Tumor-derived exosomes induce N2 polarization of neutrophils to promote gastric cancer cell migration

To investigate the role of gastric cancer cells in neutrophil phenotype and function, we treated neutrophils isolated from human peripheral blood with gastric cancer cell-derived conditioned medium (GC-CM) for 12 hours. Fluorescence-activated cell sorting (FACS) analyses showed that treatment with GC-CM inhibited the spontaneous apoptosis of neutrophils (Fig. 1a). In addition, GC-CM-treated neutrophils presented an increased expression of CD11b, an important molecule for neutrophil chemotaxis (Fig. 1b). Because tumors can modulate immune cells to acquire a pro-inflammatory phenotype, we determined the expression of inflammatory factors including IL-1β, IL-6, IL-8, oncostatin M (OSM), and TNFα in neutrophils. As shown in Additional file 1: Figure S1A, the expression of these inflammatory factors remarkably increased in GC-CM-treated neutrophils compared to controls. In addition, the expression of MMP9 and VEGF was also increased in GC-CM-treated neutrophils (Additional file 1: Figure S1B). GC-CM treatment inhibited ROS production while had minimal effect on the maturation state in neutrophils (Additional file 2: Figure S2A and B). We collected the supernatant from GC-CM-primed neutrophils and used it as chemoattracants for cell migration. The results of transwell migration assay showed that the supernatants from GC-CM-primed neutrophils promoted gastric cancer cell migration (Fig. 1c). Furthermore, GC-CM-primed neutrophils promoted gastric cancer cell proliferation and endothelial cell tube formation (Additional file 2: Figure S2C and D).

The promoting role of neutrophils in cancer progression has been linked to autophagic activation. Thus, we collected GC-CM-treated neutrophils and analyzed for autophagosomes. The results of transmission electron microscope (TEM) analyses showed that there were more autophagosomes in GC-CM-treated neutrophils than control cells (Fig. 1d). Immunofluorescent staining results confirmed the increase in the number of LC3-positive puncta in GC-CM-treated neutrophils compared to control cells (Fig. 1e). Moreover, western blot results showed an increase in LC3-II expression, an autophagosomal marker, in GC-CM-treated neutrophils (Fig. 1f). We then examined the expression of autophagy related 7 (ATG7) and Beclin 1 (BECN1) genes, important regulators of autophagosome formation, for cells treated with GC-CM. As expected, the increased expression of ATG7 and BECN1 was observed in GC-CM-treated neutrophils compared to controls (Fig. 1g). We next treated neutrophils with autophagy inhibitors and evaluated their phenotypic and functional changes in response to GC-CM treatment. As shown in Fig. 1h and i, the pre-treatment with autophagosome formation inhibitor 3-MA, but not autophagosome degradation inhibitor CQ, remarkably reversed the effects of GC-CM on neutrophil survival and CD11b expression. More importantly, the pre-treatment with 3-MA inhibited gastric cancer cell migration induced by GC-CM-primed neutrophils. This effect, however, was not seen for CQ (Fig. 1j). Taken together, these results demonstrate that gastric cancer cell-derived conditioned medium upregulates autophagy-related gene expression and induces autophagy in neutrophils, leading to pro-tumor activation of neutrophils.

To clarify mechanisms for the pro-tumor activation of neutrophils, we treated neutrophils with GC-CM and determined the responses of various pathways. GC-CM treatment increased the expression of phosphorylated p65 (p-p65), p-STAT3, and p-ERK in neutrophils in a time-dependent manner (Fig. 2a). We also found that the expression of p-p38 and p-Akt was modestly increased in GC-CM-treated neutrophils (Additional file 2: Figure S2E). Notably, NF-κB inhibitor reversed GC-CM-induced activation of STAT3 and ERK pathways (Fig. 2b). In consistent with these observations, NF-κB inhibitor completely blocked GC-CM-induced expression of inflammatory factors while only partial inhibition was noted for STAT3 and ERK inhibitors (Additional file 3: Figure S3), suggesting a dominant role of NF-κB pathway. In particular, the pre-treatment of neutrophils with NF-κB inhibitor reversed GC-CM-induced expression of LC3-II, supporting that NF-κB pathway is critically involved in GC-CM-induced autophagy in neutrophils (Fig. 2c). The increased expression of ATG7 and BECN1 in GC-CM-treated neutrophils was also blocked by NF-κB inhibitor (Fig. 2d). Finally, GC-CM-primed neutrophil-induced gastric cancer cell migration was blocked by NF-κB inhibitor (Fig. 2e). These results suggest that gastric cancer cell-derived conditioned medium induces a pro-tumor phenotype in neutrophils through activating NF-κB pathway.

