induced by physiological stimuli proceed via PAD4.³⁵ However, further studies are needed to clarify the process of chromatin decondensation.

Other PTMs of histones might occur during the process of NET formation and facilitate chromatin decondensation. Actually, both acetylation and methylation of histones have been identified during NET formation;³⁶ however, the exact role of such PTMs is still unclear. Carbamylation of histones by MPO is another PTM that could favor chromatin unravelling and NET formation. Although H1 carbamylation has been observed in the plasma of rheumatoid arthritis patients,³⁷ it is still unknown if it occurs during NET formation.

In terms of cellular pathways to NET formation, recent observations using neutrophil-like cells (DMSO-differentiated HL60 cells; dHL60) suggest that within minutes after neutrophil stimulation with ionophore or extracellular LPS, the cytoskeleton disassembles and plasma membrane vesiculation and vesicle shedding occurs followed by microtubule dissolution.³⁸ In this initial step, neutrophil enzymes, such as MPO, NE, and PAD4, often play a predominant role. Use of enzyme inhibitors upon neutrophil stimulation reduces NET release.³⁹ Chromatin expansion is an important step in the second phase of NET formation. It is believed that the swelling process causes nuclear rupture. Tears in the lamin B1 layer and distribution of lamin B1 in the cytoplasm coincide with nuclear envelope rupture.³⁹ Interestingly, deficiency in the PAD4 enzyme in dHL60 has been shown to impede nuclear envelope and lamin rupture,³⁸ suggesting a role for PAD4 in these processes.

The third phase of NET formation is characterized by membrane rupture and NET release (Figure 1). Recent reports have shown that gasdermin D (GSDMD) cleavage, resulting in membrane pore assembly, is crucial for NET release.⁴⁰ NE and, to a lesser extent, other neutrophil proteases, such as proteinase 3 and cathepsin G, can cleave GSDMD.⁴⁰ The N-terminal GSDMD fragment p30 forms pores in neutrophil granules (a loop mechanism that further increases NE in the cytoplasm and, consequently, GSDMD fragment p30) and plasma membranes.⁴¹ This allows membrane rupture and the subsequent release of NET structures. Recent studies indicate that GSDMD p30 can also affect cellular membranes. In fact, nuclear permeabilization can be achieved by incubation of the nuclei of murine neutrophils with GSDMD p30-fragment.⁴¹ It remains to be seen whether this occurs in vivo.

Cytosolic LPS and Gram-negative bacteria induce NETs through a caspase-11 dependent mechanism. Mice deficient in caspase 1/11 make less NETs compared to wild-type mice.⁴² Direct activation of caspase-11 by cytosolic bacteria triggers a non-canonical inflammasome signaling and leads to GSDMD cleavage, GSDMD-p30 pores, and NET formation. Although PAD4 is not essential in such pathways, histone citrullination seems to always accompany NET release during caspase-11 dependent NET formation.⁴¹

Our group has recently observed that nigericin or calcium ionophore induces NLRP3 inflammasome/speck assembly, and this promotes NETs. Inhibition of NLRP3 inflammasome signaling significantly reduces NET release (Münzer P., Wagner D.D unpublished, 2019). NLRP3 canonical inflammasome activation by nigericin in neutrophils triggers activation of caspase-1 and secretion of IL-1β but does not induce cell death through pyroptosis.⁴¹ Other than promoting the process of NET formation, PAD4-mediated citrullination can intrinsically regulate the expression of IL-1β and TNFα by citrullinating NF-κB p65, thus increasing its nuclear translocation.⁴³ This suggests a tight connection between NLRP3 inflammasome and NETs. This canonical pathway of NLRP3-mediated caspase-1 activation could potentially also lead to GSDMD pore formation. Further studies are required to better define the inflammasome pathway during NET formation.

The majority of the stimuli tested in vitro that induce NET formation cause rupture and lysis of the cell (so called lytic-NETs). However, it has been suggested that decondensed chromatin is also released through vesicular transport, leaving the rest of the cell intact, and thus allowing the formation of functional enucleated neutrophils called cytoplasts. This rapid (5–60 minutes) non-lytic NET release is induced under infectious conditions and requires both TLR2 and complement component 3 in mice.⁴⁴ Enucleated cytoplasts have also recently been found in vivo in the lungs and lymph nodes of asthmatic mice and in bronchoalveolar fluids collected from severe asthmatic patients.⁴⁵ These cytoplasts were associated with the activation of dendritic cells to differentiate naive CD4⁺ T cells into T helper 17 (T_H17) effector cells. This demonstrates that after NET formation, the forming cytoplasts may fulfill biological functions. Future studies are needed to determine whether cytoplasts form and function in normal tissue repair and non-infectious disease states.

NETs in autoimmunity and immune-mediated thrombosis

Aberrant NET formation and degradation is observed in several autoimmune diseases,⁴⁶ such as anti-neutrophil cytoplasmic antibodies (ANCA)-associated vasculitis (AAV), systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), and antiphospholipid syndrome (APS).

In autoimmune disorders, NETs cause tissue, organ, and vessel damage, leading to severe inflammatory and thrombotic states associated with the condition. They are also considered an important link between innate and adaptive immune response. NETs play two central roles here: release of inflammatory molecules able to propagate immune responses by stimulating cytokine release from other immune cells, like macrophages, and facilitation of auto-antibody production by exposing self-antigens or modified auto-antigens.⁴ Indeed, NETs are a source of citrullinated proteins that contribute to the production of anti-citrullinated-protein/peptide antibodies (ACPA) in RA patients.⁴⁷

It has been suggested that some individuals are more predisposed to develop autoimmune responses. A particular subset of neutrophils, called low density granulocytes (LDG), has been found in circulation in both SLE and RA patients.⁴⁸ LDGs are more prone to NET formation and are capable of stimulating cytokine release by macrophages, thus propagating inflammatory responses. Both ANCAs and ACPAs stimulate neutrophils for NET formation^{49, 50} inciting a vicious circle that leads to inflammation and, through release of neo-antigens, increased production of auto-antibodies.

