Suicidal NETosis begins with the activation of neutrophils by stimuli, such as immune complexes, certain autoantibodies, calcium-salt or cholesterol crystals, or phorbol-12-myristate-13-acetate (PMA) (39). These stimuli activate the NADPH oxidase (NOX) complex and lead to subsequent formation of reactive oxygen species (ROS) through Raf/MEK/ERK signalling, along with an increase in cytosolic Ca²⁺ concentrations. Then, NOX and ROS complexes induce the translocation of NE and MPO from neutrophil granules into the nucleus together with the activation of protein-arginine deiminase 4 (PAD4), which reduces positive charges from histones. This causes chromatin decondensation in the nuclei of neutrophils (39, 40). Decondensed chromatin enters the cytoplasm, mixes with granular and cytosolic proteins and is finally expelled outside neutrophils accompanied by cellular lysis. Here, it forms NETs. Suicidal NETosis causes the death of the respective cell due to membrane disintegration and this process can take several hours to complete (41).

Vital NET formation occurs independently of cell death in the absence of membrane disruption within minutes after stimulation of neutrophils (42). The process is initiated by stimuli such as S. aureus or lipopolysaccharide (LPS) from gram negative bacteria through toll-like receptors (TLRs) and complement receptors (CR) (36). Vital NET formation does not rely on the NOX complex or ROS. The release of nuclear DNA in vital NET formation is associated with characteristic morphological changes (1): nuclear envelope growth and the release of vesicles (2); nuclear decondensation, and (3) nuclear envelope disruption (43–45). Vital NET formation is observed more often in infectious than in non-infectious diseases. This is supported by the observation that neutrophils stay alive and are still able to perform anti-microbial functions such as chemotaxis, phagocytosis, and killing of bacteria (46). In a specific type of vital NET formation, mitochondrial DNA may also be released, and this is dependent on ROS. This process results in NET formation from 80% of the neutrophils within 15 min following stimulation with C5a or LPS (47, 48).

Regardless of the type of NET formation certain factors, such as PAD4, NE and MPO are commonly involved in NET formation (49). However, not all of them are strictly required. PAD4 is a calcium-dependent enzyme dispersed in the nucleus, cytoplasm, and secretory granules of neutrophils. Nuclear PAD4 converts arginine in proteins (e.g. in histones H3, H2A, and H4) to citrulline. Every citrullination neutralizes one positive charge of histones and decreases their affinity for nucleic acids and concomitantly supports chromatin decondensation (50). In naive neutrophils, NE and MPO are stored in azurophilic granules (51, 52). In activated neutrophils, NE enters the nucleus, where it clips the tails of certain histones further supporting chromatin decondensation (20). Although MPO has only a minor effect on chromatin decondensation on its own, it binds to DNA and catalyzes oxidative reactions that promote NE relocation (40). Thus, MPO synergizes with NE in chromatin decondensation. Furthermore, both NE and MPO reportedly decorate the DNA backbone of NET fibers (20).

Histone citrullination is a characteristic feature of NET formation and the detection of citrullinated histones on extracellular chromatin is often used to identify NETs in tissues (53, 54). As an example, citrullinated NETs were significantly associated with high histopathological tumor grades and lymph node metastasis in human CRC ( Figure 3 ) (55).

There are several reports documenting the presence of NETs in tumor tissues (56–58). In agreement with the co-cultivation of cancer cells with neutrophils resulting in NET formation within 3 hours, electron microscopy showed that neutrophils were destructed and did not provide evidence for DNA-containing vesicles budding from intact neutrophils (58). Moreover, the cancer cell-induced NET formation depended on NOX activity (58). Altogether these findings indicated that cancer cells induced suicidal rather than vital NET formation (58).