Gastric cancer cells can release exosomes to activate mesenchymal stem cells and macrophages [27, 29]. We next determined whether or not tumor-derived exosomes are involved in the pro-tumor activation of neutrophils. We found that the gastric cancer cell-derived exosomes (GC-Ex) displayed sphere-like morphology with a diameter ~100 nm and expressed the exosomal markers CD9 and CD63 (Fig. 3a and b). To determine the uptake of GC-Ex by neutrophils, GC-Ex was labeled with membrane-bound fluorescent dye CM-Dil and added to neutrophil cultures. Imaging flow cytometry results showed that GC-Ex was efficiently internalized by neutrophils (Fig. 3c). We next investigated the ability of GC-Ex to induce autophagy in neutrophils. The results of TEM and immunofluorescent staining showed that the number of autophagosomes was increased in GC-Ex-treated neutrophils compared to control cells (Fig. 3d and e). GC-Ex treatment also increased LC3-II expression as well as ATG7 and BECN1 expression in neutrophils (Fig. 3f and g). Furthermore, GC-Ex protected neutrophils against spontaneous apoptosis and induced CD11b expression in neutrophils. To demonstrate which step in the process of autophagy GC-Ex may act on, we pre-treated neutrophils with autophagosome formation inhibitor 3-MA or autophagosome degradation inhibitor CQ followed by incubation with GC-Ex. The effects of GC-Ex on neutrophil apoptosis and CD11b expression were blocked by 3-MA. In contrast, no effect was noted for CQ (Fig. 3h and i). Finally, the supernatant from GC-Ex-treated neutrophils promoted gastric cancer cell migration. This increase was blocked by 3-MA but not CQ (Fig. 3j), suggesting that GC-Ex trigger autophagosome formation in neutrophils. Furthermore, GC-Ex treatment inhibited ROS production while had minimal effect on the maturation state in neutrophils (Additional file 4: Figure S4A and B). GC-Ex-primed neutrophils promoted gastric cancer cell proliferation and endothelial cell tube formation (Additional file 4: Figure S4C and D). The expression of MMP9 and VEGF was upregulated in GC-Ex-treated neutrophils (Additional file 4: Figure S4E). These findings suggest that neutrophils can uptake GC-Ex to induce autophagy and pro-tumor activation.

Neutrophils were exposed to GC-Ex and the activation state of various pathways were analyzed. The results of western blot showed that the expression of p-p65, p-STAT3, and p-ERK was increased in GC-Ex-treated neutrophils compared to untreated controls (Fig. 4a). The expression of phosphorylated p-p38 and p-Akt was also modestly increased in neutrophils after GC-Ex treatment (Additional file 4: Figure S4E). The pre-treatment with NF-κB inhibitor blocked GC-Ex-induced activation of STAT3 and ERK in neutrophils (Fig. 4b). In addition, NF-κB inhibitor blocked GC-Ex-induced LC3-II production as well as ATG7 and BECN1 expression in neutrophils (Fig. 4c and d). Similarly, GC-Ex-induced decrease of spontaneous apoptosis and increase of CD11b in neutrophils were reversed by NF-κB inhibitor (Fig. 4e and f). NF-κB inhibitor also blocked GC-Ex-induced the expression of inflammatory factors in neutrophils (Fig. 4g). Moreover, NF-κB inhibitor impaired the migration-promoting effect caused by exposure of gastric cancer cells to the supernatant from GC-Ex-treated neutrophils (Fig. 4h). In summary, these results suggest that GC-Ex induces autophagy and pro-tumor activation of neutrophils through NF-κB pathway.

