NETworking with cancer: The bidirectional interplay between cancer and neutrophil extracellular traps.

Given the many roles for NETs in cancer progression, including secondary cardiovascular events, pharmacologically targeting of NETs would have great therapeutic benefit. Unfortunately, drugs targeting NETs remain scarce. There are, nonetheless, several enzymes in the NET formation cascade that have clinically available inhibitors, e.g., NE inhibitors 34 . One NE inhibitor is sivelestat, which was approved in Japan to treat acute respiratory distress syndrome, but unfortunately failed to improve patient outcomes 287 . One possible explanation is that NE-independent mechanisms of NET formation have been reported 288 , 289 , thus decreasing enthusiasm for NE inhibitors to block NET formation. Another important molecule in NET formation is PAD4, which regulates histone citrullination. Inhibiting PAD4 is highly successful in the experimental setting, where NET formation can be blocked using small molecule inhibitors like Cl-amidine or GSK484 62 , 72 , 290 . Unfortunately, no drug that targets PAD4 is currently approved for human use. Another possible target is gasdermin D, which is a pore-forming molecule involved in pyroptosis 291 , also required for NET release 38 – 40 . Although some reports indicate that gasdermin D is not required for NET formation 292 , the FDA-approved drug disulfiram efficiently blocks gasdermin D polymerization 293 and prevents NET formation 40 , 66 , 294 . This makes disulfiram the first FDA-approved compound to block NET formation, although this activity can be considered off-target as disulfiram is approved for alcohol abuse disorder due to its activity on aldehyde dehydrogenase. An alternative approach is to target NETs once they have been released. Recombinant DNases can digest the DNA backbone of NETs and are effective in preclinical models [e.g., 72 ]. An inhaled DNase I formulation (Pulmozyme ® ) is FDA-approved for human use in cystic fibrosis where it improves symptoms of the disease 295 . However, inhaled DNase I is unlikely to reach the circulation 66 , leading to poor efficacy in contexts where NETs are found in the vasculature or in organs other than the lungs. Furthermore, although DNases target the DNA backbone, they leave several NET components behind 296 , which could be detrimental in certain contexts 297 . Nonetheless, digesting the NET-DNA scaffold may enhance the proteolytic activity of NET-bound proteases against the NET-released histones and thereby reduce the histones’ pro-inflammatory activity 192 , 193 , 295 . Besides targeting NETs directly, an alternative therapeutic approach is to target upstream mediators of NET deployment in cancer. Antibodies against IL-8 and CXCR2 inhibitors in combination with immune checkpoint blockade have shown promising results in preclinical cancer models 298 – 301 and are actively being explored in clinical trials for cancer. Similarly, there are approved agents that could be repurposed (for example, antibodies against IL-17, used to treat psoriasis), that can regulate both neutrophil recruitment and NET formation 240 , 242 . In addition, C5a can prime neutrophils to form NETs acting on C5aR, and C5aR1 inhibitors are approved for use in antineutrophil cytoplasmic antibody-associated small-vessel vasculitis 31 , 302 . Finally, inhibiting pattern recognition receptors, e.g., TLRs or cGAS, may hold therapeutic value 303 , 304 . NETs can be triggered through TLR8, which detects nucleic acid-containing immune complexes 305 ; in turn, NET-DNA can act as a TLR9 ligand for other innate immune cells, like dendritic cells 306 , 307 . How these pathways translate to cancer remain largely unknown; however, such therapies could elicit beneficial effects by targeting NETs against NETs, and also, by dampening inflammation more broadly. The development of new NET-targeting drugs will be possible as our understanding of the molecular mechanisms leading to NET formation evolves. The COVID-19 pandemic revealed the need for NET inhibitors: although it was clear early during the pandemic that NETs played a role in severe disease, there were limited approaches available to target NETs in COVID-19 patients. So far, there are just two FDA-approved compounds that can target NETs, DNase I (in inhaled form) and disulfiram. The latter can efficiently block NET formation, but it is not specific for NETs and interferes with alcohol metabolism. Further efforts should be devoted to finding drugs able to specifically block NET formation, as they would be potentially useful in many disease contexts, including cancer.

