Glutamate blunts cell‐killing effects of neutrophils in tumor microenvironment

Abstract Neutrophils are the first defenders of the innate system for injury and infection. They have gradually been recognized as important participants in tumor initiation and development due to their heterogeneity and plasticity. In the tumor microenvironment (TME), neutrophils can exert antitumor and protumor functions, depending on the surroundings. Tumor cells systemically alter intracellular amino acid (AA) metabolism and extracellular AA distribution to meet their proliferation need, leading to metabolic reprogramming and TME reshaping. However, the underlying mechanisms that determine how altered AAs affect neutrophils in TME are less‐explored. Here, we identified that abundant glutamate releasing from tumor cells blunted neutrophils’ cell‐killing effects toward tumor cells in vitro and in vivo. Mass spectrometric detection, flow cytometry, and western blot experiments proved that increased levels of pSTAT3/RAB10/ARF4, mediated by glutamate, were accompanied with immunosuppressive phenotypes of neutrophils in TME. We also discovered that riluzole, an FDA‐approved glutamate release inhibitor, significantly inhibited tumor growth by restoring neutrophils’ cell‐killing effects and decreasing glutamate secretion from tumor cells. These findings highlight the importance of tumor‐released glutamate on neutrophil transformation in TME, providing new possible cancer treatments targeting altered glutamate metabolism.

drive the alteration of catalytic enzyme expression and activity. 5 Several AA-related receptors and transporters are also associated with cancer progression in tumor cell-autonomous and nonautonomous ways. 2,6,7 In addition to their direct role of protein synthesis in tumor cells as materials, AAs in TME are also involved in energy generation and tumor-infiltrated immune cell training, providing new ways for cancer treatment targeting AA metabolism.
The TME is tightly sculpted by tumor cells through the release of various chemokines and metabolites. 8 This leads to systemic reprogramming of surrounding cells, including fibroblast, endothelial, innate, and adaptive immune cells. 9 Neutrophils are the first defenders of the innate system for injury and infection. Neutrophils have been gradually recognized as important participants in tumor initiation and development due to their heterogeneity and plasticity. 10 Neutrophil-contained granules and proteins are shown to kill tumor cells and facilitate tumor cell migration simultaneously. [11][12][13] The role of neutrophils in TME remains controversial, with evidence for antitumor and protumor functions depending on the surroundings. 14 Understanding the molecular mechanisms that drive neutrophils' plasticity and the switch between pro-and antitumor effects is needed urgently. Based on previous evidence that AA metabolism was tightly associated with neutrophil functionality, 15,16 we hypothesized that altered AAs in TME might be one of the factors that influence neutrophil phenotype.
In this study, we found that tumor cells secreted glutamate in vitro and in vivo. High-dose glutamate stimulation on neutrophils dampened cytotoxicity toward tumor cells. Specifically, glutamate upregulates the expression of pSTAT3, ARF4, and RAB10 in neutrophils, along with decreased cytotoxicity and inhibitory effects on T cell proliferation. Moreover, riluzole (glutamate release inhibitor) effectively restored neutrophil cytotoxicity toward tumor cells by reducing glutamate secretion in TME. These results indicate that glutamate was important tumor-released AAs for neutrophil transformation in TME, which provided the possible mechanisms for neutrophil plasticity in tumors.

| Mice and treatments
Male C57/BL6 mice were purchased from TengXin Experimental Animal Company. All mice were fed and maintained at the animal facility of Army Medical University. All experiments were carried out with weight-matched (18-22 g) male mice and approved by the Ethics Committee of Army Medical University.

| Tumor cell line culture
Three murine tumor cell lines, melanoma (B16), Lewis lung carcinoma (LLC), and colon adenocarcinoma (MC48), were purchased from ATCC. Dulbecco's modified Eagle's medium (11965) and FBS (10100) were purchased from Thermo Fisher Scientific. All murine tumor cell lines (LLC, MC38, and B16) were cultured with DMEM (10% FBS supplemented). Luciferase-containing lentiviral vectors (79943; BPS Bioscience) were coated and transfected with tumor cell lines LLC, MC38, and B16. The cells that stably expressed luciferase were selected by ampicillin and passaged for the following in vitro cytotoxicity effects.