Exosomes can induce signaling pathway activation in recipient cells by delivering bioactive molecules. To investigate the molecules in GC-Ex that induce neutrophil activation, we treated GC-Ex with proteinase and evaluated the effects of digested exosomes on neutrophil activation. Compared to undigested GC-Ex, the treatment of proteinase-digested GC-Ex decreased LC3-II expression as well as ATG7 and BECN1 expression in neutrophils (Additional file 5: Figure S5A and B). In similar, GC-Ex-induced activation of NF-κB, STAT3, and ERK pathways in neutrophils was also impaired by proteinase treatment (Additional file 5: Figure S5C). Moreover, the treatment with proteinase decreased GC-Ex-induced expression of inflammatory factors in neutrophils compared to undigested GC-Ex (Additional file 5: Figure S5D). Thus, we focused on exosomal proteins in the subsequent studies.

We then tested the roles of exosomes from three other gastric cancer cell lines. Consistent with that observed in exosomes from BGC-823 gastric cancer cells, the exosomes from HGC-27, MGC-803, and SGC-7901 cells also induced autophagy and pro-tumor activation in neutrophils (Additional file 6: Figure S6). To identify protein(s) that mediate the roles of exosomes in neutrophil activation, we performed a proteomic analysis for exosomes from three gastric cancer cell lines (Additional file 7: Table S3). Several proteins of interest were identified such as high mobility group Box-1 (HMGB1), heat-shock protein 70 (HSP70), fibronectin (FN) (Fig. 5a). The results of western blot confirmed HMGB1 expression in GC-Ex from all of the tested gastric cancer cell lines (Fig. 5b). HMGB1 have previously shown to play important roles in inflammation and cancer. Thus, we chose HMGB1 as the target for further study.

We next investigated the role of exosomal HMGB1 in neutrophil activation. Neutrophils were treated with HMGB1 antagonist and then exposed to GC-Ex. HMGB1 antagonist reversed GC-Ex-induced increases in LC3-II expression and ATG7 and BECN1 expression in neutrophils compared to GC-Ex alone (Fig. 5c and d). In addition, GC-Ex-induced apoptotic inhibition as well as increases of CD11b expression in neutrophils were inhibited by HMGB1 antagonist (Fig. 5e). Moreover, HMGB1 antagonist inhibited GC-Ex-induced activation of NF-κB pathway and expression of inflammatory factors in neutrophils (Fig. 5f and g). Finally, the promoting effect of supernatant from GC-Ex-treated neutrophils on gastric cancer cell migration was impaired by HMGB1 antagonist (Fig. 5h).

HMGB1 activates NF-κB pathway by interacting with toll-like receptor (TLR) and receptor for advanced glycation end products (RAGE). We thus examined the effects of TLR2, TLR4, and RAGE inhibitors on the observed roles of GC-Ex in neutrophils. As expected, TLR4 inhibitor blocked GC-Ex-induced autophagy and NF-κB pathway activation in neutrophils (Fig. 5c, d and f). TLR4 inhibitor decreased the expression of inflammatory factors in GC-Ex-treated neutrophils and impaired its gastric cancer cell migration-promoting effect (Fig. 5g and h; Additional file 5: Figure S5E). However, the inhibitors for TLR2 and RAGE showed minimal effects (data not shown). Taken together, these results suggest that GC-Ex delivers HMGB1 to neutrophils and induces pro-tumor activation through TLR4/NF-κB signaling.

In silico gene expression analysis (Dataset GSE13911 retrieved from Gene Expression Omnibus) showed remarkably increased HMGB1 expression in gastric cancer tissues compared to the paired non-cancerous tissues (Fig. 6a). Therefore, we measured HMGB1 in paired gastric cancer tissues and non-cancerous tissues by ELISA. The expression of HMGB1 was increased in gastric cancer tissues compared to non-cancerous tissues (Fig. 6b). To determine whether increased HMGB1 expression is associated with clinical outcomes, HMGB1 was analyzed in the cancer and non-cancerous tissues from gastric cancer patients by using tissue microarray. We found that a higher level of HMGB1 in gastric cancer patients predicted poorer prognosis (Fig. 6c). In addition, we found that HMGB1 expression levels were higher in exosomes from the culture supernatants of gastric cancer tissues than non-cancerous tissues and in exosomes from the serum samples of gastric cancer patients than healthy controls (Additional file 8: Figure S7).