NETs contain diverse antimicrobial components, many of which have been directly implicated in modifying cancer biology. The most well-studied are proteases, which become bound to NET-DNA as neutrophils undergo NET formation. These enzymes elicit broad cytotoxic effects against target cells, as well as proteolytic effects on the vasculature and extracellular matrix (ECM). In the following section, we present key NET components and their effects on cancer progression ( Figure 3 ). NE is a serine protease stored in neutrophil azurophilic granules and released into the extracellular space following degranulation. It regulates NET formation by translocating to the nucleus and cleaving histones, thus promoting chromatin decondensation, and by degrading the actin cytoskeleton 34 , 36 . In mouse models of lung cancer, host deletion of Elane (encodes NE) results in a profound survival advantage compared to wild-type mice 13 , 167 . NE’s proteolytic activities affect cancer biology by modifying the vasculature or surrounding matrix. For instance, NE proteolysis releases growth factors and pro-angiogenic mediators, e.g., vascular endothelial growth factor (VEGF) or platelet-derived growth factors (PDGFs), from the ECM 168 . NE can also induce the expression of P-selectin, a vascular luminal adhesion protein, on human umbilical vein endothelial cells (HUVECs) 169 , and can cause dilation of the tumor vasculature to enable cancer cell transmigration during metastasis 170 . NET-associated NE can also modify the ECM by cleaving laminin and thrombospondin-1 (Tsp-1) 72 . NE can be taken up by cancer cells to induce aggressive phenotypes by activating the phosphatidylinositol 3-kinase (PI3K) pathway 13 , 171 , and it can proteolytically cleave the adhesion molecule E-cadherin to promote invasion 172 . Paradoxically, anti-tumor effects specific to human NE have also been reported; after uptake by cancer cells, its proteolytic activity liberates the death domain of CD95, thereby leading to cancer cell death 17 . These results highlight potential species-specific effects of neutrophil proteolytic function 173 . MMP9 is another key neutrophil protease released into the extracellular space via both degranulation and its association with NETs. MMP9 is part of the zinc metalloproteinase family of proteins that proteolyze the ECM and many other substrates 174 . MMP9 has been extensively studied, as its activity supports multiple aspects of tumor progression by regulating the TME 175 . Notably, the first report of MMP9 activity and NETs showed that NET-associated MMP9 can induce vascular dysfunction by causing endothelial cell damage and death through cleavage and activation of endothelial pro-MMP2 176 . Owing to its ability to regulate vascular phenotypes, neutrophil-produced MMP9 can also induce angiogenesis 177 , consistent with a landmark finding that MMP9 triggers an “angiogenic switch” during cancer progression 178 . Despite being abundant in neutrophils, evidence suggests that MMP9 levels may be fine-tuned depending on the state of the host; for example, MMP9 is upregulated in neutrophils in obesity where NET formation can impair the endothelial barrier 179 . In addition to effects on the vasculature, MMP9 bound to NETs can alter the ECM, specifically laminin, to induce the awakening of quiescent cancer cells 72 . Cathepsin G belongs to the chymotrypsin-related serine protease family, encoded within the chymase locus, and is highly enriched within azurophilic granules in neutrophils. Neutrophil-derived cathepsin G is important for killing pathogens 180 , either intracellularly in phagolysosomes or after release into the extracellular environment, including through NET deployment 181 . Neutrophil-derived DNA can block the protective activity of endogenous protease inhibitors within tissues such as the lung, to amplify cathepsin G proteolysis 182 . As a result, NET-bound cathepsin G can cleave and activate metalloproteases and proteolyze many ECM components, enabling cell invasion. In hepatocellular carcinoma models, NET-associated cathepsin G promotes invasive phenotypes of cancer cells 183 . Cathepsin G can also proteolytically modify chemokines and cytokines to mediate inflammation; for example, cathepsin G’s truncation of CCL15 enhances its ability to attract monocytes almost 1,000-fold 184 , which may contribute to macrophage accumulation within tumors. Similarly, cathepsin G-mediated cleavage of chemerin, a chemotactic protein, increases the chemotaxis of dendritic cells during early stages of inflammation 185 . Additionally, the activities of the IL-1 family of cytokines, including IL-18, IL-33, and IL-36 186 – 188 are also enhanced by cathepsin G cleavage. In cancer models, soluble forms of cathepsin G can stimulate angiogenesis via cleavage of pro-MMP9 and subsequent TGFβ activation, leading to VEGF induction 189 . This echoes the effects of other neutrophil serine proteases on vascular remodeling at sites of inflammation. Histones within NETs can damage endothelial cells directly, as they are inherently cytotoxic 56 , 190 . At sub-cytotoxic concentrations, histones together with DNA can induce pro-inflammatory signaling more potently than DNA alone 191 – 193 . Additionally, NET-DNA has both structural and signaling functions 194 . In mouse models of sepsis induced by cecal ligation and puncture, NET-DNA was shown to physically trap circulating cancer cells to guide them across the vascular barrier into secondary tissues, including the liver 73 , 195 . NET-DNA can also function as a scaffold to concentrate protease activity on ECM substrates 72 . Association with the DNA scaffold has also been proposed to reduce serine protease activity, giving the DNA-bound proteases a potentially different role than their soluble counterparts 34 . More recently, NET-DNA itself has also been shown to have chemotactic functions, acting as a signaling moiety via the protein CCDC25 196 . In cancer cells, NET-DNA, recognized by CCDC25, activates the ILK-β-parvin pathway, resulting in enhanced cell proliferation, adhesion, and migration 196 . NET-DNA has higher levels of 8-hydroxy-2’-deoxyguanosine (8-OHdG), a marker of oxidative damage, than normal genomic DNA 196 . Interestingly, 8-OHdG is also a recognized risk factor for cancer, atherosclerosis, and diabetes, which all are diseases linked with NETs 197 , 198 . Finally, it is likely that partially digested NET-DNA can be taken up by immune cells to trigger signaling from intracellular DNA sensors, such as cyclic GMP-AMP synthase / stimulator of interferon genes 1 (cGas/STING) 199 , absent in melanoma 2 (AIM2) 200 , 201 , or TLR9 202 . The consequences of NET-DNA’s activation of pattern recognition receptors or cGAS 203 remain an active area of research. Tandem mass spectrometry on NETs isolated from the blood of healthy volunteers has revealed more than 500 NET-affiliated proteins, including adhesion molecules belonging to the integrin (ITGAM, ITGB2, ITGAIIb, ITGAL) and carcinoembryonic Ag cell adhesion molecule (CEACAM; CEACAM1, CEACAM6, CECAM8) families 204 . Many of these proteins are likely to have both direct and indirect effects on cancer cells. For example, CEACAM1 is found on NETs and directly stimulates colorectal cancer cell adhesion and migration 204 . β1-integrin, present in NETs, can physically tether NETs to circulating cancer cells, assisting in their dissemination into secondary organs 195 . NET-bound proteins can also regulate cancer indirectly through their effects on other cells in the microenvironment. For example, PD-L1 can be detected on NETs, resulting in T cell exhaustion and an immunosuppressive TME 52 . These studies represent an evolving understanding of the roles of the many NET-associated factors in cancer.