Single nucleotide polymorphisms (SNPs) in PADI4 gene have also been associated with autoimmunity. For example, in RA, PADI4 SNPs are linked to susceptibility to the disease.^{51, 52} SNPs in other neutrophil proteins, such as protein tyrosine phosphatase, have also been shown to influence PAD4 expression in RA patients resulting in an increase of NET formation and citrullinated auto-antibodies.⁵³

Recent studies have provided mechanistic insights into how NETs promote thrombosis in some autoimmune disorders. In APS animal models, binding of antiphospholipid antibodies to neutrophils contributes to thrombosis.⁵⁴ Antibodies isolated from APS patients promote NET formation in APS neutrophils, which also spontaneously release NETs. Interestingly, APS patient antibodies could activate NET formation in vitro also in control neutrophils, indicating that this is not an intrinsic neutrophil effect in APS patients.⁵⁵ It was recently reported that agonism of the adenosine A_2A receptor strongly inhibits NET formation via cell-intrinsic production of cyclic AMP,⁵⁶ which was previously shown in infection to regulate NET formation.^{57, 58} The mechanism by which this occurs is a decrease in production of ROS. Therefore, adenosine receptor agonism could be a novel therapeutic approach not only in APS patients but also in other contexts of thrombosis where NET formation is ROS-driven.

Another thrombotic state induced by the formation of autoantibodies occurs in immune-mediated heparin-induced thrombocytopenia (HIT).⁵⁹ Antibodies formed against heparin-PF4 complexes, which in turn activate platelets in the bloodstream, lead to thrombocytopenia and microvascular occlusions.⁶⁰ Two recent studies have highlighted the importance of NETs in hybrid humanized–mouse models of HIT-induced thrombosis.^{61, 62} The first study showed induction of NETs by experimental HIT antibodies.⁶¹ PF4 binding to NETs makes them more resistant to DNase I cleavage. NETs promoted thrombus development, and the HIT milieu increased NET-dependent thrombus stability. In the second study, anti-heparin-PF4 antibodies, either experimental or from HIT patients, were shown to directly induce NET formation via FcRγIIA on neutrophils and via platelet interactions.⁶² This promoted thrombus development in a neutrophil-dependent manner. Compared to control patients, elevated NETs and NET biomarkers were found in HIT-patients along with the presence of LDGs. Of note, in both studies, deletion or inhibition of PAD4 was protective against thrombus formation.

Neutrophil infiltration and anti-neutrophil cytoplasmic antibodies (ANCAs) are also found in patients affected by vasculitis and are considered the primary cause of inflammation and vessel damage in this condition.⁴⁹ Initially, necrosis of neutrophils surrounding the vessel wall was considered the predominant cause of vessel damage and the release of neutrophil proteins, such as MPO and PR3, major targets of ANCAs in ANCA-associated vasculitis (AAV) patients. However, more recent findings have demonstrated the presence of NETs in thrombi,^{63, 64} in the plasma of AAV patients and at sites of vasculitic lesions,⁶⁵ and have been associated with severity of the disease. In AAV patients, NETs directly cause vessel damage and thrombosis, one of the most common complications of AAV. NETs directly induce endothelial damage and participate in thrombus formation. Interestingly, tissue factor–associated NETs have been found in AAV patients and are suggested to induce the thrombo-inflammatory state associated with the disease.⁶⁶ ANCAs stimulate NET formation leading, in this way, to a vicious circle.

In addition to elevated production of NETs, defects in NET clearance are also observed. Not only do patients present with low DNase I activity, but anti-NET antibodies have been found in AAV patients and are suggested to prevent access of the DNase I enzyme to the extracellular chromatin structure.^{67, 68} Anti-NET antibodies are also observed in other autoimmune disorders where NETs play a predominant role, such as SLE and RA.⁶⁹

Although inhibition of NET formation could be considered a worthwhile therapeutic target to prevent tissue and organ damage in autoimmune disorders, additional studies are still needed to analyze the effect of NET inhibitors on the crosstalk between innate and adaptive immune cells.

NETs in thrombosis

Platelets and neutrophils are quickly recruited by von Willebrand factor (VWF) to a site of vessel injury or endothelial activation.⁷⁰ In the context of normal hemostasis and immunity, this rapid activation and recruitment is advantageous because it protects against uncontrolled bleeding and microbial invasion. However, when this response is dysregulated, pathological thrombus formation may occur. In recent years, the important contribution of neutrophils, particularly via NETs, to this thrombus pathogenesis has been increasingly recognized.^71–73 NETs provide a physical scaffold for thrombus growth by binding platelets and red blood cells (RBCs). This was demonstrated by Fuchs et al. in 2010, where NETs perfused with platelet-rich plasma resulted in the adhesion and aggregation of platelets and binding of RBCs, which promoted blood clot formation. Platelets interact with NETs through DNA-histone complexes or by binding to plasma proteins, such as VWF and fibronectin bound to NETs.⁷⁴