The TME is rich in factors that can promote NET formation from both TAN and granulocytic myeloid-derived suppressor cells (GR-MDSCs), such as granulocyte-colony-stimulating factor (G-CSF) and interleukin-8 (IL-8) (59–62). G-CSF is a cytokine produced by leukocytes, macrophages, endothelium, fibroblasts and cancer cells. The expression of G-CSF is highly increased in both murine and human tumor cells (63–65). G-CSF overexpression predisposes neutrophil recruitment into metastatic lesions and enhances migration and invasion of tumor cells via generation of ROS, NET formation and production of other pro-tumor proteins (59, 66–68). IL8 is a chemokine of the CXC glutamic acid-leucine-arginine motif bearing (ELR+) family and was initially identified as a powerful chemotactic factor for neutrophils (69, 70). IL8 is produced in large amounts by several human tumors and as a main agonist of CXCR1 and CXCR2 was mostly implicated in the recruitment of neutrophils and MDSCs (61, 71). In many types of human cancer, such as bladder cancer, non-small cell lung cancer (NSCLC) and metastatic melanoma NETs show a positive association with IL-8 in tumor tissues and serum (72). In patients with diffuse large B-cell lymphoma (DLBCL), DLBCL-derived IL8 interacting with CXCR2 on neutrophils resulted in NET formation via Src, p38 and ERK signalling. Blocking of the IL8–CXCR2 axis inhibited the formation of NETs (73). Conditioned media (CM) harvested from different cancer cell lines also induced NET formation, both from neutrophils and granulocytic myeloid-derived suppressor cells (GR-MDSCs). Blocking of CXCR1 and CXCR2 with Reparixin or a CXCR1 blocking monoclonal antibody (mAb) inhibited NET formation induced by the respective CM (61, 74). Besides G-CSF and IL-8, many additional pro-inflammatory cytokines and damage-associated molecular pattern (DAMP) present in the TIME are well established to promote NET formation. In consequence, it is generally accepted that the TIME plays a critical role in the development of malignant tumors and that NETs exhibit a significant impact on tumorigenesis (75). However, how the initiation and progression of tumors is regulated by NETs and how these functions are affected by different TIMEs remains elusive.

T cells are among the most plastic cells in response to different TIMEs as they can be rendered dysfunctional and exhausted by chronic persistent antigen stimulation allowing immune escape and augmenting tumor development (76). CD8⁺ T cells of the adaptive immune system are the most potent effectors in the anti-cancer immune response and serve as executors of cancer immunotherapies with a large impact on the outcome of many different tumors (77). An interplay between NETs and CD8⁺ T cells in the TIME was suggested by de Andrea and colleagues who reported that NET density in human tumor tissues and NET concentrations in the serum of cancer patients negatively correlated with CD8⁺ T cell counts in tumor tissues (72). In analogy, Kaltenmeier and colleagues reported that NET-rich TIMEs, both in a murine liver ischemia/reperfusion (IR) metastasis model and in a murine subcutaneous tumor model induced exhaustion and dysfunction of CD8⁺ T cells (78). This was accompanied by an increase of the exhaustion markers PD-1, Tim3, and Lag3 together with a diminished production of the effector cytokines IL-2, IFN-γ, and TNF-α, and altered metabolic profiles including decreased mitochondrial function, glucose uptake, and upregulated fatty acid intake (78). Continuous exposure of murine T cells to NETs in vitro induced an exhaustive phenotype as well, supporting direct effects of NETs on CD8⁺ T cells (78). The assessment of mechanisms how NETs regulate T cell function in the TIME identified that PD-L1 was embedded in the NET chromatin of wild type (WT) bone marrow (BM) derived-neutrophils treated with PMA (79). Immune inhibitory receptor PD-1 and its ligand PD-L1 were previously recognized as an immune inhibitory axis on the surface of T cells promoting the depletion of functional T cells and tumor immune escape (79). Accordingly, NETs may directly provoke the exhaustion of T cells via the PD-1/PD-L1 axis. Consequently, targeting of PD-L1 in NETs restored functional T cells and ameliorated the tumor burden (78).

Apart from direct effects of NETs on T cell exhaustion, NETs are also able to impair the cytotoxic function of T cells by physical shielding. Teijeira et al. identified in a co-culture system and a mouse model that NETs alone cannot impact tumor cell spheroid survival or proliferation (61). However, coating of tumor cells with NETs impaired cytotoxicity of CD8⁺ T cells and protected the tumor cells from direct contact with CTLs (61). Several other studies confirmed that NETs protect tumor cells by forming a physical barrier at the tumor/stroma interface. This abrogates the infiltration of CD8⁺ T cells into the tumor cell areas (58, 80, 81). Apparently, this barrier function of NETs can protect tumors from cytotoxic immune attacks.