NETs are important players in the TME, but they impede health beyond promoting cancer. In the 19 th century, Trousseau realized that malignancies and thrombotic events were correlated 253 , and unexplainable thrombotic events are now taken to indicate a possible hidden visceral cancer in clinical practice. The link between cancer and thrombosis is clinically very important, as thromboembolic cardiovascular events are the second most common cause of death in patients with cancer 254 . Cancer survivors have a higher risk of death from cardiovascular disease than the general population 255 – 257 , and up to 50% of cancer patients show postmortem histological evidence of venous thromboembolism 258 . These relationships are further exacerbated by additional host co-morbidities, e.g., obesity and metabolic syndrome, that impede cardiovascular health. In this regard, cancer can be considered a vascular disease, and multiple host factors likely cross-talk to bolster this relationship ( Figure 5 ). The tumor vasculature is pro-coagulant, and both cancer cells and the TME release pro-thrombotic compounds 259 , 260 . Moreover, the tumor vasculature itself is leaky, mainly resulting from low pericyte coverage 261 , 262 , and it is marked by intraluminally-exposed ECM components 263 . Together, these vascular abnormalities contribute to platelet and neutrophil activation and subsequent NET formation 85 , leading to severe vascular dysfunction 264 . In fact, poor vascular integrity, which triggers NETs, indirectly promotes metastasis through varied effects on cancer cells, including hypoxia induction, EMT initiation, and MET activation 265 . Intravascular NET release also leads to increased tumor angiogenesis via, e.g., NE (discussed above) and MMP9 and latency-associated peptide of TGFβ1, which induce endothelial expression of pro-angiogenic compounds 234 . Within the vasculature, NETs play prominent roles in thrombosis: they damage endothelial cells 266 and induce aggregation by providing a scaffold for platelet deposition 267 , exposing pro-coagulant nucleic acids and polyphosphates 268 , as well as several platelet-activating ligands 269 . They also expose tissue factor (that activates the coagulation cascade) together with NE (which inhibits the tissue factor inhibition pathway), further inducing thrombosis 270 , 271 . Of note, NETs deployed in the brain vasculature can also disrupt the blood-brain barrier 272 , potentially assisting metastatic spread to the brain. Histones themselves also interact with platelets and activate the intrinsic coagulation cascade 273 , 274 . Although platelets are key players in thrombosis, NETs can also induce platelet-independent clots 63 . Together with other cytotoxic components of neutrophil granules, histones themselves damage endothelial cells 56 , causing neutrophil recruitment and establishing a feed-forward loop between NETs and platelets that ultimately drives more neutrophil recruitment, more platelet activation, and more NET formation. This feedforward loop leads to thrombi formation, and thrombi themselves can further drive NET formation and neutrophil recruitment 86 , 275 , especially in the lower-shear region downstream of the thrombi 276 . Finally, NETs can induce emboli formation, as exemplified in their role in atheroma plaque destabilization 277 . In summary, excessive NET formation can induce thrombosis and thromboembolism 64 , leading to ischemic events and potentially organ failure, and there is compelling evidence that NETs contribute to cancer-associated thrombosis [see for example 77 , 142 ]. Host metabolism is closely tied to cardiovascular health, and many aspects of metabolic syndrome—including hyperglycemia, hypertension, and hypercholesterolemia—influence neutrophil-specific biology 251 , 278 . Obesity is strongly associated with the development of metabolic syndrome and is estimated to underlie 14–20% of all cancer-related mortalities in adults 279 . Preclinical models of breast cancer have shown that obesity increases NET formation, which impairs vascular integrity in the lungs to enhance permeability and metastasis 179 , 251 . Inhibiting NET formation with PAD4 inhibitors or digesting NETs with DNase I are sufficient to restore vascular barrier integrity and reduce obesity-associated lung metastasis 179 . Clinical studies have also shown that individuals with type 2 diabetes (T2D) display increased circulating levels of NETs due to hyperglycemia-induced NET formation 70 . Similarly, homocysteine (also a risk factor for T2D when elevated systemically) can stimulate NET formation and platelet aggregation to influence vascular pathologies 280 , reinforcing the role host factors play in influencing immune and vascular states. A similar finding was reported in patients with severe coronary atherosclerosis, whereby circulating levels of NET markers were elevated in patients compared to healthy subjects 281 . Finally, endocrinological factors affected by psychological stress can also impact vascular biology and cardiovascular health, although the role of stress in cancer immunology and outcomes is still an emerging area of research 282 , 283 . In preclinical models of sickle cell disease, it was found that exposure to chronic stress negatively impacted vascular disease via an increase in aged neutrophils 284 . Although a direct connection with NETs was not explored, these findings reveal the intricate relationship between host co-morbidities