To confirm our findings from gastric cancer cell lines, we tested neutrophil activation by using exosomes from matched gastric cancer and non-cancerous tissues. Similar to GC-Ex, exosomes from gastric cancer tissues prevented the apoptosis of neutrophils and induced CD11b expression in neutrophils (Fig. 6d and e). Gastric cancer tissue-derived exosomes also induced higher ATG7 and BECN1 expression (Fig. 6f) and NF-κB activation (Fig. 6g) in neutrophils than non-cancerous tissue-derived exosomes. Neutrophils treated with exosomes from gastric cancer tissues promoted gastric cancer cell migration more efficiently (Fig. 6h). In addition, HMGB1 antagonist reversed gastric cancer tissue-derived exosomes-induced increases of ATG7 and BECN1 expression in neutrophils (Fig. 6i). Moreover, gastric cancer tissue-derived exosomes-induced expression of inflammatory factors in neutrophils was inhibited by HMGB1 antagonist (Fig. 6j). Finally, the promoting effect of supernatant from gastric cancer tissue-derived exosomes-treated neutrophils on gastric cancer cell migration was impaired by HMGB1 antagonist (Fig. 6k). Taken together, these findings indicate that gastric cancer tissue-derived exosomes could induce autophagy and pro-tumor activation of neutrophils.

To clarify the importance of HMGB1 in the pro-tumor activation of neutrophils, we silenced HMGB1 in gastric cancer cells by gene-specific siRNA (Fig. 7a) and tested GC-Ex-induced neutrophil activation. Compared to exosomes from cells transfected with scramble control siRNA, exosomes from HMGB1-silenced cells showed decreased protection against spontaneous apoptosis and CD11b expression in neutrophils (Fig. 7b and c). In addition, HMGB1 knockdown decreased GC-Ex-induced LC3-II expression as well as ATG7 and BECN1 expression in neutrophils (Fig. 7d and e). Similarly, GC-Ex from HMGB1-silenced cells had decreased ability to activate NF-κB pathway in neutrophils compared to control exosomes (Fig. 7f). Furthermore, inflammatory factor expression decreased for neutrophils that were treated with exosomes from HMGB1-silenced cancer cells compared to exosomes from cells transfected with scramble control siRNA (Fig. 7g). Finally, HMGB1 knockdown blocked the promoting effect of GC-Ex-activated neutrophils on gastric cancer cell migration (Fig. 7h). On the contrary, recombinant HMGB1 inhibited spontaneous apoptosis and increased CD11b expression in neutrophils (Additional file 9: Figure S8A and B). Recombinant HMGB1-treated neutrophils promoted the migration of gastric cancer cells and increased the expression of ATG7 and BECN1 genes in neutrophils (Additional file 9: Figure S8C and D). The expression of MMP9 and VEGF and pro-inflammatory factors was also upregulated in recombinant HMGB1-treated neutrophils (Additional file 9: Figure S8E). These results support HMGB1 as a key mediator for GC-Ex-induced neutrophil activation.

Fresh tissue specimens were obtained from patients with GC who underwent surgical resection at the Affiliated People’s Hospital of Jiangsu University between April 2016 and December 2016. None of these patients had received chemotherapy or radiotherapy before surgery. Patients with infectious diseases, autoimmune disease or multi-primary cancers were excluded. The study was approved by the ethics committee of Jiangsu University and informed consent was obtained from all patients.