CD4⁺ T cells are essential regulators in cancer immunosurveillance. They modulate the tumor microenvironment and eradicate tumor cells. NETs can exert an immunosuppressive function by inhibition of CD4⁺ T cells (78). CD4⁺ T cells are highly heterogeneous and exert different immune responses to various pathogens (82). Three major subsets of CD4⁺ T cells have been identified so far: Th cells, Tregs and follicular helper T cells (TFH) (83). Th cells have drawn a lot of attention since their integral roles in the TIME was demonstrated by several pre-clinical and clinical models (84, 85). Th1 and Th2 are the two predominant categories of CD4⁺ Th cells. Th1 and Th2 are characterized by high levels of IFN-γ and IL-4 expression, respectively (86). For Th1 polarization, the signal transducer and activator of transcription-1 (STAT1) and STAT4 activated by IL-12 and IFN-γ induce the T-box transcription factor (T-bet) that drives Th1 differentiation and suppresses Th2 differentiation. IFN-γ in turn secreted from Th1 cells stabilizes Th1 differentiation forming a positive feedback loop (86, 87). Th2 differentiation is dependent on the expression of IL-4, which results in STAT6-mediated activation of the GATA3 transcription factor that stimulates Th2 polarization and suppresses Th1 differentiation. Similar to Th1 differentiation, an autocrine positive feedback loop of IL-4 stabilizes Th2 differentiation (87, 88). The balance between Th1 and Th2 CD4⁺ T cells is critically regulating tumorigenesis. A dominance of a Th1 and Th2 TIME is associated with improved or poor prognosis, respectively (88).

It has been shown that in some non-cancer diseases, such as type 1 diabetes (T1D), rheumatoid arthritis (RA), and chronic pneumonia NETs exacerbated Th1 responses by activation of dendritic cells (89–91). Moreover, NETs may also promote a Th1-like TIME through activation of peripheral blood mononuclear cells (PBMCs), which contributed to the recruitment of T cells and monocytes-macrophages and prevented tumor growth (92). In a murine model of bladder cancer Bacillus Calmette-Guerin (BCG) caused NET-induced apoptosis, cell-cycle arrest, and inhibited migration of tumor cells (92). In contrast, Zheng and colleagues found that in a murine breast cancer model with lung metastasis Th2 cytokines stimulated NET formation by maintaining complement 3 (C3) highly expressed in lung mesenchymal stromal cells (LMSCs) through the STAT6 signalling pathway (93). The Th2-STAT6-C3-NET cascade promoted lung metastasis and NET formation was essentially required for lung metastasis in the Th2 prominent TIME (93). Th2 cells are regarded as contributors to the immunosuppression and tumor escape, corresponding to NETs, which lead to the exhaustion and dysfunction of CLTs. The partially conflicting results indicate that different TIMEs and prognoses can be associated with NETs. However, it is clear that NETs can regulate the differentiation of TIME.

Th1 cells are the main cytotoxic subtype of CD4⁺ T cells. Tregs are critical for immune tolerance and homeostasis limiting potential collateral tissue damage after the initial CTL response. The central differentiation factor of Tregs is the transcription factor forkhead box protein P3 (FOXP3) (94). Wang and colleagues observed that in the non-alcoholic steatohepatitis-hepatocellular carcinoma model (NASH-HCC model, STAM) the overall number of CD4⁺ T cells dropped while the number of Tregs sharply increased coincident with more severe tumor burden. After depletion of Tregs by diphtheria toxin (DT) Th1 cells counts increased and tumor burden decreased (95). Meanwhile, it has been repeatedly confirmed that Tregs impair cancer immunosurveillance by suppressing immunity and hampering immune responses; that fosters tumor development and progression (96–99). Altogether, Tregs establish an immunosuppressive subset of CD4⁺ T cells and play an important role in the TIME. Interestingly, in murine models of liver cancer [non-alcoholic fatty liver disease (NAFLD), non-alcoholic steatohepatitis (NASH)] NETs co-localized with Tregs, occurred in early stages, and persisted during cancer development (95). Also in humans a positive correlation between the presence of the NET marker H3cit and FoxP3 was observed in livers of patients with NASH or NASH-HCC (95). From this it has been suggested that Tregs may induce NET formation (95). Vice versa, NETs can also induce Treg differentiation from naïve CD4⁺ T cells in the NASH liver microenvironment through the reprogramming of metabolic pathways involved in the oxidative phosphorylation (OXPHOS) of mitochondrial respiration (95). NETs activated TLR4 on the surface of naïve CD4⁺ T cells and subsequently modulated OXPHOS by enhancing oxidation of NADH to NAD+, which was required for Treg differentiation (95, 100). This was further confirmed by NET ablation, which was associated with a decrease of Treg-specific FoxP3 protein levels in the liver of STAM mice followed by alleviated HCC development and progression (95).

