Neutrophils in cancer carcinogenesis and metastasis

Inflammation plays an essential role in cancer initiation by damaging tissues, and neutrophils are a crucial component of this process. Thus, neutrophils provide a link between inflammation and cancer. Cancer that develops in various mouse models of KRAS-driven ovarian cancer exhibits upregulated levels of neutrophil-related chemokines and an expansion of neutrophils. These phenotypes may result from direct upregulation of neutrophil-related cytokines such as GM-CSF and CXCL8 [24, 25]. In a zebrafish model of HRAS^G12V-driven melanoma, wounding-induced inflammation with elevated levels of prostaglandin E2 increase the formation of cancer in a neutrophil-dependent manner [26]. Depletion of the entire neutrophil population using anti-Ly6G antibodies impairs carcinogenesis in both chemically induced and spontaneous cancer models. Neutrophils overexpressing CXCR2 are attracted to cancer-prone tissues via the cytokine IL-8 and chemokine ligands CXCL1, CXCL2 and CXCL5. The application of chemical carcinogens in CXCR2-deficient mice, which show impaired neutrophil trafficking, prevents papilloma or adenoma formation [27, 28]. CXCR2-mediated neutrophil trafficking from bone marrow into peripheral blood is antagonized by CXCR4 expression due to the retention of neutrophils by CXCL12-expressing bone marrow stromal cells mediated retention. Bone marrow macrophages subsequently eliminate the retained neutrophils in a rhythmic manner.

The evidence described above has indicated that neutrophils are crucial for carcinogenesis, but the exact mechanisms by which neutrophils foster carcinogenesis require further elucidation. Neutrophils produce and release genotoxic DNA substances that increase DNA instability. In an in vitro coculture model mimicking intestinal inflammation in ulcerative colitis, neutrophils increase errors in the replication of colon epithelial cells. In individuals with chronic colon inflammation, activated neutrophils cause an accumulation of target cells in G2/M phase, consistent with the installation of a DNA damage checkpoint [29]. Neutrophil-derived elastase, neutrophil production of ROS, reactive nitrogen species (RNS) and angiogenic factors such as MMP-9 and the immunosuppressive ability of neutrophils may be associated with this process. ROS released by neutrophils during chronic inflammation, such as hypochlorous acid (HOCl, formed by myeloperoxidase (MPO)), cause DNA damage and are mutagenic in lung cells in vitro. HOCl is a major neutrophil oxidant. MPO-catalyzed formation of HOCl during lung inflammation is an important source of neutrophil-induced genotoxicity. Neutrophils cause DNA damage by releasing ROS and inducing gene mutations in premalignant epithelial cells, thus driving oncogenic transformation in lung cancer. Additionally, at physiological concentrations, HOCl induces mutations in the hypoxanthine phosphoribosyl transferase (HPRT) gene, inducing three major types of DNA lesions [30]. Haqqani and coworkers analyzed a mouse model of subcutaneous cancer and showed that inducible nitric oxide synthase (iNOS) and nitric oxide synthase (NOS) contents and neutrophil infiltration were significantly correlated with the number of mutations in the Hprt locus [31]. However, a new mechanism that does not rely on ROS was also recently identified. In clinical samples from patients with inflammatory bowel disease and injury models, activated tissue-infiltrating neutrophils release particles carrying proinflammatory microRNAs, including miR-23a and miR-155, which increase DNA double-strand breaks and genomic instability [32]. miR-155 is also responsible for neutrophils-induced DNA damage and DNA repair landscape in acute colon injury, resulting in colorectal cancer initiation even shaping the progression [33].

Coussens et al. documented that MMP-9 supplied by bone marrow-derived neutrophils and other hematopoietic cells contributes to squamous carcinogenesis [34]. MMP-9 produced by neutrophils also contributes to the carcinogenesis of pancreatic islet carcinoma and lung cancer accelerating angiogenesis [35]. NETs promote inflammation in subjects with nonalcoholic steatohepatitis, resulting in the development of hepatocellular carcinoma, which is inhibited by deoxyribonuclease treatment or peptidyl arginine deaminase type IV knockout, decreasing NET formation [36]. Furthermore, NETs positively correlate with the increased number of regulatory T cells (Tregs) in cancer by facilitating naïve CD4⁺ T cell metabolic reprogramming. Therapies targeting the interaction between these two cell types or inhibiting Treg activity may promote cancer immunosurveillance and prevent hepatocellular carcinoma formation [37].

