Intratumoral Fusobacterium nucleatum Recruits Tumor-Associated Neutrophils to Promote Gastric Cancer Progression and Immune Evasion

Abstract

Intratumoral microbiota can affect the development and progression of many types of cancer, including gastric cancer. A better understanding of the precise mechanisms by which microbiota support gastric cancer could lead to improved therapeutic approaches. In this study, we investigated the effect of intratumoral microbiota on the tumor immune microenvironment during gastric cancer malignant progression. Analysis of human gastric cancer tissues with 16S rRNA amplicon sequencing revealed that Fusobacterium nucleatum was significantly enriched in gastric cancer tissues with lymph node metastasis and correlated with a poor prognosis. F. nucleatum infection spontaneously induced chronic gastritis and promoted gastric mucosa dysplasia in mice. Furthermore, gastric cancer cells infected with F. nucleatum showed accelerated growth in immunocompetent mice compared with immunodeficient mice. Single-cell RNA sequencing uncovered that F. nucleatum recruited tumor-associated neutrophils (TAN) to reshape the tumor immune microenvironment. Mechanistically, F. nucleatum invaded gastric cancer cells and activated IL17/NF-κB/RelB signaling, inducing TAN recruitment. F. nucleatum also stimulated TAN differentiation into the protumoral subtype and subsequent promotion of PD-L1 expression, further facilitating gastric cancer immune evasion while also enhancing the efficacy of anti-PD-L1 antibody therapy. Together, these data uncover mechanisms by which F. nucleatum affects gastric cancer immune evasion and immunotherapy efficacy, providing insights for developing effective treatment strategies. Significance: Intratumoral F. nucleatum activates NF-κB signaling to facilitate gastric cancer immune evasion by promoting tumor-associated neutrophil recruitment that sensitizes tumors to immune checkpoint blockade therapy.

Graphical abstract

Figure 1.

F. nucleatum is significantly enriched in patients with gastric cancer with lymph node metastasis. A, LDA score of bacterial abundance between gastric cancer tissues (T) and normal (N) tissues at the genus level. The bacteria with values of log₁₀ (LDA score) >3 are shown (n = 40). B, The LDA plot was used to compare the bacterial abundance in the primary lesions of patients with gastric cancer with or without LN metastasis. Tn, tumor tissues without lymph node metastasis; Tp, tumor tissues with lymph node metastasis. C, The LDA plot was used to compare the bacterial abundance in the perigastric LNs of patients with gastric cancer with or without LN metastasis. D, qRT-PCR analysis indicated the abundance of F. nucleatum in gastric cancer tissues compared with normal tissues (n = 14). FFPE, formalin-fixed, paraffin-embedded. E,In situ hybridization demonstrates the localization and expression levels of F. nucleatum in gastric cancer (GC) tissues. F, FISH analysis demonstrated an enrichment of F. nucleatum within the primary lesion of gastric cancer. G, FISH results indicated the detection of F. nucleatum within the LNs in patients with metastatic gastric cancer. H, Left, representative CT scans from a patient with gastric cancer and postoperative recurrence (high F. nucleatum, metastasis). Right, survival curve displayed the PFS rates of patients with gastric cancer at different expression levels of F. nucleatum (n = 40). I, The OS of patients with lower or higher F. nucleatum abundance in gastric cancer (n = 93). Log-rank test was used. Data are presented as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Figure 2.