Central to the structure of a NET is its decondensed chromatin backbone,^{3, 75} consisting of nucleosomes: DNA coiled around a histone core. Even before NETs were linked to thrombosis, nucleic acids and nuclear components were shown to induce coagulation.⁷⁶ In sepsis patients, cell-free DNA in plasma, primarily originating from neutrophils, correlates with increased thrombin generation.⁷⁷ Extracellular DNA also amplifies thrombin-dependent factor XI activation initiated by tissue factor.⁷⁸ Histones, one of the main protein components of NETs,⁷⁹ are highly cytotoxic to endothelial and epithelial cells and have been shown to contribute to mortality in sepsis.⁸⁰ The same histones that are particularly cytotoxic, namely histones H3 and H4, are also able to induce platelet aggregation,⁷⁴ either directly by binding of histones to platelet TLR2 and 4,⁸¹ or indirectly via fibrinogen.⁸² This histone-mediated platelet activation promotes thrombin generation⁸³ by preventing protein C activation via thrombomodulin.²¹ This is not solely an in vitro effect, as in vivo, histones in circulation induce profound thrombocytopenia.⁸² In addition to individual DNA and histone components, intact NETs themselves can also promote thrombin generation, via platelets and factor XI/XII.⁷⁷ However, complete NETs were shown to be less procoagulant than their separate components individually, particularly DNA and histones.⁷⁸ Moreover, NETs bind factor XII, and with the cooperation of platelets, they support its activation to factor XIIa, presumably through its negative charge.⁸⁴ Neutrophils can in fact themselves also be a source of factor XII.⁸⁵ Therefore, NETs can promote both the intrinsic and extrinsic coagulation pathways.⁸⁶ Interestingly, neutrophil-derived FXII can promote NET release by urokinase plasminogen activator receptor-mediated Akt2 phosphorylation, indicating a broader role for FXII beyond promoting coagulation.⁸⁵

In addition to promoting thrombin generation, NETs can also directly bind fibrinogen and, thus, provide a scaffold for fibrin deposition.⁷⁴ Moreover, the serine proteases neutrophil elastase and cathepsin G, also released on NETs,⁷⁹ help thrombin to promote fibrin formation in the presence of platelets. This occurs locally within a developing arterial thrombus, via the ability of these neutrophil serine proteases to degrade the platelet-derived tissue factor pathway inhibitor (TFPI), enhancing tissue factor and factor XII-driven coagulation.⁸⁷ This enhanced coagulation also represents a mechanism by which NETs can control local bacterial infections, as both NETs and fibrin can trap bacteria and prevent their spreading from a local wound^{44, 88, 89} or their dissemination in sepsis.⁹⁰ Of note, in systemic bacteremia, NET-deficiency did not have an impact on bacteremia,^{91, 92} indicating that this procoagulant effect is likely not sufficient to contain pathogens in conditions of high bacterial burden. Also, in large vein thrombosis, neutrophil elastase deficiency did not have impact on thrombus formation.²⁹

Animal models have provided a wealth of evidence of a role for NETs, both in arterial and in venous thrombosis,^{84, 87, 93} and in the decade since NETs were first described in thrombosis, there have been several key insights into the mechanisms by which NETs form during thrombosis. The interaction between platelets and neutrophils, mediated by P-selectin/PSGL-1, has been shown to be important in NET formation.¹⁵ PSGL-1 signaling during the process of NET formation occurs with dimerized P-selectin,⁹⁴ and in experimental DVT PSGL-1 together with CXCR2 signaling promotes thrombus formation via NETs.⁹⁵. Platelets can enhance NET formation also by means of their secretome, such as through the release of damage associated molecular pattern (DAMP) high-mobility group box 1 protein (HMGB1).⁹⁶ This is dependent on the oxidation of HMGB1 to disulfide HMGB1, which occurs after its release from platelets.⁹⁷ Although HMGB1 can be produced by leukocytes, during thrombosis the majority of HMGB1 deposition responsible for enhanced NET formation originates from platelets.^{98, 99} Furthermore, cathelicidins released on NETs, such as LL-37, also promote platelet activation and subsequent interaction with neutrophils in a P-selectin-dependent manner.¹⁰⁰ Thus, released NETs bind platelets and promote their activation, while the activated platelets also promote NET formation, which then trap additional platelets. This creates a vicious cycle promoting pathological thrombus development, also seen in conditions of sepsis.¹⁰¹ In addition to platelets, new insights into the role of other cell types, including mast cells¹⁰² and NK cells,¹⁰³ in promoting venous thrombosis have recently been described. NK cell-derived interferon γ promotes NET formation, and it remains to be determined if mast cells also influence NET formation in venous thrombosis.

The role of oxidative stress on NET formation during thrombosis has recently been investigated. Deficiency of Sirtuin 3, a NAD+-dependent protein deacetylase which acts as a negative regulator of ROS production, results in increased intracellular ROS generation in both platelets and neutrophils.¹⁰⁴ This enhanced ROS production did not impact venous thrombus development in an experimental DVT model using young, healthy mice. However, in an arterial thrombosis model, which involved prior induction of an inflammatory response with injection of LPS, Sirt3-deficiency resulted in accelerated thrombus formation.¹⁰⁵ Therefore, oxidative stress-dependent NET generation may depend on the thrombotic environment and the type of vascular bed.

Many reports now show NETs also in thrombi from patients, both in venous thromboembolism and in arterial thrombi.¹⁰⁶ Thrombi harvested from venous thromboembolism patients, mainly post-mortem, had NETs primarily in organizing regions rather than in fresh or organized areas of the clots.¹⁰⁷ Advances in the treatment of arterial thrombosis by mechanical thrombectomy has opened the door to the study of thrombus composition in stent thrombosis,¹⁰⁸ myocardial infarction,^{109, 110} and in stroke,^111–113 where a majority of thrombi contain NETs. Clearly, NETs have a large impact on thrombotic disease in patients. Therefore, better understanding of the triggers of the process of NET formation, and also the regulation of NET induction or of released NETs, has important therapeutic implications. Endogenous DNases present in plasma likely further regulate NET formation by degrading NETs released in the bloodstream.¹¹⁴ In DNase I and DNase I/13-double deficient animals, fibrin-poor, DNA-rich occlusive clots formed in septic mice and, spontaneously, in neutrophilic mice.¹¹⁴ A reduction of DNase activity in bacteremia patients, such as that seen in E. coli-associated hemolytic uremic syndrome,^{115, 116} is thus likely to promote thrombus formation.