In summary, neutrophils recruited to inflammatory sites promote cancer initiation mainly by increasing DNA damage, angiogenesis and immunosuppression. However, the mechanism underlying neutrophil-dependent carcinogenesis is complicated and cannot be reduced to one specific molecule. Even the same molecule often exerts different effects on diverse stages. Although CXCR2 promotes neutrophil migration into pro-cancer sites, knockdown of CXCR2 in neutrophils increases ROS production and exerts pro-cancer effect [38]. Thus, in future studies, genetically engineered mouse models (GEMMs) will be extremely valuable for research in the field of cancer-related neutrophil biology, as they enable neutrophils and neutrophil-derived factors to be manipulated as cancer arises de novo.

The mechanisms by which neutrophils promote cancer growth are diverse. Neutrophils are characterized by rich granules, which perform different functions (Table 1). Some granule proteins (MMP-9 and ARG-1) released by activated neutrophils are associated with cancer progression. For example, MMP-9 released by neutrophils degrades the extracellular matrix, which in turn releases vascular endothelial growth factor (VEGF) and promotes angiogenesis [42]. Depletion of neutrophils or blockade of CXCR2 signaling to affect neutrophil recruitment inhibits cancer growth and reduces angiogenesis [43]. In contrast, an injection of cancer cells with neutrophils from cancer-bearing mice increases cancer growth and angiogenesis. In addition, the release of ARG-1 from neutrophils depletes arginine in T cells, causing the downregulation of CD3ζ. This process inhibits CD3-mediated T cell activation and proliferation, creating an immunosuppressive environment that also contributes to cancer growth [44]. In addition, the H⁺-pumping ATPase on tertiary granules causes cancer acidosis when it is mobilized to the cell surface, which may lead to cancer progression. Furthermore, an acidic pH inhibits the anticancer activity of T cells and natural killer (NK) cells, resulting in immune escape. Neutrophils also promote cancer growth and progression by recruiting macrophages and Tregs [45]. The structure of NETs formed by granule proteins and DNA induces the proliferation of cancer cells through high mobility group protein B1 (HMGB1) and NE [46‐48]. In hematological malignancies, levels of NETs are found to positively correlated with lymphoma progression or childhood acute leukemia development [49, 50].

In addition to granular proteins, neutrophils also play a role in promoting cancer growth by releasing growth factors, including epidermal growth factor, hepatocyte growth factor (HGF) and platelet-derived growth factor. Another study has shown that neutrophils eliminate senescence through IL-1 receptor antagonist (IL-1RA) and thus promote the progression of prostate cancer. Based on cancer promotion effect of neutrophil in pancreatic ductal adenocarcinoma (PDAC), lorlatinib inhibiting FES kinase, which is activated in neutrophils by PDAC cells, can attenuate cancer growth [104] (Fig. 1A).

Although fewer studies have assessed the inhibitory effects of neutrophils on cancer, very interesting data have been reported. For example, in models transplanted with different cancer cell lines or spontaneous cancer models, changes in neutrophil recruitment induced by specifically knocking out neutrophil MET, the HGF receptor, increase cancer growth [105, 106]. In mice transplanted with mouse mammary cancer virus promoter-driven polyomavirus middle T antigen (MMTV-PyMT) or MMTV-myc mammary cancer, neutrophils may exert a cytotoxic effect by producing H₂O₂ and subsequently inhibit cancer growth. Antibody-dependent cellular cytotoxicity (ADCC) during antibody therapy may be another mechanism by which neutrophils kill cancer cells [107, 108]. Neutrophils express Fcγ receptors, which mediate cancer cell elimination through ADCC. Depletion of neutrophils reduces the efficacy of treatment with anti-CD52 mAb (alemtuzumab) and anti-CD20 mAb (rituximab) in mouse lymphomas [109]. IgA induces the killing of cancer cells by neutrophils much more strongly than IgG [110]. In addition, neutrophils slow cancer growth by controlling microbial populations and cancer-associated inflammation [111]. However, since endogenous antibodies usually activate the anticancer effects of neutrophils via Fc receptors, researchers have not determined whether ADCC occurs in vivo in the absence of exogenous antibodies (Fig. 1B).