| Neutrophil isolation
Neutrophils from blood were purified by using LY6G MACS

| In vitro cytotoxicity assay
The cytotoxicity effects of neutrophils were determined as previously described. 18

| Glutamate detection
The glutamate concentration was determined by a fluorometric glutamate assay kit (STA-674; Cell Biolabs). For cell culture medium (10 6 cells in 2 ml culture medium), supernatants were centrifuged at 6800 g for 10 min and diluted with assay buffer. For tumors, the tissues were homogenized and sonicated with hydrolysis buffer (STA-674, 200 μl for 20 mg tissue; Cell Biolabs). The glutamate concentration was calculated by comparing the sample RFU value to the standard curve.

| Immunofluorescence
Immunofluorescence analysis was carried out as previously described. 19 Briefly, isolated neutrophils were smeared on glass slides and incubated with primary Abs diluted with 5% BSA containing 0.1%

| Mass spectrometry
Murine neutrophils were isolated from blood using the Percoll gradient method 20 (P4937, pH 8.5-9.5; Sigma) and treated with/without glutamate (100 μM, G0355000, CAS 56-86-0; Sigma) for 24 h. Then 10 7 neutrophils were lysed in 200μl lysis buffer combined with sonication on ice. After digestion with trypsin (V5280; Promega), samples were shifted to Luming Biological Technology for the following steps, according to a previous study 21 : iTRAQ labeling and hydrophilic interaction chromatography fractionation, LC-MS/MS, and data analysis. Proteins with changes greater than 1.5-fold and p < 0.05 were considered differentially expressed.

| Tumor Immune Single-cell Hub analysis
Tumor Immune Single-cell Hub provides detailed cell-type annotation at the single-cell level, enabling the exploration of TME. 22 In this study, single cell-sequencing data of human NSCLC were retrieved and compared (GSE127465 and GSE99254), and mRNA relative expression of intended genes in different cell types from the GSE127465 dataset are presented. The mRNA expression levels are shown as log 2 values of TPM.

| Gene Expression Profiling Interactive Analysis database
We tested the mRNA correlation between GLS and human neutrophil marker genes (CD11b/ITGAM and CD66b/CEACAM8) from human LUAD and LUSC samples in the GEPIA database, as well as comparable normal tissues. The Pearson correlation coefficient values (R) and p values were analyzed and presented according to a previous study. 23

| Statistical analysis
All experiments were repeated at least three times in duplicate.
Results are presented as mean ± SEM. When comparing two groups, p values were calculated using two-tailed Student's t-tests. For multiple treatment group comparisons, significance was determined by one-way ANOVA, followed by Tukey's post hoc multiple comparisons test. All statistical analyses were undertaken with GraphPad Prism software.

| Glutamate is released from tumor cells in TME
We analyzed the AA release from three murine tumor cell lines (LLC, MC38, and B16) in vitro using the High-Speed Amino Acid Analyzer AminoSAAYA. According to the standardized AA sample curve, 20 AAs/metabolites were detected in cell medium ( Figure S1).
After 48 h, supernatants were collected for the AA detection and cells were harvested for cell recounting. The standardized data ( Figure 1A) showed that Glu and Ala were concordantly increased after culturing. We then measured Glu levels from tumor tissues and serum of tumor-bearing mice (LLC, MC38, and B16). Using the fluorometric glutamate assay kit assay, we analyzed the Glu concentration in tumors and paratumor tissues (control). As shown in Figure 1B, glutamate levels in tumors were significantly increased compared to control groups, indicating that tumor tissues were glutamate-abundant. We did not see statistical differences in Glu levels in serum ( Figure 1C), whereas the glutamate concentration in tumor tissue was increased approximately 2-7-fold compared to serum.
L-glutamate is converted from L-glutamine by GLS. 24 The increased L-glutamine consumption and GLS expression driven by oncogenes (cMyc) are hallmarks of cancer cells, 25 leading to excessive glutamate production ( Figure 1D). Tumor-released glutamate was shown to promote tumor cell proliferation in an autocrinedependent manner. 26,27 However, the effect of extracellular glutamate on neutrophils in TME is still unknown. Using TISCH (http://  Figure 1F) and CEACAM8 (CD66, r = 0.26, p < 0.0001; Figure 1G) in tumor tissues, which were not observed in normal tissues. Here, we proved that tumor cells were capable of secreting glutamate in vitro and in vivo, making TME into a glutamate-abundant environment.