NETs and atherosclerosis

An increasing body of clinical and experimental evidence suggests a role for NETs in atherosclerosis. The first indication of this came from a clinical study showing that circulating nucleosomes and plasma MPO/DNA-complex levels in patients with severe or extremely calcified coronary artery disease (CAD) are significantly increased compared to the respective parameters in healthy individuals or patients without CAD.¹¹⁷ Furthermore, this study has also established that levels of circulating nucleosomes, dsDNA, MPO/DNA-complexes, TAT complexes, and VWF positively correlate with the number of diseased coronary artery segments, and that the concentration of nucleosomes and VWF may independently predict the presence of CAD.¹¹⁷ Interestingly, neutrophils isolated from the blood from infarct-related coronary artery areas are more susceptible to NET release than blood neutrophils selectively sampled from non-infarct-related vessel segments in patients with ST-segment elevation acute myocardial infarction.¹¹⁸ This could indicate that interactions with the diseased vessel wall may prime neutrophils for NET release.

Studies in ApoE-deficient mice have shown the presence of NETs in atherosclerotic lesions of the aortic root as early as three weeks after initiation of a high fat diet (HFD).¹¹⁹ This indicates that NETs form early in the disease process. Apoe^−/− mice, which in addition lack enzymes critical for NET formation (the serine proteases neutrophil elastase and proteinase 3¹¹ or PAD4¹¹⁹), have a reduced atherosclerotic plaque size after eight to ten weeks of HFD.^{11, 119} Similar results were observed in ApoE-deficient mice on HFD treated with DNase I¹¹⁹ or the pan-PAD inhibitor Cl-amidine.¹²⁰ In the Ldlr^−/− murine atherosclerosis model, PAD4-deficiency in hematopoietic cells had little effect on atherosclerotic plaque formation¹²¹ but was associated with reduced endothelium denudation, similarly to results obtained in DNase I-treated Apoe^−/− mice.¹²¹ This experimental data suggests that NETs have an important function in promoting superficial erosion that exposes thrombogenic material in plaques, rendering them more dangerous.

In line with these findings, NET-releasing neutrophils in both mouse and human atherosclerotic lesions are predominantly localized in the smooth muscle cell-rich fibrous cup. Histone H4 externalized on NETs may cause lysis of smooth muscle cells and thus contributes to plaque vulnerability.¹²²

Cholesterol uptake by macrophages can induce inflammasome activation.¹²³ In addition, cholesterol crystals present in atherosclerotic lesions may trigger NET release, which, in turn, can induce cytokine production in macrophages (e.g. of pro-inflammatory IL-1β) that promotes further neutrophil recruitment to inflammatory sites,¹¹ suggesting that NETs may support an important crosstalk between immune cells in chronic inflammation.

Atherosclerosis, hypercholesterolemia, and hypertension are also risk factors for development of abdominal aortic aneurysm (AAA). Immune cells, in particular neutrophils, are thought to play an important role in the early phase of AAA formation,¹²⁴ and recent studies support the notion that the neutrophilic IL-1β signaling pathway essentially contributes to disease progression.¹²⁵ NETs are present in the adventitia of human AAA.¹²⁶ In an elastase perfusion-mediated mouse model of AAA, NETs were also detected in the adventitia of aorta as early as two days,¹²⁷ and in the intima three days post initiation of elastase treatment.¹²⁵ Cl-amidine treatment mitigates AAA formation,¹²⁵ as does DNase I treatment.¹²⁷ In line with these reports, AAA progression was attenuated in genetically engineered mice deficient for all three neutrophil serine proteases (neutrophil elastase, cathepsin G, proteinase 3), whose neutrophils exhibited defective NET formation.¹²⁷ These experimental data suggest that inhibition of NET formation might represent a promising novel therapeutic approach in patients with AAA,¹²⁸ a potentially life-threatening disease condition for which surgical interventions currently represent the single effective treatment but are also associated with a significantly elevated risk of systemic inflammation and postoperative mortality.

NETs and VWF in ischemia-reperfusion injury and development of fibrosis

Neutrophils, platelets and inflammatory cytokines have been implicated in ischemia/reperfusion (I/R) injuries, such as myocardial infarction, stroke, and peripheral vascular disease. In the heart, a thrombotic occlusion in a major cardiac artery, or a series of microvascular occlusions resulting from an injurious event, cause extensive cell death within the myocardial tissue. The subsequent reperfusion of the vessel activates the endothelium downstream propagating the injury (Figure 2).

Figure 2. VWF and NETs released during ischemia/reperfusion injury contribute to occlusive thrombus formation in the microcirculation and to tissue fibrosis.

The relative roles of VWF and NETs to the complex process of fibrosis development are depicted in the following schematic: (A) Ischemia leads to the rapid release of storage granules containing P-selectin and UL-VWF by endothelial cells that, by self-aggregation, produce extremely long strings (red) which bind platelets (yellow). VWF is also released basolaterally into the underlying tissue (light pink). (B) P-selectin and VWF recruit neutrophils that, under the effect of hypoxia and/or platelet activation, readily release NETs. The large DNA strands (blue) covered with histones bind to VWF polymers through specific interactions, forming an extensive scaffold. VWF and P-selectin promote leukocyte activation and transmigration into the surrounding tissue. (C) Recruitment of additional platelets by NETs produces an occlusive clot that is rich in both VWF and DNA. This propagates ischemic response, while the transmigrated neutrophils produce NETs in the tissue. NETs stimulate other cells, such as fibroblasts (green), to release collagen and other extracellular matrix molecules in response to the perceived injury. Resolution of the clot proceeds by plasma and cellular release of DNases, cleaving the DNA scaffold, and ADAMTS13 fragmenting VWF. ADAMTS13 also detaches the scaffold from the vessel wall. (D) Clot resolution restores blood flow. The endothelium recovers its anti-thrombotic character; however, the fibrotic scar (dark green) persists with possible consequences on organ function.

