Neutrophil-derived reactive oxygen species promote tumor colonization

A single-nucleotide polymorphism of neutrophil cytosolic factor 1 (Ncf1), leading to an impaired generation of reactive oxygen species (ROS), is a causative genetic factor for autoimmune disease. To study a possible tumor protection effect by the Ncf1 mutation in a manner dependent on cell types, we used experimental mouse models of lung colonization assay by B16F10 melanoma cells. We observed fewer tumor foci in Ncf1 mutant mice, irrespective of αβT, γδT, B-cell deficiencies, or of a functional Ncf1 expression in CD68-positive monocytes/macrophages. The susceptibility to tumor colonization was restored by the human S100A8 (MRP8) promoter directing a functional Ncf1 expression to granulocytes. This effect was associated with an increase of both ROS and interleukin 1 beta (IL-1β) production from lung neutrophils. Moreover, neutrophil depletion by anti-Ly6G antibodies increased tumor colonization in wild type but failed in the Ncf1 mutant mice. In conclusion, tumor colonization is counteracted by ROS-activated and IL-1β-secreting tissue neutrophils. Zhong et al. investigate how mutations in the NOX2 subunit Ncf1 lead to reduced ROS formation and affect B16F10 lung metastasis. The study highlights neutrophils as key to detrimental ROS-formation and suggest the involvement of IL-1 beta and RAGE.

I mmunotherapy has proven successful to reduce the cancer mortality in metastatic melanoma 1 , although the 5-year survival rates of patients with advanced melanoma are usually still <30% (ref. 2 ). It highlights the importance to study the role of earlier steps in immune responses to tumor colonization.
A recent finding is the identification of a single-nucleotide polymorphism in the Ncf1 gene to be a genetic basis for autoimmunity [3][4][5] . The Ncf1 gene encodes neutrophil cytosolic factor 1 (NCF1, earlier denoted p47 PHOX ), a cytosolic component of the NADPH oxidase 2 (NOX2) complex 6 . NCF1 is required for the NOX2 complex to trigger oxidative burst, i.e., an induced production of reactive oxygen species (ROS) 7 , and identified Ncf1 mutations result in a decrease of ROS production 8,9 . NCF1 is highly expressed in neutrophils and CD68-positive cells 10 . CD68expressing cells, predominantly macrophages, are a critical mediator regarding autoimmune disease susceptibility, exhibiting the properties of both innate and adaptive immune responses. The downstream effect of the Ncf1 mutation includes regulation of T cells in autoimmune arthritis 10,11 and encephalomyelitis 12 , exaggerated type I interferon signaling in lupus autoimmunity 13,14 , innate interleukin 17 (IL-17) induction in psoriatic diseases [15][16][17] , and impairment of the capacity to form neutrophil extracellular traps in patients with gout 18 and gallstones 19 .
The induced immune responses by Ncf1 mutations may on the other hand prevent cancer. It has been shown that myeloid cells producing ROS could suppress the cytotoxic T-cell response to tumor cells 20 . We and others have found that impaired NCF1 and NOX2 function are protective against the melanoma development 21,22 . We have now chosen the shared melanoma model B16F10 cell line, supplied by ATCC, to probe the mechanisms of cell-specific Ncf1 mutations in tumor colonization.
In this study, we transferred the Ncf1 mutation (Ncf1 */* ) into mouse strains with different major histocompatibility complex (MHC) haplotypes on the C57BL genetic backgrounds. Using the MHC congenic mice in the mouse melanoma model, we observed a reduced number of lung colonies in Ncf1 mutant mice. We excluded a critical role of T cells, B cells, and macrophages, but found that neutrophils were the most important cell for the Ncf1 mutation-mediated protection. MRP8-restricted expression of a functional Ncf1 restored the release of IL-1β from neutrophils and the susceptivity to tumor colonization. Thus, this study provides insights into the role of NOX2-derived ROS in driving tumor colonization through a neutrophil and IL-1β-dependent pathway.

Results
Protection despite the deficiencies of T and B cells. To study the role of redox regulation of tumor colonization, we introgressed the Ncf1 mutation in mice with the MHC b, q, and r haplotypes on the shared C57BL genome background. Lung tumor foci were counted at day 10 after intravenous inoculation of B16F10 cells. We found that Ncf1 mutant mice (Ncf1 */* ) were protected from metastatic colonization, regardless of the MHC haplotype (Fig. 1a), arguing against redox control of allogenic-mediated rejection.