In conclusion, most studies summarized in the chapter above indicate that NETs exert immunosuppressive functions on CD8⁺ T cells and CD4⁺ T cells augmenting the progression of tumors, further supporting an important role of NETs in the TIME. Enigmatically, in certain non-cancer diseases NETs appear pro-inflammatory through eliciting immune functions.

The innate immunity establishes the first line of anti-microbial defence with a major impact on the TIME. Various kinds of macrophages together with dendritic cells, histiocytes, Kupffer cells, and mast cells are the gate keepers of innate immunity. Besides long-lived sessile tissue macrophages, monocytes from peripheral blood differentiate into macrophages after recruitment to tissues. Macrophages make up the majority of myeloid immune cells in the TIME (101, 102). They possess phagocytic functions and serve as antigen presenters in the TIME. Dependent on their activation, macrophages can be polarized into many distinct subsets, with M1 and M2 macrophages being the extremes of a multidimensional continuum (29).

M1 macrophages are induced in a Th1 TIME by pro-inflammatory cytokines, such as IFN-γ. They mediate the defence of the host from a variety of antigens and have roles in the anti-tumor immunity (103). M2 macrophages are induced in a Th2 TIME and exert an anti-inflammatory function and immunosuppressive activity (103).

Beyond their antigen-presenting capabilities and phagocytic functions, macrophages regulate NET formation and clearance. Haider and colleagues analysed abdominal aortic aneurysm (AAA) and found macrophage counts to be inversely correlated with NETs in both the intraluminal thrombi and the vessel wall (104). Macrophages, when seeded on blood-clots coated by NETs in vitro showed the ability to digest NETs by secreted deoxyribonucleases (DNases) followed by phagocytosis of the NET remnants (104). Specially, M1 macrophages were most active in NET clearance by macropinocytosis (104). Expanding on macrophage-dependent NET degradation, Fadeel and colleagues reported that physiological concentrations of extracellular DNase I were not sufficient to completely clear NETs and the cytosolic exonuclease TREX 1 (DNase III) in macrophages was required for the digestion of NETs after phagocytic uptake (105, 106).

In contrast, other reports indicated that macrophages can also induce the formation of NETs. It was shown that exosomal miR-146a from oxidized low-density lipoprotein (oxLDL)-stimulated macrophages promoted the excessive release of NETs, concomitant with the overexpression of ROS and high levels of pro-inflammatory cytokines such as IL-8, TNF-α, IL-6, and IL-1β (107). This indicated that according to their stimulation macrophages may decrease or increase the concentrations of NETs in tissues. In a third scenario macrophages may also mediate the activity of NETs. For example in an in vitro co-culture system of macrophages and A549 lung cancer cells, Zhang and colleagues reported that addition of purified NETs promoted migration and invasion of A549 lung cancer cells (108). They suggested that this partly depended on the cytokines IL-1β, IL-6, IL-18 and TNF-α, which were released from the macrophages (108). Interestingly, a subsequent study showed that the treatment of macrophages with NETs also increased extracellular M1 macrophage-derived DNA, likely representing METs (109). Accordingly, METs may have contributed to the activation of migration and invasion of tumor cells in the co-culture experiment.

In conclusion, it is best established that macrophages exert disintegrating functions on NETs. However, sporadic reports on different effects of macrophages on the formation and the activity of NETs indicate that the specific outcome of the interaction of NETs with macrophages may be TIME related, warranting further investigations.

Not only neutrophils but also macrophages can form net-like structures, referred to as METs. Similar to NETs, METs are composed of DNA and extracellular proteins and are produced in response to various pathogens and chemical stimuli. METs are generated in the course of a unique cell death program of macrophages (METosis) and until now have only been observed in the inflammatory M1 subset after treatment with IFN-γ and LPS (109, 110).