The damaged heart provides an ideal environment for NET formation: elevated proinflammatory cytokines,¹²⁹ local hypoxia,¹³⁰ ROS,⁷⁵ and release of DAMPs, such as mitochondrial DNA¹³¹ and HMGB1.⁹⁶ All of these have been shown to promote NET formation (Figure 1) either directly, by activating neutrophils, or indirectly, via other cells, such as platelets.^{97, 98} Platelets are crucial drivers of NET formation in thrombo-inflammation during cardiac injury. Importantly, in the aortic constriction model of cardiac fibrosis, either NET-deficiency or excessive cleavage of VWF by delivery of the proteolytic enzyme ADAMTS13 prevented platelet aggregate accumulation within the damaged vasculature of the heart.^{132, 133} Platelets were previously shown to be the main source of TGFβ, which promotes collagen deposition during cardiac remodeling.¹³⁴ In addition, platelets and their secretome propagate inflammation.^135–137 Both VWF and NETs recruit platelets, and it is therefore not surprising that they are implicated in fibrosis. In addition, VWF released from Weibel-Palade bodies at sites of endothelial activation might bind and attach NETs to the vessel wall (Figure 2).

In a striking parallel to ultra large VWF (UL-VWF) released from endothelium by hypoxia,¹³⁰ neutrophils release their chromatin in response to low oxygen levels.^{138, 139} Hypoxia induces HIF-1α expression,¹⁴⁰ and translation of HIF-1α, usually regulated by mTOR, promotes NET formation.¹⁸ Indeed hypoxia, such as that encountered by mountaineers at high altitudes, may induce thrombosis and promotes DVT in mice,¹⁴¹ a pathological process requiring both VWF and NETs.^{93, 142, 143}

The release of NETs is highly injurious to the surrounding tissue, resulting in direct cytotoxicity on endothelial cells via histones.^{80, 144} As mentioned previously, histones are also prothrombotic, promoting thrombin generation⁸³ and platelet aggregation⁷⁴ and enhancing the stability of fibrin clots.¹⁴⁵ Cardiac fibrosis results from excessive deposition of extracellular matrix components, such as collagen, in the cardiac muscle, making it stiffer.¹⁴⁶ The degree of fibrotic tissue is correlated to a decline in heart function both in mouse models and in patients.^{132, 147} Fibrosis can occur either perivascularly or within the interstitium of the myocardium. NETs and neutrophil IL-17 release were shown in vitro to promote differentiation of lung fibroblasts to collagen-producing myofibroblasts.¹⁴⁸ Whether this also occurs within the heart remains to be determined.

The importance of VWF and PAD4 has been addressed in murine models with resultant cardiac fibrosis, and the possible therapeutic value of ADAMTS13 and DNase I have been evaluated for their role in diminishing the fibrotic process. VWF release and VWF-dependent platelet recruitment have been observed early after induction of pressure-overload injury, along with neutrophil infiltration.¹³³ Treatment of wild-type mice at the onset of ascending aortic constriction (AAC) injury with therapeutic doses of ADAMTS13 reduced VWF and platelet aggregates in the myocardial tissue. ADAMTS13 was shown to also lower plasma levels of active TGF- β1. Interestingly, both ADAMTS13 and DNase I treatment have been shown to reduce left ventricular collagen content after AAC and improve systolic function.^{132, 133} Parallels can be drawn regarding the influence of NETs on fibrotic remodeling in the injured heart and scar tissue formation in the skin. Slow-healing wounds deposit more scar tissue.¹⁴⁹ NETs delay wound repair, especially in situations where their production is excessive, such as in diabetes.¹⁵⁰

Also in spontaneous age-related organ fibrosis development, mice defective in releasing NETs due to PAD4-deficiency were protected from detrimental collagen deposition in both the heart and lung.¹³² Strikingly, a decline in heart function was present in wild-type aged mice, while aged PAD4-deficient mice maintained systolic and diastolic function similar to that of young mice. PAD4 is not only important with respect to NET formation, but it is also a protein-modifying enzyme, converting positive arginine residues to neutral citrulline¹⁵¹ and is expressed, to a lesser extent, by other cells besides neutrophils. Therefore, the effect of PAD4 could extend beyond its role in NET formation. This possibility remains to be investigated using tissue-specific PAD4-deficient mice. Cardiac tissue damage likely promotes additional local NET formation and further leukocyte recruitment. Recently, the chemokine midkine was shown to be a crucial player in the development of fibrosis in autoimmune myocarditis leading to heart failure¹⁵² by promoting neutrophil and CD4⁺ T cell recruitment, with a presence of NETs observed both in the autoimmune myocarditis animal model and in the patient cardiac tissue.¹⁵²

NETs also have detrimental implications in fibrotic remodeling after myocardial infarction. Both DNase I administration and PAD4-deficiency were similarly protective against myocardial I/R-injury, limiting neutrophil infiltration into the infarcted area.¹⁵³ DNase I treatment also provided a long-term benefit in LV remodeling in a rat MI/R model,¹⁵⁴ as determined by echocardiography. NETs were identified within the culprit thrombus in MI patients,¹⁰⁹ and NET thrombotic burden positively correlates with infarct size.¹⁵⁵ This is not limited to the heart. DNase I treatment improves I/R outcome after stroke, and recent papers demonstrate that NETs have a detrimental effect also in kidney,^156–158 hepatic,¹⁵⁹ and intestinal I/R injury.^{160, 161}

Here, again, we can draw parallels between NETs and VWF. Several mouse studies have provided solid experimental evidence for a protective role of ADAMTS13 in MI/R injury. Studies show larger infarcts in Adamts13^−/− mice than in wild-type mice, and a reduction of infarct size in wild-type mice treated systemically with a supraphysiological, anti-inflammatory dose of ADAMTS13. These effects were fully dependent on the presence of VWF, as a reduction in infarct size by ADAMTS13 treatment was not observed in VWF-deficient mice.^162–164 Very similar results were obtained in stroke I/R models. Again, ADAMTS13 treatment resulted in better behavioral outcomes and smaller infarcts,^165–168 and improvement was also obtained with DNase I treatment.¹⁶² The results with I/R further substantiate the existence of a relationship between VWF and NETs, as depicted in Figure 2, and the possibility to combine anti-NET with anti-VWF therapy in the future.