Human gastric cancer cell lines BGC-823, MGC80-3, SGC-7901, and HGC-27 were purchased from the Institute of Biochemistry and Cell Biology at the Chinese Academy of Sciences (Shanghai, China). Cells were cultured in low-glucose Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA, USA) at 37 ^oC in humidified air with 5% CO₂. When cells reached 80% confluence, medium was replaced with serum-free medium. Following a 24-hour incubation, medium was collected, centrifuged to remove cell debris, and stored at -80 ^oC in aliquots as conditioned medium (CM).

Peripheral blood was collected from healthy volunteers after obtaining written informed consent. The study was approved by the ethics committee of Jiangsu University. Neutrophils were isolated by using Polymorphprep (Axis-Shield PoC AS, Norway) as previously described [21]. RBCs were lysed using hypotonic lysing procedure. The purity of neutrophils was 98% after this procedure. Neutrophils were seeded in RPMI 1640 (Invitrogen) supplemented with 10% FBS and 1% penicillin/streptomycin at a density of 1×10⁶ per well and treated with or without CM or exosomes from gastric cancer cells for 12 h. For experiments that use inhibitors, cells were pretreated with inhibitors for 1 h prior to adding CM or exosomes. The inhibitors used in this study include autophagy initiation inhibitor 3-methyladenine (3-MA, 5 mM), autophagosome degradation inhibitor chloroquine (CQ, 20 μM), HMGB1 inhibitor Glycyrrhizin (Gly, 10 μM), TLR4 antagonist TAK-242 (TAK, 10 μM), ERK inhibitor U0126 (10 μM), NF-κB inhibitor Bay11-7082 (10 μM), and STAT3 inhibitor WP1066 (10 μM). For HMGB1 treatment, cells were treated with recombinant human HMGB1 (10 μg/ml; Biovision, Shanghai, China) for 12 hours.

Exosomes were isolated from the conditioned medium of gastric cancer cells as previously described [29]. In brief, cells were cultured in exosome-depleted medium and the supernatants were collected after 48 h. The conditioned medium was centrifuged at 1,000 g for 10 min to remove cell debris followed by 30 min at 10,000 g using 100 KDa MWCO before the concentrated solutions were filtrated through a 0.22-μm pore filter (Millipore, Shanghai, China). Exosomes were precipitated by adding the exosome quick extraction solution (System Biosciences, Palo Alto, CA, USA) at a ratio of 1:5 at 4 ^oC for 12 h. Exosomes were dissolved with PBS and stored at -80 ^oC. Protein concentration was determined by BCA protein assay kit (ThermoFisher Scientific, Shanghai, China). The size and concentration of exosomes were measured by Nanoparticle Tracking Analysis (NTA). The morphology of purified exosomes was identified by transmission electron microscopy (Tecnai 12; Philips) and the expression of exosomal markers CD9 and CD81 by western blot.

Tissue array was purchased from Shines Pharmaceuticals (Shanghai, China). A total of 76 pairs of tumor tissues and non-tumor tissues were included in the tissue array. Tissue array was incubated with antibody against HMGB1 (Cell Signaling Technology). Immunohistochemical staining was performed as described elsewhere. IHC scoring was assessed by two pathologists in a double-blinded manner.

The supernatants from gastric cancer tissues and adjacent non-cancerous tissues were collected for ELISA. The concentrations of HMGB1 in tissue supernatants were determined by using ELISA kit according to the manufacturer’s instructions (Chondrex, Redmond, WA, USA).

Neutrophils treated with or without CM or exosome from gastric cancer cells for 12 h were collected and resuspended in serum-free medium. Cells were stained with DCFH-DA (10 μM; Beyotime Biotechnology, Shanghai, China) at 37 ^oC for 30 min and subjected to analyses of florescence intensity by flow cytometry.

Total RNA was extracted from cells using Trizol reagent (Thermo Fisher Scientific) and 1 μg of RNA was reverse transcribed to cDNA by using reverse transcriptase (Vazyme). Real-time quantitative PCR was performed by using the SYBR Green I real-time detection kit (Cwbio, Beijing, China) on a Bio-Rad CFX96 Detection System. The relative gene expression was normalized to β-actin. The primers for target genes were listed in Additional file 10: Table S1.