| Tumor cell-killing effect by neutrophils is significantly blunted in TME, which could be restored by glutamate release inhibitor riluzole
Neutrophils show both protumor and antitumor effects, depending on various surroundings. 10,11,[27][28][29] A previous study and our data both showed that glutamate was abundant in tumor tissues. 2,28 We hypothesized whether glutamate could modulate neutrophil functions in tumor. Hence, we compared the cytotoxic effects between bloodderived neutrophils and tumor-derived neutrophils. The isolated neutrophils were cocultured at varying ratios with luciferase-labeled tumor cells. The blood-derived neutrophils presented obvious cytotoxic potential in a ratio-dependent manner, which was not observed in intratumoral neutrophils (Figure 2A-C). This phenomenon was also observed previously, [29][30][31] indicating TME possibly switched the neutrophil phenotype. In order to explore glutamate effects on neutrophils, we used glutamate release inhibitor riluzole 32 to treat LLCtumor bearing mice. As shown in Figure 2D-F, riluzole treatment significantly inhibited glutamate release from tumor cells in vitro and in vivo. LY6G + intratumoral neutrophil cells from control mice were found to be noncytotoxic up to a ratio of 1:10, whereas neutrophils isolated from the tumors of riluzole-treated mice showed dose-dependent cytotoxicity, with more significantly killing effects than control groups at ratios of 1:10 and 1:50 ( Figure 2G). These data preliminarily showed that glutamate release blocked by riluzole increased neutrophil cytotoxic effects.

| Glutamate decreases neutrophil cytotoxicity toward tumor cells in vitro
We next used glutamate-pretreated neutrophils to coculture with tumor cell line LLC. As shown in Figure 3A

| Expression of pSTAT3, ARF4, and RAB10 highly induced after glutamate treatment in neutrophils
We next used mass spectrometry to screen out differential proteins in neutrophils after glutamate treatment. As shown in Figure 4A- patients. The data revealed that all these genes were widely distributed ( Figure 4D-G). For neutrophils in TME ( Figure 4D-G, red circle), the SCaMC-1 mRNA level was extremely low, while STAT3, ARF4, and RAB10 levels were considerably high. We also used immunofluorescence and western blot analyses to confirm this. As Figure 5A,B shown, protein levels of ARF4, RAB10, and pSTAT3 (functional form of STAT3) in neutrophils were significantly induced by glutamate. Also, levels of these three proteins in neutrophils from LLC tumor tissues were strikingly increased compared to bloodderived neutrophils ( Figure 5C). Quantified data ( Figure 5D,E) supported that this induction was consistent with increased glutamate concentration in tumor tissues ( Figure 1B,C) and decreased neutrophil cytotoxic effects (Figure 2A-C). Specifically, Vasquez-Dunddel et al. 39 confirmed that activated STAT3 drove neutrophils into an immunosuppressive phenotype by enhancing arginase-I activity directly. Fan et al. 40 and Gu et al. 41 also proved that RAB10 phosphorylation in neutrophils leads to activation of leucine rich repeat kinase 2, which was associated with poor prognosis in cancer patients. The detailed mechanisms that explain how STAT3/ARF4/ RAB10 influenced neutrophil-mediated cell killing effects were the subject of the following investigations.

| Riluzole decreases pSTAT3 level on neutrophils and increases CD8 + T cell frequency in TME
The above data suggested that glutamate abundance in TME was possible for suppressing neutrophil antitumor effects. Riluzole treatment significantly inhibited glutamate release, which was speculated to reshape immune responses in TME. To further investigate how riluzole influenced the TME and the antitumor response, we characterized neutrophils and T cells in an LLC tumor model by flow cytometry. As shown in Figure 6A,B, the number of neutrophils (CD11b + Gr-1 + ) was significantly decreased in LLC tumors after riluzole treatment. The intracellular pSTAT3 levels of neutrophils in TME were also dramatically decreased after riluzole treatment ( Figure 6C,D). The pSTAT3 levels of neutrophils from blood were relatively low compared to those from The TME reshaping after riluzole treatment, especially on neutrophils and CD8 + T cells, was possibly attributable to decreased glutamate release in tumor tissue.