NETs in pulmonary disease

In recent years, NETs have been recognized as an underlying pathomechanism of multiple lung diseases, including cystic fibrosis (CF), neutrophilic asthma, and transfusion-related acute lung injury (TRALI), and as a key contributing factor in complications linked to lung transplant. The lung and its function of blood oxygenation seem to be particularly threatened by NET formation in the alveoli. Acute NET-deposition there may cause death like in TRALI^{169, 170} and in influenza when combined with secondary bacterial lung infection.⁴²

CF is an autosomal recessive disorder caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene.¹⁷¹ Hallmarks of the disease include the presence of thick mucus obstructing the airways and facilitating bacterial colonization. Neutrophils are the main immune cell type in the lungs of CF patients; however, they are unable to efficiently remove invading pathogens. This leads to frequent lung infections associated with excessive neutrophil recruitment to the lungs, which eventually may result in chronic inflammation and further decline in function.¹⁷² NETs are frequently observed in CF patients and are responsible for the majority of extracellular DNA in CF sputum.^173–176 Both NETs and concomitant bacterial colonization are injurious to the lung tissue. Nebulized recombinant human (rh) DNase I (Pulmozyme) is the mainstay for treatment of CF patients. Importantly, in vitro studies suggested that DNase I treatment may liberate neutrophil proteases that can exert deleterious proteolytic effects on lung tissue,¹⁷⁴ whereas DNA-bound NE may improve sputum solubilization.¹⁷³ Thus, the therapeutic potential of agents modulating the activity of neutrophil serine proteases requires further investigation.

Extracellular traps originating from both neutrophils and eosinophils have been detected in bronchial biopsies from asthma patients.¹⁷⁷ The sputum of neutrophilic asthmatics displayed higher levels of extracellular DNA and LL-37 compared to a non-neutrophilic asthma patient population, and functional studies revealed that lung function was inversely correlated to NET marker levels, suggesting that NET formation exacerbates asthma severity.¹⁷⁸ Likewise, patients with a low asthma control test score indicative of increased disease severity presented with high levels of extracellular DNA which were linked to increased airway inflammation, elevated concentrations of NET components, and interestingly, markers of inflammasome activation (caspase-1 and IL-1β).¹⁷⁹ NETs induce cytotoxicity in airway epithelial cells that can be reversed by DNase I treatment.¹⁷⁹ A recent study described the presence of enucleated neutrophil cytoplasts in the bronchoalveolar lavage fluid of patients with severe neutrophilic asthma.⁴⁵ Likewise, cytoplasts were detectable in the lungs of mice concomitantly exposed to house dust mite (allergen) and LPS (endotoxin), which led to NET formation.⁴⁵ Unlike DNase I treatment, PAD4-deficiency ameliorated allergen-induced airway neutrophilia and cytoplast generation.⁴⁵ Notably, neutrophil-derived cytoplasts were found in mediastinal lymph nodes, and a distinct novel function innate to cytoplasts—and not neutrophils—in the initiation of adaptive immune responses has been identified.⁴⁵

TRALI is a rare but serious complication that may occur during blood transfusions and can lead to transfusion-related mortality. Histologically, TRALI is characterized by the presence of lung edema and substantial neutrophil infiltration.¹⁸⁰ Some severe cases of TRALI have been attributed to the presence of anti-neutrophil antibodies in the transfused products.¹⁸¹ Serum levels of circulating DNA, nucleosomes, and MPO are significantly elevated in TRALI patients.^{169, 170} NETs were detected in the lungs of mice with TRALI, and intranasal DNase I treatment increased arterial blood oxygenation¹⁶⁹ and survival,¹⁷⁰ suggesting that, similarly to CF, targeting extracellular DNA might be a promising treatment strategy in acute lung injury. Of note, the route of administration (inhalation vs. intravenous treatment) may have important consequences on therapeutic efficacy, and targeted delivery to the site of excessive NET release might be superior to systemic administration.¹⁶⁹

The role of NETs in pathogen-induced lung injury is likely more complicated, as recently demonstrated in experimental studies. Lung injury was reduced in mice with impaired NET formation (Padi4^−/−) upon methicillin-resistant S. aureus lung infection; however, overall survival rates were unaltered between wild-type and Padi4^−/− mice.¹⁸² By contrast, DNase I treatment initiated after disease onset both dampened lung injury and reduced pathogen-induced mortality,¹⁸² demonstrating that treatment approaches may vary in their efficacy depending on the disease settings.

Primary graft dysfunction after lung transplantation is a form of acute lung injury characterized by massive leukocyte infiltration. The organ injury can be counteracted by neutrophil depletion.^{183, 184} NETs have been identified in the bronchoalveolar lavage fluid of patients with primary graft dysfunction and in the lungs of mice subjected to an experimental model of the disease.¹⁸⁵ Disrupting NETs with DNase I decreased protein leakage in the lungs, reduced the number of neutrophils present in the bronchoalveolar lavage fluid, and improved lung function at early time points after transplantation in a mouse model recapitulating key features of human primary graft dysfunction after lung transplantation.¹⁸⁵ Recent work utilizing intravital microscopy in a mouse model of transplant-induced I/R injury and an elegant set of in vitro experiments shed light on the consequences of NET fragment generation at later time points after DNase I treatment.¹⁸⁶ DNase-generated NET fragments have been identified as key contributing factors in allograft rejections and have been shown to play an important role in the induction of pro-inflammatory cytokine production by macrophages and in dendritic cell maturation and T cell proliferation.¹⁸⁶ Based on these early observations, further investigations are clearly warranted; however, a growing body of experimental data provides compelling evidence that different strategies aimed to interfere with NET formation and degradation might be beneficial depending on the specific clinical settings.