Remarkably, many studies using the same transplanted cell lines reported the opposite results. This discrepancy may be caused by the use of different experimental methods or sampling times in each experiment. For example, the different antibodies used to deplete neutrophils have different corresponding targets and efficiencies. Neutrophils will have different functions in different stages of cancer progression and will gradually transform from exerting anticancer effects to producing cancer-promoting effects. All of these factors may have led to inconsistent conclusions. Therefore, future studies should focus on how the context affects neutrophil function.

In the early stages of metastasis, neutrophils release MMP-9 to promote angiogenesis, playing an important role again by not only facilitating cancer growth but also providing more routes for cancer cells to escape. Neutrophils also direct cancer cells to endothelial cells, prompting them to enter the bloodstream. One mouse model of melanoma showed that cancer cells clustered around blood vessels and increased lung metastasis but had no effect on the growth of the primary tumor. In this model, cell damage increased HMGB1 levels, leading to the recruitment of neutrophils that subsequently promoted the migration of cancer cells toward blood vessels [117‐119]. In vitro, neutrophil-derived tumor necrosis factor (TNF) stimulates melanoma cell migration, suggesting that TNF is one of the factors related to neutrophil-induced metastasis.

Next, neutrophils guide cancer cells into blood vessels. Cathepsin G, a neutrophil-derived serine protease, induces cell migration, activates insulin-like growth factor 1, increases E-cadherin-mediated intercellular adhesion and cancer cell aggregation, and promotes cancer cell entry into blood vessels [120]. NETs trap circulating cancer cells (CTCs), helping them spread to distant sites and promoting their adhesion to distant sites [121, 122]. The interaction between neutrophils and CTCs promotes cell cycle progression in the blood and expands the metastatic potential of CTCs [123]. According to a recent study, ROS produced by neutrophils increase NETs, especially in obese cancer-bearing mice, which weakens endothelial junctions and promotes the extravasation of cancer cells[124]. In addition, several studies have shown that direct interaction between neutrophils and cancer cells activates neutrophils, increases the migration of cancer cells, promotes the anchoring of cancer cells to endothelial cells, and ultimately helps cancer cells exit blood vessels [123, 125].

Finally, metastatic cancer cells in distant tissues typically remain dormant for an extended period, during which infiltrating neutrophils release MMP-9 to promote angiogenesis, triggering the growth of dormant metastases. In addition, continued inflammation induces the formation of NETs, which are needed to wake dormant cancer cells. A mechanistic analysis has shown that two NEs and MMP-9, which are associated with NETs, cleave laminin. Cleaved laminin induces the proliferation of dormant cancer cells by activating α3β1-integrin signaling [72].

A related interesting phenomenon has been observed. Before disseminated cancer cells arrive, neutrophils accumulate in distant organs, forming the premetastatic niche. Neutrophils have been observed to aggregate in the lungs prior to the occurrence of metastasis in mouse models of MMTV-PyMT mammary cancer, breast cancer with nicotine exposure and melanoma, all of which are closely associated with the occurrence of pulmonary metastasis [79, 126, 127]. Neutrophils also contribute to ovarian cancer metastasis to the omentum by premetastic niche formation [128]. In cancer-bearing mice, cancer tissues modulate the microenvironment in the distal organ by releasing various cytokines, including vascular endothelial growth factor A (VEGFA), TNF, transforming growth factor-β (TGFβ), and G-CSF, to prepare for subsequent cancer metastasis [126, 129]. Blockage of neutrophil recruitment to the premetastatic sites or NET formation often prevents metastasis. However, whether targeting this phenomenon can prevent cancer metastasis into other organs or tissues, which are common as metastatic sites such as brain, breast, and lymph nodes, remains to be further investigated.