Susceptibility to colonization restored by neutrophils. In the continuous search for the critical cells in mediating the Ncf1 regulatory role, we investigated the role of neutrophils. We designed a conditional knock-in mouse model expressing a functional copy of the Ncf1 gene in neutrophils by a Cre under control of the MRP8 promoter on Ncf1-deficient background (Ncf1 TN3/* .Mrp8 Cre/+ ). In prior to experiments, we validated the conditional expression of NCF1 in bone marrow neutrophils sorted from naive Ncf1 TN3/* .Mrp8 Cre/+ mice (Fig. S1a). We observed oxidative burst of blood neutrophils in the naive Ncf1 TN3/* .Mrp8 Cre/+ (Fig. 2a), whereas the ROS production of blood monocytes was not restored (Fig. S1b). We performed the tumor colonization assay with the knock-in mice. A higher number of tumor colonies were found in the lungs at day 10 of Ncf1 TN3/* .Mrp8 Cre/+ mice (Fig. 2b).
Previously, we and others have found a role of ROS in the regulation of IL-1β expression 12,23 . In this study, we show that ROS-sufficient neutrophils express an enhanced IL-1β in conditional knock-in mice (Fig. 2c, d). Consistently, expression of IL-1β was enhanced in lung neutrophils from wild-type mice (Ncf1 +/+ ; Fig. S1c), as well as CD54 known as a molecular bridge 24 at the interaction between tumor cells and neutrophils (Fig. S1d). These results suggest that neutrophils play a critical role to promote lung colonization by ROS through activating IL-1β secretion.
Downstream mediators through IL-1β signals. We next studied the effects of IL-1β in the tumor colonization model. For this purpose, we collected bronchoalveolar lavage fluid (BALF) at day 10 after intravenous injection of B16F10 cells. The analysis of flow cytometry data shows a decreased level of IL-1β in BALF from Ncf1 mutant mice (Ncf1 */* ; Fig. 3a). In addition, we observed the cytokine expressions of TNFα, IFN-γ, IL-4, and IL-1α in the same BALF, but there was no significant difference between Ncf1 mutant group and wild-type controls (Fig. S1e).
To address a possible functional role, we treated the mice with recombinant mouse IL-1β (rmIL-1β) half an hour prior to the inoculation of B16F10 cells. We show that IL-1β treatment enhanced lung colonization, independently of the Ncf1 polymorphism (Fig. 3b). The oxidative burst capability of lung neutrophils was also not affected by injection of rmIL-1β (Fig. 3c). Interestingly, we observed a reduced number of mature natural killer (NK) cells (CD45 + CD3 − NK1.1 + ) in lungs (Fig. S2a), i.e., CD27 − KLRG1 + NK cells in an Ncf1-dependent manner, after a single injection of rmIL-1β in vivo (Fig. S2b, c). Mature NK cells can have a major role in controlling B16F10 lung colonization 25 , and then we studied the effect of NK cells by injecting NK1.1 antibody [26][27][28] . As expected, depletion of NK cells leads to an increase of the lung colonization of B16F10 tumor cells (Fig. S2d). However, the Ncf1 mutation-mediated protection was consistent regardless of NK cell depletion (Fig. S2d). These results suggest that IL-1β promote tumor colonization as a downstream effect of ROS. NK cells may play a role in IL-1β signaling, but not critically in regulating the Ncf1 effect.
Enhanced colonization after neutrophil depletion. Our data indicate that neutrophils play a critical role in ROS-mediated tumor colonization. To deplete neutrophils, we injected anti-Ly6G antibodies (clone 1A8) in wild-type mice, MN mice, and Ncf1 mutant mice, and evaluated neutrophil infiltration at day 10 in the lungs of B16F10-exposed mice (Fig. 4a). The numbers of lung neutrophils (CD45 + CD11b + Gr1 + Ly6C mid ) were reduced in all strains (Fig. 4b, c), while an increased number of monocytes (CD45 + CD11b + Gr1 mid Ly6C hi ) was observed in wild-type mice (Fig. S3a). Contrary to the previous results that neutrophils depletion with anti-GR1 antibody (clone RB6-8C5) reduced the tumor growth 29,30 , and that 1A8 antibody treatment had no impact on lung colonization in C57BL/6 mice 31 , our data shows that 1A8 antibodies enhanced lung colonization in Ncf1 wild-type mice on the C57BL/6 genetic background. Importantly however, this phenomenon was Ncf1 dependent as it was not seen in either Ncf1 mutant or MN mice (Fig. 4d). A possible contribution to an increase of tumor colonies could be the expression of the receptor for advanced glycation endproducts (RAGE) on monocytes from wild-type mice (Fig. S3b). RAGE is known to be a receptor of MRP8 at the interaction between tumor cells and myeloid cells 24,32 . To determine the role of RAGE in oxidative regulation, we treated mice with the RAGE blocker FPS-ZM1 (ref. 33 ). Interesting, FPS-ZM1 reduced lung colonization in wild-type mice to the same low level as in Ncf1 mutant mice (Fig. S3c, d). These results suggest that MRP8-RAGE signaling could activate monocytes to support tumor colonization after the ROS-sufficient neutrophil depletion.