F. nucleatum promotes the malignant progression of gastric cancer. A, C57BL/6 male mice were orally gavaged with H. pylori SS1 (n = 27 for 12 months), F. nucleatum (n = 9 for 6 months), PBS (n = 9 for 12 months; n = 9 for 6 months), or E. coli (n = 9 for 6 months). i.g, intragastric gavage. B,In situ image of the stomach of mice infected with F. nucleatum, H. pylori, or E. coli.C, Representative H&E staining images of the stomach of mice demonstrated higher atypical hyperplasia in the gastric mucosa of the mice in the F. nucleatum-positive (F.n⁺) group. D and E, Representative macroscopic images and H&E staining images of peripheral gastric lymph nodes in the primary gastric cancer model. H&E results showed the status of LN infiltration. F, The survival curves illustrate the survival times and proportions of mice in different groups. G, In this model, the PBS (n = 5), E. coli (n = 5), and F. nucleatum (n = 5) were cocultured with gastric cancer cells, and the cell suspension was then injected subcutaneously into BALB/C mice. H and I, The macroscopic images of tumor tissues and growth curves displaying the growth status of subcutaneously xenografted tumor models in BALB/c nude mice following F. nucleatum infection. J, Construction diagram for the subcutaneous tumor model in 615 mice. K and L, The macroscopic images of tumor tissues and growth curves displaying the growth status of subcutaneously xenografted tumor models in 615 mice following F. nucleatum infection. M, The model construction pattern of the mouse tail vein–lung metastasis model and popliteal LN metastasis model. IF, immunofluorescence. N, Representative in vivo imaging system staining, macroscopic images, and H&E staining of lung tissues from mice in the tail vein–lung metastasis model. O, The in vivo imaging systems revealed the fluorescence of popliteal LNs within the mouse popliteal LN metastasis model. P, The popliteal LN metastasis anatomy and LN tissue and the H&E staining of popliteal LNs in mice. Q, Representative quantification of LN-negative and LN-positive tissues in PBS, E. coli, and F. nucleatum groups. R, The LN volume contrast histogram. S, The survival curve of mice with implanted footpad lymph node metastasis model. Data are presented as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant. A, G, J, and M, Created with BioRender.com. Zhang, T. (2025), https://BioRender.com/r65m917.

Figure 3.

F. nucleatum invades gastric cancer cells and establishes intracellular colonization. A, Representative SEM images of MKN1 cells after coculture with F. nucleatum. B, Representative 3D fluorescence confocal images of F. nucleatum infection in gastric cancer cells. C, Representative confocal images of cystic protrusions and folds on the surface of gastric cancer cells infected by F. nucleatum. D, Representative TEM images of gastric cancer cells after coculture with F. nucleatum. The higher magnification images demonstrate the invasion of gastric cancer cells by F. nucleatum. E, Representative TEM images show that gastric cancer (GC) cell membranes extended tentacles to wrap F. nucleatum. F, Representative living cell fluorescence images of F. nucleatum being endocytosed by gastric cancer cells. G, CFSE staining was employed to demonstrate the impact of F. nucleatum infection on the viability of gastric cancer cells. H, Antibiotic experiments verified the intracellular survival status of F. nucleatum. I, The application of endocytosis inhibitors led to a noteworthy reduction in the quantity of F. nucleatum that invaded gastric cancer cells. J, The macroscopic image of the subcutaneous tumor xenograft model with endocytosis inhibition. K, The specific values of tumor tissue weights for different groups of mice in a histogram. L, Growth curves displaying tumor volumes. M, Dissect the popliteal LN tissue of the mouse foot pad LN metastasis model for culture to confirm the presence of viable F. nucleatum in the LNs. Data are presented as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant. CHI, chlorpromazine; GEN, genistein; E-C-G, EIPA-chlorpromazine-genistein.

Figure 4.

F. nucleatum–infected gastric cancer cells induce TAN recruitment and differentiation. A, Overview of experiment design in scRNA-seq. B, The main cell types and clusters annotated by known cell lineages in PBS and F. nucleatum (F.n) samples are illustrated by Uniform Manifold Approximation and Projection (UMAP) plots. C, Bar plot showing a comparison of the proportions of the main cell types between PBS 1, PBS 2, F.n1, and F.n2. Among them, 1 and 2 represent different periods. D, Flow cytometry results showing the expression level of the neutrophil marker protein CD11b and Ly6G in subcutaneous tumors of mice. E, MIF showing the expression level of Ly6G in subcutaneous tumors of mice. F, Circle diagram demonstrating the interaction between different cell types in different groups. G, UMAP plot showing the distributions of these neutrophil clusters in different groups. H, Bubble plot illustrating the mean expression levels of immune checkpoint inhibition–related genes across distinct TAN clusters. I, UMAP plot showing the expression of CD274 (PD-L1) in different samples. J, The mean fluorescence intensity (MFI) analysis of flow cytometry revealed the expression level of PD-L1 within TANs in mice. K, The single-cell trajectory plots by Monocle demonstrating the reconstructed developmental path of tumor cells from neutrophils. L, Heatmap showing the expression of genes in a branched-dependent manner. Each row represents the dynamic expression of a gene. M and N, Bar chart displaying the results of GO and KEGG enrichment analysis on genes of corresponding clusters in J. Data are presented as mean ± SEM. **, P < 0.01; ***, P < 0.001; ns, not significant. BP, biological progress; SMC, smooth muscle cell. A, Created with BioRender.com. Zhang, T. (2025), https://BioRender.com/d33a784.