Only few differences have been reported between NETs and METs. King and colleagues noted at the morphological level that chromatin fragments are longer in NETs as compared to METs (111). In addition, METs are associated with the macrophage marker protein CD68 and the matrix metalloproteinase (MMP) 12 (111–113). Moreover, differences have been reported at the signalling processes responsible for extracellular trap formation. For example, Aulik and colleagues reported that after stimulation of cells MET formation occurs faster than NET formation (114). Finally, PAD4 and PAD2 are considered to be the main promoters of NET and MET formation, respectively (115, 116). Notably, elastase and MPO, which are often regarded as neutrophil-specific markers, have also been identified in METs (117, 118). Accordingly, co-staining of H3cit together with CD68 seems to be at present the most specific approach to distinguish METs and NETs (111–113). The differences between NETs and METs may cause their different roles in host defence.

NETs and METs generally protect against microorganisms by immobilizing them (119). However, at present it is not clear whether this immobilization is specifically microbicidal, because MET formation has also been found to promote pathogen growth by providing a scaffold for aggregation rather than killing pathogens (114, 119, 120).

In tumor tissues, macrophages are often present in higher numbers as compared to neutrophils (121). However, the presence of METs and NETs was not found to be significantly related with the infiltration of the respective cell types (122). Accordingly, the presence of METs and NETs may be more strongly related to the responsiveness of the respective cell types to TIME-derived stimuli and to different stability of NETs and METs. The first hypothesis was supported by the fact that higher concentrations of PMA are required to induce the formation of METs compared with NETs and that this process is accompanied by a higher ROS production in macrophages (111, 114).

To specifically address the relevance of METs in cancers, a retrospective study of 135 patients after radical resection of non-functioning pancreatic neuroendocrine tumors (pNETs) was analysed and showed that both NETs and METs were associated with shorter recurrence-free survival (RFS) (122). In univariate and multivariate Cox regression analyses NETs and METs were independent prognostic factors for shorter RFS and better indicators than WHO grade and TNM stage to predict the prognosis for patients (122). This finding was confirmed in another study with 116 CRC patients. In a training and a validation cohort (n=94), METs were identified as an independent prognostic factor to predict the 5-year overall survival rate in CRC and significantly associated with distant metastasis but not with local tumor invasion and lymph node metastasis (35).

Co-cultivation of the CRC tumor cell lines HCT-116 and SW480 with macrophages and a neutrophil-deficient mouse model were used to investigate the crosstalk between METs and CRC cell lines. These studies revealed that METs in the absence of NETs can activate the migration and invasion of CRC tumor cells and promote liver metastases (35). For example, the inhibition of PAD2 impeded MET formation from macrophages, blocked the crosstalk between CRC cells and METs and reduced liver metastasis (35). From this it can be concluded that targeting of both METs and NETs expands the palette of new therapeutic targets for cancer patients.

NK cells are innate cytotoxic lymphoid cells acting against tumor cells and pathogens. Activated NK cells exert their cytotoxic functions by releasing perforin and granzymes. They secrete cytokines, such as IFN-γ that participate in the orchestration of the adaptive immune response (123, 124). NK cells, unlike CTLs, are not restricted by the major histocompatibility complex (MHC). They can identify and destroy tumor cells without exposing tumor-specific antigens. This considerably accelerates the anti-cancer immune response (125, 126). The interaction between neutrophils and NK cells has emerged as an important mechanism for the modulation of immunological responses (127). NK cell-derived IFN-γ, GM-CSF and TNF-α not only prolong neutrophil survival and induce neutrophil activation, but are also essential for NET formation (128, 129). Of note, surgery promotes the production of fibrin and platelet clots coating tumor cell emboli. This limits NK cell-mediated tumor clearance (130, 131). In addition, surgery was found to stimulate NET formation and exacerbated distal organ injury by the activation of a systemic procoagulant state and diffuse microvascular immune thrombi (132–134). Clinical trials showed convincing therapeutic effects of NK cell infusion in various hematological malignancies (135). However, NET-coated tumor spheroids protected tumor cells from cytotoxicity of NK cells (61). In vitro, NETs inhibited migration and motility of NK cells indicating a direct effect of NETs (61). In a NET-rich TIME the therapeutic efficiency of NK cells was clearly impaired (135). Moreover, the inhibition of NETs in a murine model of HCC enhanced the anti-tumor immunity mediated by NK cells (135). Besides physical shielding tumor cells from attacks of NK cells, further mechanisms may exist that impair the anti-tumor function of NK cells. For example, NETs may activate platelets to impair the NK cell mediated tumorilytic function by increasing the secretion of transforming growth factor (TGF)-β, which can inhibit the mobilization of NK cells (136, 137). MMP-9 in NETs may also be implicated in NK cell dysfunction resulting in tumor immune evasion (138). All of these results indicate a mutual interaction of NETs and NK cells, which in an inflammatory TIME fosters the tumorigenic functions of NETs by counteracting NK cell activity.