In contrast, other researchers have shown that neutrophil depletion facilitates metastasis. CCL2 and G-CSF secreted by the primary tumor activate the cytotoxic functions of these antimetastatic neutrophils mediated by H₂O₂. The type of tumor-entrained neutrophils is only observed in patients with cancer and not in healthy people; neutrophils migrate from primary breast tumor sites into the lung before metastatic cancer cells and then exert an inhibitory effect on metastatic colonization [130]. Neutrophils produce chemokines that recruit T cells and other leukocytes to indirectly kill cancer cells [131]. In a mouse model of breast cancer cell metastasis to the lung, the inhibitory effect of neutrophils required the presence of NK cells. In the absence of NK cells, the tumoricidal activity of neutrophils switched into metastatic facilitation [132]. Moreover, neutrophil expression of thrombospondin 1, IL-1β and the receptor tyrosine-protein kinase MET limit the formation of metastases by blocking the cancer cell mesenchymal-to-epithelial transition and releasing NO individually [69, 105, 133]. Neutrophils acquire the characteristics of antigen-presenting cells (APCs) in the early stage and thus might stimulate the proliferation of T cells to protect against tumor metastasis [134].

In cancer, neutrophils exert both pro-cancer and anticancer effects. The diversity of neutrophils is very common in cancer. A transcriptomic analysis revealed that tumor-associated neutrophils (TANs) and neutrophils from patients with cancer or cancer-bearing mice, which showed a higher proportion of neutrophil progenitors and a tendency toward immunosuppressive properties, differed significantly from those from healthy people or mice [166]. This diversity results from the high plasticity of neutrophils due to the effects of complex cancer microenvironments. The cancer and tissue microenvironments, conventional therapies and immunotherapy shape neutrophil function.

In a GEMM of lung adenocarcinoma, TGFβ polarized neutrophils in a cancer-promoting direction, and TGFβ blockade reversed the neutrophil protumor phenotype to an antitumor phenotype. These two types of neutrophils with opposite functions are named N2 and N1, respectively, which are similar and comparable to tumor-associated macrophages, such as M2 and M1 [167]. In the early stage of non-small cell lung cancer, the anticancer state of neutrophils is also induced by interferon-γ (IFNγ) and GM-CSF. Induced neutrophils indeed develop from immature progenitors through the negative regulation of the transcription factor Ikaros and acquire APCs properties, which as APC-like hybrid cells, promote T cell antitumor responses [152]. Another study has shown that hypoxia is a potent determinant of the TAN phenotype and direct neutrophil-cancer cell interactions. After the removal of hypoxia, the number of neutrophils recruited by the cancer decreased significantly, but the recruited cells were more effective at killing the cancer cells. This activity is mediated by the production of NADPH oxidase-derived ROS and MMP-9. At the same time, the ability of neutrophils to promote cancer cell proliferation, which appears to be mediated by their production of NE, is also reduced [155]. The general trend is that TANs belong to a network of anticancer cells in the early stages of carcinogenesis, but with cancer progression, neutrophil function shifts to immunosuppressive and cancer-promoting states.

Neutrophils among TANs with proven immunosuppressive function have been extensively studied and have been named granulocytic myeloid-derived suppressor cells (G-MDSCs) or polymorphonuclear myeloid-derived suppressive cells. G-MDSCs appear as neutrophils at different stages of maturation [168]. G-MDSCs flexibly adapt to the cancer microenvironment. The most important of these adaptations is the metabolic shift, which exerts a substantial effect on cell function.