Figure 5.

F. nucleatum–mediated malignant progression of gastric cancer is significantly associated with neutrophil recruitment. A, Overview of neutrophils transwell experiment design. B, The transwell experiment and live cell fluorescence staining demonstrating the induction of dHL60 and aNB4 cell migration by the supernatant after F. nucleatum infects gastric cancer cells (the human acute promyelocyte HL60 and NB4 were induced to neutrophils). C, Pattern diagram showing the construction of the neutrophil depletion model. In this model, neutrophils were depleted using the anti-Ly6G antibody via intraperitoneal injection. Each group of mice consists of five individuals. D, General anatomy of subcutaneous tumors in different groups of 615 mice with Ly6G clearance. E, Growth curves displaying tumor volumes. F, The pattern diagram depicting the construction of the neutrophil depletion model in 615 mice. Each group of mice consists of six individuals. SC, subcutaneous injection. G, Anatomical illustration of the mouse footpad and popliteal LNs. H, The data pertaining to LN volume in the popliteal fossa demonstrating a notable reduction in the metastatic potential mediated by F. nucleatum following the depletion of neutrophils in mice. I and J, H&E staining revealing the distribution of tumor cells within different LN groups. K and L, MIF staining showing the expression level of CD8, Ly6G, CD49b, NK1.1, and CD3 in tumor tissues during neutrophil depletion following F. nucleatum infection. Data are presented as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant. Ln (−), LNs without metastasis; Ln (+), LNs with metastasis. A, C, and F, Created with BioRender.com. Zhang, T. (2025), https://BioRender.com/n44z466.

Figure 6.

F. nucleatum activates the IL17 signaling pathway within gastric cancer cells, mediating TAN recruitment and differentiation. A, The heatmap displaying the relative activation and significance of TAN-related immune pathways in different clusters. Orange, relative activation; blue, relative inactivity. The size of the bubble represents the relative level of activity or inactivity, with larger bubbles indicating a higher degree of alertness. B, The volcano map showing differential genes between the PBS group and the live F. nucleatum–infected group. F.n(L), live F. nucleatum. C and D, The visualization of a bubble plot representing the outcomes of KEGG enrichment analysis conducted on the genes associated with gastric cancer cells infected with live F. n(L) and F. n(D). F.n(D), heat-inactivated F. nucleatum. E, Representative IHC images of IL17A expression in human gastric cancer tissues with high and low abundance of F. nucleatum. F, Representative IHC images of IL17A expression in gastric mucosa of mice infected by F. nucleatum. The specimen used in this figure is derived from mouse gastric mucosal tissue, as shown in Fig. 2C. G, IHC staining showing the expression level of IL17A protein in gastric cancer tissues of mice after F. nucleatum infection. H, The ELISA analysis revealed the secretion level of IL17 by gastric cancer cells following infection with F. nucleatum. I, The Western blot results demonstrating the expression of IL17 protein in gastric cancer cells subsequent to infection with F. nucleatum and across varying MOI values. J, Western blot analysis indicates the expression levels of IL17A protein in gastric cancer cells following F. nucleatum infection at different time intervals. K, The pattern diagram showing the construction of the IL17 clearance model in mice. Each group of mice consists of five individuals. L, General anatomy of subcutaneous tumors in different groups of 615 mice with IL17 clearance. M, Growth curves displaying tumor volumes. N, Immunofluorescence results showing the expression of IL17A in tumor tissues with IL17A clearance. O, IHC staining demonstrating significantly weakened expression of Ki67 and Ly6G when IL17A was cleared in the tumor tissue with F. nucleatum infection. P, The pattern diagram demonstrating the construction of the IL17 and Ly6G clearance model. Each group of mice consists of five individuals. Q, General anatomy of subcutaneous tumors in different groups of 615 mice with IL17 and Ly6G clearance. R, Growth curves displaying tumor volumes. Data are presented as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. K and P, Created with BioRender.com. Zhang, T. (2025), https://BioRender.com/b17e915.