Discussion
A natural mutation in the Ncf1 gene leading to an impaired production of ROS by the NOX2 complex protects against B16 melanoma tumor colonization of the lungs. The protection was reversed by conditional expression of a functional Ncf1 gene in neutrophils. Subsequently, the neutrophil-derived IL-1β was demonstrated to promote lung colonization of B16F10 cells.
ROS could be produced by both tumor cells and infiltrating immune cells during metastatic colonization, and the location of ROS may determine oxidative effect on tumor growth. A recent study provided evidence that metabolic ROS support the metastasis process of tumor cells in clusters, compared with single cells 34 . The cluster could also be formatted between tumor cells and neutrophils. A phenomenon was reported that neutrophil-tumor clusters exhibit the capability to expand the metastatic potential of B16F10 cells 24 , but the role of ROS was not addressed. Normally, the NCF1-NOX2 complex is efficiently activated, dominating to the production of ROS in neutrophils. In this study, we validated with mouse models for the first time that a copy of functional Ncf1 gene expression in neutrophils driven by the MRP8 promoter was enough to increase the level of lung metastatic foci.
ROS deficiency on tumor colonization, increasing tumor colonies to a high level. Recently, it has been reported that deficiency of the developmental endothelial locus-1 in the host may promote neutrophil accumulation in the lung metastatic niche, where neutrophil-released ROS mediates a direct interaction with B16F10 cells 40 .
Another direct interaction between neutrophils and tumor cells could involve the RAGE receptor on B16F10 cells, of which the MRP8 is a potent ligand 41 . The myeloid cells, including neutrophils, monocytes, and the interstitial macrophage-like phenotype, might express both the NCF1-NOX2 complex and MRP8, referred to the formation of the pre-metastatic microenvironment Mrp8 +/+ , n = 7) to be analyzed at day 10 after intravenous injection of B16F10 cells. All results are shown as mean ± SEM. Mann-Whitney U test, *p < 0.05, **p < 0.01. Fig. 3 Administration of rmIL-1β increases lung colonization in Ncf1 mutant mice. The bronchoalveolar lavage (BAL) fluids and lungs were harvested from wild-type mice (Ncf1 +/+ ) and Ncf1 mutant mice (Ncf1 */* ) at day 10 after intravenous injection of B16F10 cells. a The level of IL-1β in the BAL fluid was assessed by flow cytometry. b The number of tumor colonies in lungs per mouse was counted in wild-type mice with injections of rmIL-1β (n = 5) versus PBS (n = 11), and Ncf1 mutant mice with injections of rmIL-1β (n = 8) versus PBS (n = 11). c ROS production by neutrophils in the lungs was assessed by flow cytometry with DHR, after these cells were incubated ex vivo with PMA and DMSO as the control. The lung tissues were collected from wild-type mice with injections of rmIL-1β (n = 5) versus PBS (n = 5), and NCF1-deficient mice with injections of rmIL-1β (n = 5) versus PBS (n = 5) at day 10 postinjection of B16F10 cells. All results are shown as mean ± SEM. Mann-Whitney U test, *p < 0.05, **p < 0.01 and ***p < 0.001. that is required for blood tumor cells to engraft at the lung niche 42 . Thus, it is difficult to distinguish the neutrophils from monocytes/macrophages on the immunoregulation. Expression of a functional Ncf1 gene in CD68-positive cells can reduce the pathogenicity of T cells in experimental autoimmune arthritis models 10 . In this study, we show that the Ncf1 mutation protected from tumor colonization, irrespective of the CD68-expressing cell status. What is more, depletion of ROS-sufficient neutrophils with 1A8 antibodies was likely to trigger the ROS-dependent RAGE expression on Ly6C monocytes.