NETs in cancer progression and cancer-associated thrombosis

Neutrophils play an important role in cancer biology, acting as major inflammatory cell infiltrates in several human tumor types and mouse cancer models, with both pro- and anti-tumor functions.¹⁸⁷ The role of neutrophils depends on their capacity to produce distinctive effector molecules and is highly influenced by the tumor microenvironment.^{187, 188} The antitumor activity of neutrophils is mainly due to the release of pro-inflammatory and immune-stimulatory molecules, like IL-12 and TNF-α, that induce T cell recruitment and support tumor suppression; in contrast, neutrophils, through oncostatin M, ROS production¹⁸⁹, MMPs, NE, and many other factors,¹⁹⁰ sustain tumor angiogenesis, invasion, and metastasis, and encourage tumor progression.¹⁸⁷

NETs, found in both human and mouse tumors,^{191, 192} are associated with poor prognosis.¹⁹¹ In fact, citrullinated histone H3, a hallmark of NET formation, is detected at elevated levels in the plasma of cancer patients when compared to healthy individuals and even to severely ill patients without cancer^{193, 194} and high levels may predict mortality in cancer patients.¹⁹⁵

Postsurgical infections are common complications that occur in cancer patients,¹⁹⁶ and severe infections, such as pneumonia and sepsis, are associated with an increased rate of death in metastatic patients. Initially, infection was considered a necessary hallmark to induce NET formation in cancer. Minor or severe systemic infections in tumor-bearing mice activate neutrophils and induce NET release, which was shown to promote metastasis in the animals.¹⁹⁷ Tumor progression and metastasis in mice was decreased after NET degradation or inhibition of NET formation.¹⁹⁷ However, it is now established that cancer creates a systemic environment that promotes NET release. In a leukemic mouse model and in solid tumors models, isolated neutrophils produced NETs more readily than control neutrophils.¹⁹⁸ Priming is mediated by cytokines released by the tumor cells, primarily granulocyte colony stimulating factor (G-CSF) that can accumulate in the blood of tumor-bearing mice.¹⁹⁸ Moreover, neutrophils from patients with myeloproliferative neoplasms (MPNs), characterized by a mutation constitutively activating JAK2 signaling, are also primed for NET formation, likely by enhancing synthesis of PAD4 in the cells.²⁴ Inhibition of constitutively active JAK2 abrogated NET formation and reduced thrombotic events associated with the mutation.²⁴

Stimulation of neutrophils towards NET formation by cancer-derived G-CSF has been shown to promote tumor growth. Cancer mouse models in which the tumor cells do not produce G-CSF, such as B16 melanoma, do not show a reduction in tumor growth when implanted in PAD4-deficient mice that do not produce NETs.¹⁹⁹ Other than supporting tumor growth, excessive NET structures produced in cancer are shown to sequester the circulating cancer cells favoring their adhesion, invasion, and migration through the endothelium,¹⁹⁷ allowing the subsequent growth of tumors in the distant organs and tissues.^{200, 201} Indeed, migration of neutrophils to pre-metastatic niches and subsequent NET formation allows entrapment of circulating tumor cells and leads to the formation of metastatic implants.^{200, 201} In ovarian, colon, and breast cancer mice models, depletion or inhibition of NET formation markedly decreased the number of tumor metastases.^{197, 200–202}

Recent studies show that not only is the DNA scaffold involved in tumor progression, but NET-associated proteases also play a fundamental role in the spread of cancer.²⁰³ NET release upon sustained inflammation allows remodeling of the extracellular matrix (ECM). ECM modification in cancer was shown to promote mesenchymal to epithelial transition and to drive tumor progression.²⁰³ Cleavage, changes in glycosylation, and crosslinking²⁰⁴ are well-known modifications that disorganize the ECM in cancer. Recently, NETs have been found to bind the ECM, allowing close association of ECM proteins, like laminin, with neutrophil enzymes, such as NE and matrix metalloproteinase 9.²⁰³ Both enzymes are able not only to promote ECM degradation, thereby favoring tumor metastasis, but also generate, through cleavage of laminin, an α₂β₃ integrin epitope that may awaken and induce proliferation of dormant cancer cells.²⁰³ Whether other neutrophil enzymes participate in ECM modification and remodelling is unknown. It has been shown that PAD4 is able to citrullinate and modify the ECM in human colorectal cancer liver metastasis. However, the extracellular PAD4 in this study was of cancer cell origin.²⁰⁵ It would be of interest to determine if in other metastatic cancers, PAD4 released with NETs has similar effects on the tumor microenvironment.

Interestingly, NETs not only promote tumor progression but can also have systemic effects, altering peripheral vessels and distant organs that are not targets for metastasis.^{206, 207} Reduction in vascular function has been observed in the kidney and heart of mice with cancer compared to healthy littermates. This phenotype was associated with the increase of NETting neutrophils.²⁰⁶ Indeed, depletion of neutrophils and NETs completely restored perfusion of the organs.²⁰⁶

Increased neutrophil count and biomarkers of NETs in cancer patients are also risk factors for cancer-associated thrombosis. Venous thrombosis is the second most common cause of cancer-related death;^{208, 209} however, it is now clear that excessive NET release in cancer also contributes to arterial thrombotic events in the brain and heart injury through arterial microthrombosis.²¹⁰ Increase of plasmatic chromatin and spontaneous lung microthrombi formation has also been observed in a murine mammary cancer model.¹⁹⁸ Simulation of the pro-thrombotic state seen in cancer in tumor-free mice was accomplished by G-CSF and low-dose LPS treatment, a mixture that induces NET release. This confirmed the important role of neutrophils and NETs in cancer-associated thrombosis.¹⁹⁸ NET-associated microthrombi and high circulating levels of G-CSF have also been found in patients with ischemic stroke with underlying cancer, further linking cancer, thrombosis, and NETs.²¹⁰

NETs promote cancer-associated thrombosis in several ways. The extracellular chromatin stimulates coagulation by the activation of platelets and coagulation factors, as described in the thrombosis section. Moreover, tumor-derived extracellular vesicles, found in cancer patients with a pro-thrombotic state, have been shown to stimulate NET formation²¹¹ and to bind to NETs²¹⁰ accelerating venous thrombosis in tumor-free mice.^{210, 211} Treatment with chloroquine, an autophagy inhibitor affecting NET formation, has been shown to reduce hypercoagulability and lower venous thrombosis in patients affected by pancreatic cancer.^{212, 213}

Although not much is known about the specific role of neutrophil enzymes in cancer-associated thrombosis, we have observed that increased levels of PAD4 in plasma of mice accelerate the time to vessel occlusion in a thrombosis model and inhibit, through citrullination, the antithrombotic enzyme ADAMTS13 that cleaves VWF.²¹⁴ Investigating the role of PAD4 and other NET-associated proteins during cancer propagation and thrombosis could allow the identification of novel therapeutic targets.