Figure 7.

F. nucleatum activates the IL17 signaling pathway through RelB–IL17A interaction. A, The gene set enrichment analysis revealed a significant upregulation of the NF-κB signaling pathway in association with live F. nucleatum infection. B, qRT-PCR results showing RNA expression of differentially expressed genes in the IL17 signaling pathways after F. nucleatum infection with gastric cancer cells. C and D, Western blot results showing the levels of NF-κB pathway-related proteins increased after F. nucleatum infection with gastric cancer cells, among which, p52 and RelB were more prominent. E, Western blot was used to compare the changes in the levels of p52 and p50 proteins. F, Western blot results demonstrating the expression level of RelB in gastric cancer cells infected with live F. nucleatum (denoted as “L”) or upon deactivation of F. nucleatum (denoted as “D”). G, The qRT-PCR results showing the expression level of RelB within gastric cancer cells infected with F. nucleatum. H, IHC staining revealed the expression level of RelB in human gastric cancer tissue enriched with F. nucleatum. I, IHC staining results displaying the expression of RelB in gastric cancer tissues of mice after F. nucleatum infection. J, Immunofluorescence. showing the status of RelB–nuclear translocation in gastric cancer cells after F. nucleatum infection. K, Western blot results depicting the status of RelB–nuclear translocation in gastric cancer cells after F. nucleatum infection. L, Overview of the neutrophil transwell experiment design. GC, gastric cancer. M, The transwell experiment results demonstrating the chemotactic ability of dHL60 and aNB4 cells after knocking down RelB in gastric cancer cells. NC,  negative control. N, ELISA results displaying the level of IL17A secretion after knocking down gastric cancer cells and neutrophil RelB genes. O, ELISA results showed the secretion level of IL17A in gastric cancer cells caused by F. nucleatum infection after RelB knockdown. P, Immunofluorescence. results showing the expression and positioning of IL17A, Ly6G, and RelB in tumor tissues. Data are presented as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. L, Created with BioRender.com. Zhang, T. (2025), https://BioRender.com/g61x838.

Figure 8.

F. nucleatum mediates TAN recruitment to promote gastric cancer immune evasion and boost PD-L1 blockade efficacy. A, CellPhoneDB analysis of scRNA-seq displaying the interaction relationship of proteins between designated receptor cell taxa. B, Bar plot showing a comparison of the proportions of the main cell types in different samples. C, A comparison of the proportions of the main cell types between PBS1, PBS2, F.n1, and F.n2 was displayed by the bar plot. D, Flow cytometry analysis revealed the expression levels of PD-1 on CD8⁺ T cells in tumor tissues of the mouse neutrophil depletion model. E, The pattern diagram depicting the construction of the PD-L1 clearance model in mice. Each group of mice consists of five individuals. F, General anatomy of subcutaneous tumors in different groups of 615 mice with PD-L1 clearance. G, Growth curves display tumor volumes. H, The pattern diagram depicting the construction of the IL17A and Ly6G clearance model in mice. Each group of mice consists of five individuals. SC, subcutaneous injection. I, General anatomy of subcutaneous tumors in different groups of 615 mice with IL17A and Ly6G clearance. J, Growth curves display tumor volumes. K and L, MIF staining revealed the proportion of CD8⁺ T cells and Ly6G⁺ neutrophils in F. nucleatum–infected tumor tissue with PD-L1 clearance. M and N, MIF staining revealed the proportion of CD66b⁺ neutrophils in gastric cancer tissue with high F. nucleatum density. Np, paranormal tissue of gastric cancer with lymph node metastasis; Tn, tumor tissues without lymph node metastasis; Tp, tumor tissues with lymph node metastasis. O,In situ hybridization revealed the abundance of F. nucleatum in gastric cancer tissues of patients responding to anti–PD-L1 therapy. FFPE, formalin-fixed, paraffin-embedded. P, The qRT-PCR analysis indicated the abundance of F. nucleatum in the primary tumors of patients with gastric cancer (GC) exhibiting a response to PD-L1 therapy. Q, Mechanism diagram of the whole article. Data are presented as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. E, H, and Q, Created with BioRender.com. Zhang, T. (2025), https://BioRender.com/o74e010 (E and H) and https://BioRender.com/h16l685 (Q).