and neutrophil states. Taken together, it is important to recognize that co-occurrence of certain disease states may confound our current understanding of the mechanisms contributing to cancer disease processes. It is plausible that underlying host conditions perturb neutrophil and NET biology, with cancer exacerbating these effects, but further research is needed. Nevertheless, data strongly support the idea that NETs can drive cancer-associated thrombosis in humans. First, elevated blood levels of citrullinated histones are associated with increased risk of venous thrombotic events in patients with cancer 285 . Second, cancer-associated arterial microthrombi in autopsy samples contain citrullinated histones 286 . The fact that neutrophils release NETs in the context of cancer likely contributes to the abnormally high incidence of cardiovascular events, including thromboembolic events, in patients with cancer.

The neutrophil lifecycle and neutrophil maturation are tied to the acquisition of effector functions, including the ability to form NETs and migrate. Therefore, disentangling how cancer alters neutrophil maturation is key to understanding the impact of NETs on cancer. Neutrophils are produced in the bone marrow within hematopoietic cords surrounded by venous sinuses. One fascinating feature of neutrophils is their short lifespan 95 , which under normal conditions, is less than one day following release into the bloodstream 2 . The reason for their short lifespan is unclear, but may be a guard against the damaging potential of their vast cytotoxic arsenal 58 . Many factors that regulate the neutrophil lifecycle under normal conditions are highly expressed by tumors, leading to dysregulated neutrophil maturation, lifespan, and effector functions in cancer ( Figure 2 ). For instance, tumors can induce the early release of immature, immune-suppressive neutrophils by producing G-CSF 96 . In the following section, we summarize the main steps of the neutrophil lifecycle, the influence of lifecycle on NET formation, and how cancer changes the lifecycle. Neutrophils are produced from hematopoietic stem cells that give rise to proliferative multipotent, common myeloid, and granulocyte-monocyte progenitors (GMPs). The latter, via committed proliferative neutrophil precursors 97 , 98 , ultimately differentiate into mature neutrophils. Neutrophil maturation is governed by the complex interplay of several transcription factors: PU.1, CCAAT/enhancer-binding protein alpha (C/EBPα), CCAAT/enhancer-binding protein epsilon (C/EBPε), Growth Factor Independent 1 (Gfi-1), and GATA-binding factor 1 (GATA-1) 99 . Immature neutrophils are less effective at forming NETs than their mature counterparts 100 – 102 . Recently generated neutrophils spend 4–6 days in the bone marrow before their release into the circulation 3 . This release is controlled by C-X-C motif chemokine receptor 2 (CXCR2, mobilization-promoting) and CXCR4 (retention-promoting) signals 103 , 104 and follows a circadian rhythm: in mouse, neutrophils are released from the marrow into circulation during the night 2 , 105 . In homeostasis, sympathetic innervation controls the rhythmic, circadian expression of CXCL12, the ligand for CXCR4, in bone marrow stromal cells 106 , 107 . G-CSF promotes mobilization by decreasing CXCL12 expression in the bone marrow 108 while simultaneously increasing the expression of CXCR2 ligands (CXCL1 and CXCL2) in bone marrow endothelial cells 103 . “Fresh” or “young” neutrophils, newly released from the bone marrow, stay in circulation for less than one day before being cleared into tissues. Within the blood, fresh neutrophils undergo neutrophil “aging”, a process of phenotypic changes, before being cleared out of circulation 2 . This process is also circadian: it maximizes immune readiness during the active phase of the organism, when pathogen encounter is more likely, while minimizing potential collateral damage to healthy tissues from overactive neutrophils during the resting period. Control of neutrophil aging is cell-intrinsic and -extrinsic. Cell-intrinsic mechanisms are driven by the core molecular clock machinery to induce circadian CXCL2 expression, which promotes neutrophil aging through autocrine CXCR2 signaling 58 , 109 . External cues also guide the process: engaging the CXCL12-CXCR4 axis dominantly blocks CXCR2 signaling 103 in an AKT- and mTOR-dependent manner 110 , effectively inhibiting aging. As CXCL12 is highly expressed in the bone marrow 107 , the aging process does not start until neutrophils are released into circulation 105 . CXCL12 levels in plasma show a circadian pattern that peaks in antiphase with the number of aged neutrophils, serving as a second circadian control checkpoint 109 . As neutrophils age, their transcriptomic and proteomic profiles change, as do their effector functions. Fresh neutrophils have high surface levels of L-selectin (CD62L) and CXCR2 and low levels of CXCR4. As they age, CD62L and CXCR2 expression decreases, while CXCR4 expression increases 111 . They also slowly degranulate, reducing their cytotoxic potential before being cleared to the tissues 58 . Fresh neutrophils have higher migration capacity to inflamed tissues, while aged neutrophils are preferentially recruited to tissues in homeostasis 109 , where they can be reprogrammed to perform tissue-specific functions 112 . Aged neutrophils can cause more intravascular damage 109 , 113 and have increased