The siRNA against HMGB1 was produced by Genepharrm (Suzhou, Jiangsu, China). The sequences of HMGB1 siRNA and the scramble control were shown in Additional file 11: Table S2. BGC-823 cells (1×10⁵ cells/well) were grown in 6-well plates and transfected with siRNAs by using LipoFiter transfection reagent (Hanbio, Shanghai, China) for 36 hours.

Cells were lysed in RIPA buffer containing proteinase inhibitors. Equal amount of proteins was separated by a 12% SDS-PAGE gel. Following electrophoresis, proteins were transferred to a PVDF membrane, blocked in 5% non-fat milk, and incubated with primary antibodies at 4 ^oC overnight. Antibodies for ERK1/2, p-ERK1/2, NF-кB p65, p-p65, STAT3, p-STAT3, p-p38, p38, p-Akt, Akt, LC3, CD9, CD63, Alix, and TSG101 were purchased from Cell Signaling Technology (Louis Park, MN, USA). After washing with TBST for three times, membrane was incubated with HRP-conjugated goat anti-rabbit or anti-mouse secondary antibodies (Bioworld Technology) at room temperature for two hours. The protein bands were visualized by enhanced chemiluminescence. GAPDH served as the loading control.

Neutrophil apoptosis was analyzed by using an Annexin V apoptosis detection kit according to the manufacturer’s instructions (Invitrogen). The binding of Annexin V-FITC and PI to the cells was analyzed on FACS Calibur (BD Biosciences, NJ, USA) by using Cell Quest software.

For transmission electron microscopic analysis, neutrophils treated with or without CM or exosome from gastric cancer cells were washed and fixed in 4% glutaraldehyde, followed by post-fixation in 2% osmium tetroxide. Thereafter, cells were dehydrated, treated with propylene oxide, and embedded. The sections were subsequently stained with uranyl acetate and lead citrate and examined in a Tecnai 12 transmission electron microscopy. For immunofluorescent staining, neutrophils treated with or without CM or exosomes were stained with an autophagy detection kit (Enzo Lifesciences, NY, USA) and analyzed by Cytasion 3 cell imaging multi-mode reader (BioTek, Shanghai, China).

Cell migration assay was performed in a 24-well Boyden chamber with an 8-μm pore size polycarbonate membrane (Corning, Union City, CA, USA). Cancer cells (2×10⁴ in 100 μl of serum-free medium) were added into the upper chamber with 600 μl of supernatant from neutrophils in the lower chamber. After a 12-h incubation, cancer cells on the upper surface of the membrane were removed. The migrated cells on the lower surface of the membrane were fixed by paraformaldehyde, stained with crystal violet, and counted under a microscope.

Neutrophils were treated with or without CM or exosome from gastric cancer cells for 12 h, followed by washing with PBS once and culturing in fresh medium for additional 12 h. The supernatant from activated neutrophils were collected and filtered through a 0.22 μm filter. Human umbilical vein endothelial cells (HUVEC) were seeded at 2×10⁴ cells/well and incubated with or without conditioned media from neutrophils at 37 ^oC for 12 h. The formation of tube-like structure by HUVECs were observed under a phase-contrast microscope and photographed at 100× magnification. The number of tubules from five random fields in each well was counted. The experiments were repeated for three times.

Gastric cancer cells were seeded at 4×10³ cells/well and incubated with or without supernatant from the activated neutrophils. CCK8 reagent (10 μl; Vazyme, Nanjing, China) was added at 3 hours before the end of the experiments. The absorption was measured at 450nm in a microplate reader. The experiments were repeated for three times.

Fresh surgically removed gastric cancer tissues and adjacent non-cancerous tissues (at least 5 cm distant from the tumor site) were cut into 1 cm³ and placed in 1 ml serum-free RPMI 1640 medium for 24 h. The supernatants were centrifuged at 300 g for 10 min and filtered through a 0.22 μm filter and stored at -80 ^oC until use. Neutrophils were cultured in 50% T-CM or N-CM for 12 h. Tumor-derived exosomes were extracted from the conditioned medium as described in previous section.