Although DCs constitute a rare immune cell population in tumors and lymphoid organs, these cells are crucial for specialized antigen-presentation, the regulation of innate immunity, and the initiation of the adaptive immune response (129, 139). DCs as antigen-presenting cells are required for T cell responses and are widely used in vaccination. DCs include several subsets, such as plasmacytoid DCs (pDCs), conventional DCs (cDCs) [also referred to as myeloid DCs (mDCs)], and monocyte-derived DCs (moDCs). cDCs are the major DC population and they play a critical role in anti-tumor responses. They endocytose apoptotic and necrotic tumor cells and present tumor-related antigens to T cells (140). This may happen either in an activating or a tolerogenic manner. While pDCs mainly produce large amounts of IFN-α and IFN-β and can also be stimulated to directly activate T cells (141, 142). In patients with psoriasis and systemic lupus erythematosus NETs modulate the crosstalk between innate and adaptive immune responses and activate pDCs via TLR9 (143, 144). Sangaletti and colleagues described co-cultures of mDCs with inflammatory polymorphonuclear leukocytes (PMNs) were prone for NET formation, which induced a stable interaction of NETs and mDCs, subsequently resulting in mDCs loaded with NET components, including extracellular proteins and DNA (145). Thus, mDCs were capable to take up antigens of NETs for potential antigen processing and presentation (145). In vivo, immunization with mDCs co-cultured with NETs or apoptotic/necrotic PMNs induced anti-neutrophil cytoplasmic antibodies (ANCAs) and anti-dsDNA autoantibodies, while mDCs or NETs alone had no effect (145). Grippingly, ANCA-related autoimmune vasculitis was exclusively found in the renal and pulmonary parenchyma of mice inoculated by NET-loaded mDCs, although measurable amounts of ANCA were also detected in mice inoculated by mDCs co-cultured with apoptotic/secondarily necrotic PMNs (145). This indicated that the structural integrity of NETs was required for the transfer of cytoplasmic neutrophil antigens to mDCs (145). The autoimmune features observed in immunized mice were shared by human autoimmune systemic vasculitis (145). The interaction between NETs and mDCs might lead to autoimmunity and underlies the dynamics of ANCA induction in humans.

In cancer DCs acquire, process and present tumor-associated antigens (TAAs) on MHC molecules and shape adaptive immune responses (141). However, the investigation of the interaction of NETs and DCs in tumors is still at the beginning. Tripodo and colleagues described that NETs can boost DC vaccination against acute myeloid leukemia (AML) (146). They established a h-MRP8-NPM1+ (NPMc+) transgenic mouse model, which developed myeloproliferation without inducing overt AML (146). DCs co-cultured with NETs from NPMc+ mice were used to treat NPMc+ mice, thereby reducing NPMc+ myeloproliferation with a deceleration of myeloid expansion and a reduction of myeloid blasts by triggering immune activation (146). To further evaluate the impact of DCs uploaded with NPMc+ NETs on AML, NPMc+ mice were implanted subcutaneously with the leukemia cell line C1498 with mutant NPM1 (C1498-NPMc+) followed by a treatment with NPMc+ NETs/DC vaccine. A reduced tumor growth and stronger CTL cytotoxicity against NPMc was observed (146). NPMc+ NETs/DC vaccination enhanced anti-tumor immunity and prevented growth of leukemia transplants (146). These data indicated that NETs improved anti-tumor vaccination and tumor antigens trapped by NETs could be used to boost immune responses to cancer vaccines.