phagocytic capacity 114 than fresh neutrophils. In terms of NET formation, there are conflicting reports: fresh neutrophils release NETs more efficiently in response to inflammatory stimuli than their aged counterparts 58 , 110 , but increased NET formation by aged neutrophils has also been reported 113 . This discrepancy may be due to differences in experimental models, such as whether aged neutrophils were isolated under physiological conditions 58 or from animals where aged neutrophils in blood were enriched by blocking extravasation into tissues 113 . Once neutrophils have fully matured and been cleared out of circulation, they enter their final lifecycle stage. It was previously thought that neutrophils were exclusively cleared from blood to the bone marrow, liver, and spleen 115 , 116 . However, recent data show that during homeostasis, neutrophils are cleared into virtually all tissues 117 , and are removed at the end of their life by tissue-resident macrophages 118 – 120 . Before elimination, neutrophils in tissues provide immune surveillance (mainly during the night, the active phase in mice) 109 . In the lung, liver, and intestine, neutrophils gain pro-angiogenic functions; in the spleen, they support B cells; and in the skin, they exert epithelial and connective tissue-supporting functions 112 , 121 . Neutrophil lifespan varies dramatically between organs, likely reflecting differences in tissue-specific functions 112 . Finally, despite their short lifespan, neutrophils are continuously being produced, rhythmically released, and cleared in great numbers, allowing them to function as a phenomenal collective workhorse. Cancer can affect the neutrophil lifecycle, as cancer cells often express G-CSF and/or CXCL1 122 , 123 , which can promote granulopoiesis and drive neutrophil mobilization into the bloodstream. Other cytokines expressed by tumors—e.g., IL-6, IL-1β, interferon gamma (IFN-γ), and tumor necrosis factor alpha (TNF-α)—also enhance the production and mobilization of neutrophils from the bone marrow. This process, known as emergency granulopoiesis, can lead to an expansion of mature, fresh neutrophils in circulation—neutrophils that have increased NET formation ability 58 , 124 . However, emergency granulopoiesis can also lead to the release of immature neutrophil populations with reduced NET formation capacity 101 . It is unclear exactly what determines whether mature versus immature neutrophils leave the bone marrow, but the process likely depends on a balance between the signals that promote granulopoiesis versus release. Emergency granulopoiesis, in both infection and cancer, also induces the release of hematopoietic stem and progenitor cells (HSPCs) that can circulate and seed in other tissues to induce extramedullary hematopoiesis, which further increases hematopoietic output 125 – 128 . Skewing of immune cells toward myeloid lineages is correlated with worse prognosis in breast, cervical, esophageal, gastrointestinal, and lung cancers 129 , likely caused by cytokines, including G-CSF, secreted by cancer cells 122 . High numbers of neutrophils or an elevated neutrophil-to-lymphocyte ratio (NLR) in blood are associated with poor prognosis in many cancers 130 – 133 . Consistently, during treatments, a low NLR in blood may indicate improved responses e.g., following immune checkpoint blockade in non-small cell lung cancer 134 . As humans age, HSPCs become skewed toward myelopoiesis 135 , resulting in a bias toward myeloid cells over lymphoid cells. Age also contributes to clonal hematopoiesis of indeterminate potential (CHIP) 136 , a condition in which genetic mutations in hematopoietic stem cells grant a selective proliferative advantage without meeting the definition of cancer 137 , although it can sometimes progress to leukemia 138 . Cancer patients with CHIP are at higher risk of progression and recurrence after treatment 139 , 140 . Interestingly, one common mutation driving CHIP is an activating mutation of JAK2 that enhances proliferation and is associated with a 12-fold higher risk of coronary heart disease and thrombosis 141 . Moreover, clonal hematopoiesis with JAK2 mutation enhances neutrophils’ ability to form NETs 142 , which could explain the increased risk of atherosclerosis, given the role of NETs in atherogenesis 143 . Other alterations at the level of the bone marrow can also affect neutrophil behavior. “Trained immunity” is a long-term memory mechanism that causes a shift in the function of innate immune cells against subsequent (and often unrelated) immune challenges 144 . It is mediated by long-term epigenetic and metabolic rewiring of progenitor cells in the bone marrow 145 , 146 . Unlike adaptive immune system memory mechanisms, trained immunity is not antigen-specific 147 . Trained immunity can be elicited by exogenous pathogen-associated molecular patterns, endogenous danger-associated molecular patterns 148 , and even by dietary components 149 . Studies have shown that canonical trained immunity (e.g., caused by β-glucan or Bacillus Calmette-Guérin (BCG) vaccination) induces a myeloid bias in hematopoiesis 144 , 146 , leads to long term effects on neutrophil effector functions 150 , and enhances anti-tumor responses 151 . In cancer, childhood BCG vaccination significantly decreases lung cancer incidence 152 , while in bladder cancer, it is used as an adjuvant immunotherapy 153 where it can trigger NET formation to mediate tumor-control 