Metabolic features include the upregulation of fatty acid transport protein 2 (FATP2) [143], increased levels of arginase I [169], high NADPH oxidase activity [155] and active NOS [170]. These factors have all been shown to inhibit T cell function. Regarding the accumulation of high lipids in cancer microenvironment, G-MDSCs increase the uptake of exogenous fatty acids through STAT3- or STAT5-mediated upregulation of lipid transport receptors. Increased fatty acid oxidation induces G-MDSCs to undergo metabolic reprogramming from glycolysis and become immunosuppressive [171]. Accordingly, inhibition of the neutrophil metabolic reprogramming by blocking fatty acid oxidation can synergize with the immunotherapeutic effect of T cells. Thus, neutrophils not only utilize diverse metabolic strategies to meet the energy requirements for survival but also exhibit functional alterations in cancer based on changes in the cancer microenvironment, such as decreased glucose levels, low oxygen pressure and low pH values [172]. Cancer can produce many factors such as IL-1β, CCL2, TGF-β, G-CSF and GM-CSF influencing innate immune cells, including neutrophils [173]. In particular, G-CSF, GM-CSF and IL-6 secreted by cancer and/or by stromal cells surrounding cancer cells induce potent activation of G-MDCs by activating the myeloid transcription factor C/EBPβ.

Researchers have attempted to identify neutrophil subsets. Specific surface markers proposed to identify neutrophil subsets in cancer include CD101 and CD177 [174, 175], which are associated with cancer regression, and CD117, PDL1, CD170, LOX1, CD84 and JAML [176], which are associated with T cell immunosuppression and cancer progression. In PDAC, the purinergic receptor P2RX-negative neutrophil subset exhibits immunosuppressive role with enhanced PD-L1 expression and mitochondrial metabolism [177]. However, an unequivocal method to detect immunosuppressive neutrophils and other neutrophil subsets using flow cytometry or other strategies remains to be developed. Since the subsets of neutrophils show continuous changes and are highly phenotypically and morphologically similar (even between MDSCs and other cells), a reasonable assumption is that these hypothetical subsets are actually the same type of cells, with larger or smaller changes induced by different local environments. These neutrophils are a single cell type with many different functional phenotypes. The high plasticity of neutrophils enables them to respond quickly to external stimuli, leading to their heterogeneity. Because different stimuli mobilize different cytoplasmic granules, different degrees of exposure of the membrane proteins of each granule to the cell surface can change the cell surface composition of neutrophils, potentially leading to the misidentification of new cell types. Taken together, TANs appear to be more flexible than circulating neutrophils, which enables them to adapt to diverse cancer microenvironments.

Moreover, TANs or normal neutrophils have consistently been shown to be trained to become anticancer neutrophils through various methods to achieve the goal of killing cancer cells. For example, PPM1D/Wip1 is a negative regulator of the cancer suppressor p53 and is overexpressed in several human solid cancers. Ppm1d knockout or chemical inhibition of Wip1 in human or mouse neutrophils exacerbates anticancer phenotypes and increases p53-dependent expression of costimulatory ligands and the proliferation of cocultured cytotoxic T cells [178]. Another study showed that exposure to β-glucan [179], a fungal-derived prototype agonist of trained immunity, trained neutrophils in mice to enhance the anticancer activity of neutrophils. These results, in turn, prove that neutrophils are highly plastic (Fig. 1C).

Cancer is highly heterogeneous and is considered one of its hallmarks. The tumor contains cancer cells and noncancerous cells such as neutrophils, macrophages, T cells, adipocytes, stromal cells and others constituting the microenvironment. All these cells communicate directly or indirectly. Thus, neutrophils in cancer not only have a relationship with the T cells mentioned above but also affect or are affected by other cells. During advanced colorectal cancer progression, cancer stem cell-derived exosomes containing triphosphate RNAs prime neutrophils for cancer development and depletion of neutrophils with antibodies attenuate the tumorigenicity of these cancer stem cells [180]. In obese patients with pancreatic cancer, crosstalk among pancreatic stellate cells, neutrophils and adipocytes mediated by IL1β promotes PDAC. Genetic or pharmacological targeting of this circuit provides a potential method for pancreatic cancer treatment [181]. Cancer-associated fibroblasts are considered one of the important stromal cells contributing to cancer development. A recent report identified that one of the underlying mechanisms as NET induction. This induction is driven by increased amyloid and β-secretase expression in fibroblasts [182].