Positive and negative consequences of NETs: thoughts for the future

The capacity of neutrophils to release NETs in response to inflammatory stimuli is fundamental to protect the body against harmful invasion by pathogens. It has been shown that NETs capture bacteria and provide a barrier avoiding pathogens’ access to circulation.^{89, 90} NETs and their components are considered a beneficial “alarm” for the organism capable of indicating when something is wrong; however, an exaggerated response can result in chronic inflammation. Excessive NET release in diseases, such as diabetes,¹⁵⁰ augments chronic cardiovascular problems in the patients and interferes with wound healing. NETs and the many neutrophil proteins released during NET formation, as shown in Figure 3, also activate the immune system interacting with both innate and adaptive immune cells, allowing propagation of the immune response. NETs stimulate macrophages to produce important pro-inflammatory cytokines.¹¹ In addition, in vitro studies show that the released proteases on NETs modulate cytokine levels either by destroying them, thus possibly resolving inflammation,^{215, 216} or activating them, resulting in the opposite effect.²¹⁷ Moreover, recently it has been shown that NETs have a positive influence during immunization.²¹⁸ Injection of adjuvants, given in most vaccines, recruits neutrophils, which produce NETs, that in turn support the adjuvant activity.²¹⁹ In this way, NETs help to promote antibody formation and could possibly prevent future infections. However, the capacity of NETs to sustain antibody production could also have a negative impact in autoimmune disease such as lupus and RA.

Figure 3. Emerging targets for NETs inhibition and direct consequences of NETs removal.

(A) Overview of potential druggable targets that interfere with NET formation. (B) Depiction of molecules associated with NETs injurious to tissues and summary of diseases discussed in this review with experimental and clinical evidence for direct involvement of NETs in their progression. (C) Cell types whose behavior NETs are known to influence. (D) Summary of beneficial and deleterious effects of NETs degradation by DNases.

NETs not only signal trouble but can also cause it. As discussed in the review, excessive NET formation leads to many prominent pathological conditions (Figure 3B) and escalates inflammatory responses that can promote autoimmunity.⁴⁶ Binding of NETs to platelets and RBCs activates the coagulation system leading to highly efficient fibrin deposition^{74, 87} and the formation of stable thrombi.⁷³ The formation of large platelet plugs however, could be beneficial in some conditions such as during large wounds, where a large and stable plug could stop re-bleeding. Furthermore, both fibrin and NETs together would prevent entry of infectious agents into the wound. Although intuitively logical, this positive role of NETs has not been addressed experimentally.

The significant pro-inflammatory and disease promoting activities of NETs justify the development of anti-NET therapies. Currently, several inhibitors targeting proteins that impact the process of NET formation are being developed (Figure 3A). In particular, PAD4 inhibition has been shown to be beneficial in many disease models.⁶ Also Jak1/2 inhibitor, which reduces NET formation in both mice and humans, has been shown, experimentally and epidemiologically, to reduce the incidence of thrombosis.²⁴ However, these targets are not unique to neutrophils and have additional functions other than participating in the process of NET formation. Therefore, the choice of how to impede NET formation should be further investigated. Other then inhibition of NET formation, nature has developed a system that allows their degradation. In fact, both plasma and cellular (macrophage) DNases degrade extracellular DNA directly in plasma or after phagocytosis. Decreased clearance of NETs has been shown to occur or even cause autoimmune disease, like lupus and RA.²²⁰

In chronic inflammatory diseases or in autoimmune disorders, inhibition of NET formation (figure 3A) would likely have beneficial results because it would dampen the inflammatory response allowing tissue repair and restoring tissue functionality. However, in some chronic inflammatory conditions, such as otitis media,²²¹ and gingivitis,²²² NETs contribute to DNA-based biofilm that promotes bacterial growth and causes antibiotic resistance. Here, it seems that bacteria hijack our body’s neutrophils for their own benefit and protection. Treatment with DNase combined with the specific antibiotic might interrupt the chronic nature of the disease. In fact, DNases, as NET treatment, are currently being developed. Nevertheless, formation of bacterial biofilm in other circumstances might be beneficial for the host. Perhaps in the gut, NETs may play a protective role for the “good” symbiotic bacteria. However, this has not been examined and mice unable to make NETs do not have obvious digestive problems.

During pathological conditions in which NETs lead to occlusive thrombi formation (Figure 2) and interfere with the functionality of the organ, such as in heart or brain, timely DNase I treatment becomes fundamental in order to dislodge the NET-based clots. Similarly, in acute lung injury when NETs rapidly form in the alveoli (such as TRALI) DNase I inhalation and/or systemic infusion could be highly useful. However, complete blockage of NET formation could result in the inability to mount defenses against infection. Resolution of NET structures by DNase has been shown to be mainly beneficial, yet the formation of “NET debris” could potentially stimulate secondary immune responses or possibly increase the pro-coagulant state of the individual.⁷⁷

Inhibition of NET formation therefore represents a double-edged sword.⁴⁶ The medical choice of treatment may depend on the severity of the condition and on the beneficial or harmful consequences of the treatment. With the overwhelming supporting role of NETs in many prominent diseases, future studies on NETs and their inhibition are much needed.

Non-standard Abbreviations and Acronyms

PAD4

Peptidylarginine deiminase type IV

NETs

Neutrophil extracellular traps