154 . Thus, in this context, trained immunity and NETs are anti-tumor. On the other hand, trained immunity can also be pro-tumor, as innate immune training driven by myocardial infarction drives immunosuppressive phenotypes in myeloid cells that can promote breast cancer 155 . Thus, trained immunity can positively and negatively affect cancer progression, and further research will be needed to understand these opposing effects. In addition to the interactions between tumors and the bone marrow, several studies implicate potential interactions with the process of aging and clearance of neutrophils in cancer. As neutrophils age within the circulation, an intriguing possibility is that they may facilitate metastasis by allowing circulating cancer cells to ‘hitchhike’ and follow the neutrophils into tissues. Neutrophil-cancer cell interactions can be governed by CD11b 156 , which is elevated on the cell surface of aged neutrophils 111 . Additionally, disseminating cancer cells in blood can be physically guided by NETs 73 , 157 . Many cancers express high levels of CXCL12 158 , 159 and may, therefore, attract aged neutrophils, which express higher levels of the CXCL12 receptor CXCR4 104 , 111 . Neutrophils infiltrating tumors may be reprogrammed by tumor-derived factors similarly to their reprogramming in normal tissues 112 . One example is the pro-tumor phenotype of neutrophils that have been exposed to transforming growth factor (TGFβ) in solid tumors 12 . Aged neutrophils are less efficient at forming NETs during homeostasis, but it is conceivable that the pro-inflammatory milieu of many tumors, which can extend neutrophil lifespan 160 – 163 , also increases NET formation. In the context of cystic fibrosis, delayed neutrophil apoptosis leads to an increased likelihood of NET formation 164 . The normal circadian changes in neutrophil migration patterns mirror recent findings that breast cancer cells have greater metastatic potential during the “resting” period of both humans and mice, driven by an increase in mitotic gene expression 165 . This phenomenon raises the possibility that neutrophils and NETs play a role in promoting rest phase-driven metastasis, given that neutrophil numbers in circulation are greatest during this period. Additional studies suggest a role for aged neutrophil clearance in metastasis: more cancer cells home to the lungs during the time of day when neutrophils are actively cleared into tissues, and this diurnal difference in cancer cell homing is lost upon neutrophil depletion 117 , 166 . The extent to which these diurnal effects act within the circulation or the target tissues remains unclear 72 In summary, cancer can co-opt neutrophils at different stages of their lifecycle, including changing neutrophil aging to alter the neutrophil’s behavior and ability to form NETs. Further research is needed to fully understand the role of bona fide neutrophil aging and aging-related changes in NET formation in cancer.

In the past decade, neutrophils and NETs have emerged as central regulators of tumor progression, with diverse effects on cancer cells and the microenvironment. Yet many knowledge gaps remain regarding how NETs influence cancer biology. First, although most studies identified tumor-promoting effects of neutrophils and NETs, other studies have shown tumor-inhibiting effects, most often in early-stages of cancer or metastasis 244 , 308 . Factors that dictate whether NETs are pro- or anti-tumor are unknown. There are likely to be tissue-specific determinants, given the profound variation in tissue composition, and thus protease substrates or cells, that the NETs can act on. Of note, NETs are present in both the tumor tissue and plasma of cancer patients 90 , 309 , and it is likely that NETs released within blood vessels versus tissues have different consequences. The context-dependent effects of NET formation raise another fundamental question: are all NETs are equal? We know very little about whether NET components (e.g., their proteolytic or immunoregulatory components) differ in different contexts, although the diurnal changes of neutrophil granule contents 58 suggest that NET-associated proteins may also vary. Finally, how NETs are cleared from tissues and blood is not well understood. Plasma DNases can degrade intravascular NETs 63 , 310 , and internalization by macrophages and subsequent intracellular degradation has been proposed as well 311 , but how NET degradation or its potential dysregulation influences cancer remains largely unexplored. It is conceivable that NET degradation could have a major impact by regulating the amount of time that NETs are present and by generating NET degradation products: proteases and histones released from the NET-DNA scaffold, as well as nucleic acid fragments, may have different functions than intact NETs. Perhaps the biggest remaining goal in the field is to be able to move neutrophil- and NET-targeted therapies into clinical practice. So far, advancements have been limited, yet we are hopeful that leveraging efforts to target NETs in other conditions, such as COVID-19, or repurposing drugs, such as disulfiram, may accelerate progress toward this goal. Nevertheless, our insights into the functions of neutrophils and NETs in cancer are growing at unprecedented speed, and there is a growing realization that they are attractive targets due to their multifaceted effects on the microenvironment and on the body as a whole.