In addition, NK cells could play a prominent role in the protection against lung colonization. Aydin and colleagues studied the role of NK cells in Nox2-knockout mouse models by anti-NK1.1 antibody injection, finding an NK cell-associated antimetastatic effect of histamine dihydrochloride in Nox2 wild-type mice, but not in Nox2-knockout mice 26,27 . Li and colleagues investigated the role of the cytokine G-CSF in tumor models on use of both NOD-scid and NSG mice, suggesting a neutrophilmediated role of NK cells in lung metastasis 28 . It is known that innate lymphoid cells are also deficient in NSG mice 43 . To study the importance of neutrophils, Boivin and colleagues found that neutrophils could even expand in blood and spleens after anti-Ly6G treatments at a low dose 29 . Clearly, NK cells play an important role in the protection of lung colonization, but it remains unknown whether NK cells are involved with the protective effect of the NCF1-NOX2 complex. It should be pointed out that the small compound histamine dihydrochloride might not specifically affect NK cells, and there is also genetic difference between the knockout mouse strains and naturally mutant strains. For an example, both Nox2-knockout mice and Ncf1knockout mice have been reported to be protected from experimental autoimmune encephalomyelitis 44 , but the Ncf1 mutant mice exhibited an enhanced autoimmune encephalomyelitis 12,45 . Therefore, we further studied the role of NK cells in the Ncf1 mutant mice using anti-NK1.1 antibody 26 . However, we found that NK cell depletion, by anti-NK1.1 treatments, enhanced tumor metastasis independent of the Ncf1 mutation.
In summary, we used naturally mutant mouse models of the Ncf1 gene to screen the cell-type-specific effect of the NOX2 complex deficiency, and then found out that only Ncf1 expression in neutrophils can restore the susceptibility to tumor colonization. We found that the Ncf1 competent neutrophils, exhibiting the capability of functional ROS induction and IL-1β signaling, can promote lung colonization. As the Ncf1 expression is polymorphic in both humans and experimental animals, and since redox regulation is a key pathophysiologic process, this observation could be of key importance not only to understand the basis of tumor metastasis, but also to improve cancer treatment.
Administration of RAGE blocker FPS-ZM1. Mice were administrated intraperitoneally with 25 μg FPS-ZM1 (Merck, Catalog Number 553030, Sweden) dissolved in 200 μL of PBS per time. PBS was used as a control. Mice received FPS-ZM1 treatments on days −1, 0, 1, and 2 after intravenous injection of B16F10 cells.
Flow cytometry analysis. The single-cell suspension derived from blood and lung tissues was analyzed with flow cytometry. To prepare the single-cell suspensions, the lungs were dissected into smaller fragments and digested in PBS with 2 mg/mL collagenase type IV (Sigma, C5138, CAS no. 9001-12-1, Sweden), 100 U/mL DNase I (Roche, Sigma, Catalog Number 04716728001, Sweden), and 2 mM EDTA (Sigma, CAS no. 60-00-4, Sweden). The cell density was counted by using Sysmex KX-21N automated hematology analyzer (Sysmex Corporation, NY). The cell samples were stained with a LIVE/DEAD® fixable near-IR dead cell stain (Ther-moFisher, catalog no. L10119, Sweden). After an anti-mouse CD16/CD32 Fc block, extracellular antigens were stained 20 min at 4°C in PBS with 1% fetal bovine serum (FBS, Gibco, ThermoFisher, catalog no. 26140079, USA). To measure intracellular ROS, the staining of 3 μM DHR (ThermoFisher, Catalog Number D23806, Sweden) was conducted, respectively, after cell surface markers staining, followed by stimulation of 100 ng/mL of PMA (Sigma-Aldrich Co., CAS No. 16561-29-8, Sweden) for 30 min. To detect the intracellular expression of cytokines, the cells were stimulated with 100 ng/mL of PMA and 1 μg/mL of ionomycin in the presence of 5 μg/mL of brefeldin A (BFA) for 4 h at a humidified 37°C, 5% CO 2 incubator. The stock solutions of PMA, ionomycin (ThermoFisher, Catalog Number I24222, Sweden), and BFA (ThermoFisher, catalog no. B7450, Sweden) were prepared with dimethylsulfoxide (DMSO, Sigma-Aldrich Co., CAS no. 67-68-5, Sweden). For intracellular cytokine staining, cells were fixed and permeabilized using BD cytofix/cytoperm solution (BD Biosciences, catalog no. 554714, Sweden). Samples were acquired using BD LSR II flow cytometer. The workstation is managed by FACSDiva software version 8.0 (BD Biosciences), and the data were analyzed using the FlowJo software version 10.5.3 (TreeStar, Inc., OR).
Statistics and reproducibility. Statistical analyses were performed with Graph Prism software, version 8.2.1 (GraphPad Software, San Diego, USA). Unless otherwise stated, the Mann-Whitney U test was used to compare the means of two groups. All results are shown as mean ± standard error of the mean. P value < 0.05 was considered as significant: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
Study data are included in the article and the Supplementary Materials. No datasets were generated or analyzed during the current study.