| Antitumor effects of riluzole is partially dependent on the presence of neutrophils and CD8 + T cells in TME
Our data ( Figure 7A) and other studies both showed that riluzole inhibited tumor growth in a dose-dependent manner. 17,33,35 However, the antitumor effects in SCID mice (T and B cell deficiency) were F I G U R E 3 Cytotoxic effect of neutrophils on tumor cells mediated by riluzole and glutamate. (A) Cell viability of LLC after coculturing with glutamate treated neutrophils at different ratios. (B) Cell viability of LLC after coculturing with riluzole pretreated neutrophils at different ratios. */**Significant differences between glutamate (100 μM) and blank neutrophil groups. (C) Cell viability of three tumor cell types after riluzole treatment for 24 h. (D-F) Cell viability of LLC/B16/MC38 cocultured with/without neutrophils at 1:10 ratio combined with gradient doses of CB-839 treatments (0-2.5 μM) for 24 h. **Significant differences between tumor cell alone groups and neutrophil coculture groups. Riluzole (2.5 μM) or glutamate (100 μM) pretreated neutrophil/tumor cells for 6 h and resuspended with complete DMEM. Subsequently, these pretreated neutrophils/tumor cells were cocultured at different ratios for 24 h. Tumor cell viability was determined by luciferase signal intensity. (G-I) Cell viability of LLC/B16/MC38 after coculturing with neutrophils. */**Differences between riluzole group and blank group. #/##Differences between glutamate group and blank group. n = 5 reduced when riluzole was injected alone. 17,34 Given that glutamate stimulation was capable of modulating T cell activation and neutrophil migration, 42,43 these two populations were both tightly engaged with immune responses against cancer. 8, 10 We explored whether riluzole-mediated tumor growth inhibition was dependent on neutrophils and CD8 + T cells. As shown in Figure 7B, injection of either anti-LY6G or anti-CD8 mAb dramatically abolished the tumor growth inhibition in riluzole-treated tumor-bearing mice. This implied that both neutrophils and CD8 + T cells were essential to the antitumor effects of riluzole. One of the immunosuppressive potentials of neutrophils was through suppressing stimulated T cell proliferation. 44 Given the fact that riluzole alone did not influence CD8 + T cell proliferation rate directly ( Figures 7C and S4), we wondered whether neutrophils from riluzole-treated tumors could affect CD8 + T cell proliferation. We assessed the proliferation of anti-CD3/CD28stimulated T cells cocultured with neutrophils either from blood or tumor tissues for 24 h. When the neutrophil/CD8 + T cell coculture ratio was 1:4 and 1:2, tumor-derived neutrophils dramatically decreased the T cell proliferation rate compared to blood-derived neutrophils ( Figure 7D). The inhibitory effects were significantly blocked after coculturing with neutrophils from riluzole-treated tumor tissues. These results proved that riluzole treatment increased CD8 + T cell proliferation by decreasing neutrophil suppressive effects. The presence of neutrophils and CD8 + T cells were indispensable for riluzole-mediated tumor growth inhibition.