As prime effector cells of the immune system, neutrophils possess a wide array of functions to fight invading pathogens: phagocytosis 21 , ROS generation 22 , degranulation and the consequent release of proteases and other cytotoxic granule components 23 , the recruitment of other immune cells 24 and the formation of NETs 25 . NETs are web-like filamentous extracellular structures released by neutrophils in response to supernumerary 25 or oversized 26 pathogens. NETs entrap pathogens in a network of DNA, histones, proteases, and other cytotoxic and highly inflammatory compounds, including myeloperoxidase (MPO), lactotransferrin, LL-37, calprotectin, bactericidal/permeability increasing protein, and pentraxin 3 25 , 27 – 32 . To form NETs ( Figure 1 ), neutrophils decondense their chromatin in a process that requires ROS, neutrophil elastase (NE), myeloperoxidase activity, and histone citrullination 33 – 35 . Within NET-forming neutrophils, NE degrades the actin cytoskeleton, thereby impeding their ability to move and phagocytose 36 . The nuclear membrane is also dismantled, causing the nuclear contents to blend with the cytotoxic cargo released from the granules 37 . Lastly, NETs are released into the extracellular space following the loss of plasma membrane integrity, a process involving gasdermin-D polymerization 38 – 40 . NET formation usually leads to the lytic death of the neutrophil, termed “NETosis” 37 ; however, non-lytic forms of NET formation have also been reported 41 , 42 . Given that NETs form through a cellular “suicidal” process and indiscriminately affect both pathogens and host tissues upon their release, the formation of NETs can be thought of as a last-resort pathogen defense mechanism. However, NETs are produced in response to many different pathogens 25 , 26 , 43 – 46 , and using extracellular DNA traps to defend against pathogens is an evolutionarily conserved innate immune mechanism found in social amoeba 47 and plants 48 . Besides the pathogen response, NETs are also released in response to sterile injuries 49 , although their purpose in this setting is less clear. Many inflammatory mediators that induce NETs (e.g., interleukin 8 [IL-8, or C-X-C Motif Chemokine Ligand 8, CXCL8], CXCL1, mitochondrial DNA, or nitric oxide) are shared between sterile and non-sterile injuries, suggesting that NET deployment under sterile conditions may be coincidental 50 . However, NETs also have anti-inflammatory functions. For example, NETs with associated proteases can act as cytokine- or chemokine-degrading scaffolds to dampen further inflammation 51 . Additionally, they can promote T cell exhaustion through immunosuppressive ligands embedded within their chromatin, such as programmed death-ligand 1 (PD-L1) 52 . Nevertheless, due to their high cytotoxicity, NETs are most commonly reported to be detrimental in sterile injuries 53 . NETs are coated with granule-derived, lytic, cationic antimicrobial peptides 54 ; proteases; and histones, which are highly cytotoxic to pathogens and bystander host cells alike 55 , 56 . Dysregulated NET formation can inflict collateral damage to healthy tissues; highly vascularized organs, e.g., lung, kidney, and liver, are particularly vulnerable 31 , 32 , 57 , 58 . As a consequence, NET-induced damage has been implicated in several conditions: lupus 59 , periodontitis 60 , atherosclerosis 56 , 61 , 62 , thrombosis 63 , 64 , acute lung injuries 26 , 26 , 58 , 65 – 67 , sepsis 68 , psoriasis 69 , diabetes 70 , 71 , and cancer 72 – 76 . NETs are induced in cancer through various mediators. Ex vivo , inflammatory molecules released from cancer cells [e.g., IL-8/CXCL8, granulocyte colony-stimulating factor (G-CSF), CXCL1, CXCL2, Cathepsin C, and Toll-like receptor (TLR) ligands 76 – 81 ] can induce NETs, although data confirming that these molecules induce NETs in vivo are largely lacking. Tumor-derived exosomes can also induce NETs, coinciding with enhanced cancer-associated thrombosis 82 . Additionally, cells within the TME, such as cancer-associated fibroblasts (CAFs) and platelets, can induce NETs. In mouse models of melanoma, and lung and pancreatic cancer, CAFs induce NETs by secreting amyloid β 83 . Tumors also activate platelets 84 , which, in turn, promote intravascular NET formation 85 . Activated platelets induce NETs, e.g., by releasing high mobility group box 1 protein (HMGB1) 86 or via P-selectin 87 , which may act on neutrophils in both soluble or cell-bound forms. Besides specific cell populations, the state of the microenvironment is also relevant; for example, cancer cells experiencing hypoxia—as frequently observed in solid tumors—may have an enhanced ability to induce NETs 88 . The presence of NETs in aggressive tumors is therefore common, and hallmark features of NETs, including a citrullinated form of histone 3, are prognostic in the clinical setting 89 , 90 . Intriguingly, several NET-inducing molecules supplied by cancer cells serve multiple functions in neutrophil biology. For example, CXCL1, CXCL2, and IL-8 are major neutrophil recruitment factors 91 , while G-CSF and CXCL1 regulate granulopoiesis and neutrophil efflux from bone marrow 92 – 94 . Thus, tumors can hijack many aspects of basic neutrophil biology, including their maturation, trafficking, and effector functions.

Neutrophils and NETs play roles in cancer initiation, either indirectly by exacerbating inflammation or directly by perpetuating genotoxic stress ( Figure 4 ). A direct role for neutrophils in cancer initiation was found in zebrafish models of melanoma, where wounding-induced inflammation increased cancer formation in a neutrophil-dependent manner 205 . A classical analogy in cancer biology is that tumors are like “wounds that do not heal” 206 . During wound healing, neutrophils, NETs, and their associated proteases are highly abundant 207 , 208 ; however, NETs’ role in tissue repair is poorly understood, as neutrophil depletion accelerates wound closure in animal models 209 . Indeed, NETs can induce chronic wounding and directly impede wound closure, as exemplified by their defining role in the diabetic wound, another “wound that does not heal” 71 . It is estimated that at least 20% of all cancers arise as a direct consequence of various chronic inflammatory conditions 210 . Specific examples include inflammation stemming from infections with Helicobacter pylori, hepatitis virus B and C, or Kaposi’s sarcoma-associated herpesvirus, or from schistosomiasis, endometriosis, inflammatory bowel disease, thyroiditis, prostatitis, or asbestos 211 . For each of these inflammatory conditions, neutrophils 212 , 213 and NETs 214 are part of the innate immune response to clear pathogens and/or re-establish tissue homeostasis. In a mouse model of nonalcoholic steatohepatitis (NASH)-induced tumor development, NETs promoted the onset of hepatocellular carcinoma by orchestrating monocyte-derived macrophage infiltration to the inflamed liver and inflammatory cytokine production 215 . Thus, NETs can be critical components of a microenvironment that favors cancer initiation. In addition to supporting tumor formation by exacerbating chronic inflammation, neutrophils may also have direct carcinogenic capacity. Classical literature linked neutrophil production of ROS (which precedes NET formation) with an ability to cause malignant transformation through the induction of mutations and sister chromatid exchanges 216 – 218 . More recently, neutrophils were shown to be critical for tumor formation in several chemical carcinogenesis mouse models 219 – 221 , although it was not clear whether the neutrophils acted by generating a tumor-permissive microenvironment or by directly inducing mutations. In addition to ROS, neutrophils have been proposed to promote neoplastic transformation by producing genotoxic hypochlorous acid and other reactive molecules generated by MPO during oxidative bursts 222 , 223 . Finally, micro-RNAs (miR-23a and miR-155) released by neutrophils promote double-stranded break formation, leading to genomic instability in cancer 224 , 225 . Intriguingly, miR-155 can also induce NETs 226 .

Neutrophils are the main leukocyte population in human blood and “first responders” under inflammatory conditions 1 . They exist in circulation for approximately 12h 2 , 3 and have, until recently, been considered uniform, endpoint effector cells protecting against invading pathogens. This oversimplified view of neutrophil biology has slowed progress on understanding their role in chronic diseases, including cancer 4 . However, over the last decade, the roles of neutrophils in cancer biology have been receiving more attention; they are now considered major players within the tumor microenvironment (TME) and have been functionally implicated in all stages of cancer progression. Neutrophils promote tumor initiation, mainly through the inflammatory production of reactive oxygen species (ROS) or protease release 5 – 7 . They also regulate tumor progression, with both pro- and anti-tumor functions 8 . Depending on the context, neutrophils can either inhibit metastasis directly through cytotoxic activity or support metastasis by promoting immunosuppression, angiogenesis, cancer cell motility, and epithelial to mesenchymal transition (EMT) 9 – 17 . Nonetheless, most studies on neutrophils in cancer have reported on pro-tumor role. Indeed, intratumoral neutrophils were reported to be the most adverse prognostic cell type of all tumor-infiltrating leukocyte populations, based on a pan-cancer assessment of over 3,000 solid tumors from 14 different cancer types 18 . Here, we discuss the involvement of neutrophils in cancer initiation, progression, and metastasis. We focus on the role of neutrophil extracellular traps (NETs), as the role of neutrophils in cancer more generally has been recently reviewed elsewhere 19 , 20 . We first discuss the effects of cancer on NET formation, including how tumors alter the neutrophil maturation lifecycle, then explore how different NET components affect specific steps of cancer initiation and progression. Finally, we discuss how NETs may underlie the link between cancer and co-morbidities (particularly cardiovascular events and obesity), and provide insights into developing therapeutic strategies to target NETs in the clinical setting.