| DISCUSS ION
In this study, we established the connection between glutamate release and neutrophil reshaping in TME. Abundant glutamate re-  Figure 1E). Purple bar represents log 2 (transcripts per million) of intended gene neutrophils forming immunosuppressive phenotypes (decreased cytotoxicity toward tumor cells and increased inhibitory effects on T cells) in TME. Moreover, riluzole (glutamate release inhibitor) effectively restores the antitumor effects of neutrophils by reducing glutamate abundance in TME. This confirms the important roles of glutamate in neutrophil transformation and TME modulation.
Glutamate, an important excitatory neurotransmitter in the brain, is also a major bioenergetic substrate for proliferating neoplastic cells. 24 A previous study confirmed that glutamate over-release from glioma cells is associated with fast tumor growth. 27 Here, we further prove that glutamate release is prevailing in tumor cells, which can possibly be attributed to the enhanced glutamate uptake/release systems in tumor cells. 45 Tumor cell-released glutamate promotes its proliferation in an autocrine-dependent manner. 46 Overexpression and activation of glutamate receptors are highly involved in tumor malignancy. 47 The effectiveness and pharmacological properties of glutamate receptor antagonists are still under evaluation. 2 However, most of these investigations mainly focused on tumor cells themselves, neglecting glutamate-mediated effects on TME. In this study, we proved that glutamate stimulation could inhibit the cell-killing effects of neutrophils, during which STAT3 activation was possibly involved. Glutamate generally exerts biological functions by binding to GluRs or iGluRs. 2 Abnormal activations of GluRs/iGluRs are responsible for tumor development. 48,49 The glutamate mediated pathways, such as GluRs/β-arrestin/STAT3 and iGluRs/Ca 2+ /STAT3, 50,51 are possible mechanisms for explaining neutrophil or other immune cell modulation in TME.
The origins and relationships between neutrophils and MDSCs in TME are still debatable. 52 Recently, it was reported that the overexpression of LOX-1, ARG, and FATP2 are possible molecules to distinguish them. 18,53,54 However, some researchers believe that MDSCs represent polarized immature neutrophils, 55 making F I G U R E 5 Immunofluorescence and western blot (WB) analysis of ADP-ribosylation factor 4 (ARF4)/Ras-related GTP-binding protein 10 (RAB10)/signal transducer and activator of transcription 3 (pSTAT3) on neutrophils. Blood-derived neutrophils were treated with/ without glutamate (100 μM) for 24 h. (A) Triple immunostaining of neutrophils. Green, LY6G; red, specific proteins (ARF4/RAB10/ pSTAT3); blue, nucleus. Scale bar, 100 μm. (B) WB analysis of ARF4/RAB10/pSTAT3 expression on blood-derived neutrophils treated with gradient glutamate (0, 25, and 100 μM) for 24 h. In LLC-tumor bearing mice, neutrophils were collected from blood and tumor tissues. (C) WB analysis of ARF4/RAB10/pSTAT3 expression on neutrophils isolated from blood and tumor tissues. β-Actin was detected as loading control. (D) Quantifications of target proteins in (B) compared to loading control. */**Differences between glutamate and control group. (E) Quantifications of target proteins in (C) compared to loading control. */**Differences between blood and tumor group. n = 3 it possible for cell reprogramming and transformation mediated by surroundings. 54 Others argue that neutrophils are a separate cell lineage of MDSCs, in which activation and recruitment are different. 56 Even though the origins and relationships between them are controversial, the functions and phenotypes of tumor-associated neutrophils are well-defined. Based on the effect of neutrophils toward tumor cells, two neutrophil subtypes are introduced. 18 The N1 subset represents those neutrophils that are recruited to kill tumor cells by H 2 O 2 secretion and myeloperoxidase granule transfer, 12,13 and N2 cells are capable of facilitating tumor growth in a TGFβ or ARG-dependent manner. 18 The tumorigenesis generally transformed N1 into the N2 cell type by reshaping TME. 10,11,57 Here we proved that tumor cell-released glutamate was one of the factors that drove this transformation. The inductions of pSTAT3, ARF4, and RAB10 mediated by glutamate are possible molecules to distinguish N1 and N2 combined with other markers, such as LOX-1, ARG, and FATP2.
Riluzole has been approved by the FDA for amyotrophic lateral sclerosis treatment. 32,58 Interestingly, riluzole also shows obvious inhibitory effects on several types of cancers. 17,33,35 Apart from the direct inhibitory effects on tumor cells, riluzole was deduced to inhibit tumor growth by modulating immune cells in TME as riluzole did not show significantly inhibitory effects in SCID-tumor bearing mice (T and B cell deficiency). 17,34 Here, we confirm that riluzole restores neutrophil cell-killing effects and CD8 + T cell proliferation in TME. This action is mainly attributed to inhibition of glutamate release in TME instead of direct effects on immune cells. Apart from tumor cells, T cells and macrophages are also the source for glutamate in TME., 59,60 It needs to be verified in the future whether these immune cells could influence the expansion and polarization of neutrophils by regulating glutamate secretion and reuptake.
Collectively, our data indicate that glutamate is a crucial tumor-released AA for TME. During this process, induction of STAT3, ARF4, and RAB10 mediated by glutamate embodies the neutrophil immunosuppressive phenotype. Glutamate release inhibitor riluzole dramatically blunts neutrophil transformation by reducing glutamate abundance in TME. Therefore, glutamate is an Targeting glutamate metabolism could improve immunotherapy for cancer patients.