Introduction

Gastric cancer (GC) is the fifth most commonly diagnosed cancer and the fourth leading cause of cancer-related deaths worldwide1. Peritoneal metastasis (PM) is a common form of recurrence and the survival of patients with PM remains poor2,3. Immune checkpoint inhibitors (ICIs) have recently attracted considerable attention as novel treatment options for several malignant tumors. In GC, several phase III trials have demonstrated the efficacy of ICIs in prolonging survival, although some patients respond poorly to ICI therapy4,5. However, a multicenter retrospective study revealed that patients with PM had a poor prognosis when treated with nivolumab6, and real-world data from a multicenter prospective study showed that PM was a key poor prognostic factor in nivolumab-treated GC patients7.

Cold tumors, characterized by an immunosuppressive tumor microenvironment (TME), pose a major obstacle to immunotherapy8. Immunosuppressive cells within the TME are of interest for modifying the immune environment of cold tumors and overcoming therapeutic resistance to ICIs. We have previously shown, using clinical GC specimens, that PM is a form of cold tumor with reduced infiltration of CD8+ T cells and increased infiltration of M2 macrophages compared to primary tumors, and that this immune evasion may be responsible for PM resistance to ICI therapy9. However, the precise mechanism of how the polarization of M2 macrophages is induced and CD8+ T cells are suppressed in PM of GC patients has remained unclear.

Recently, a more refined characterization of tumor immunity identified that tumor-infiltrating B cells play a crucial, synergistic role in the modulation of the immune response in the TME10. The existence of a subset of B cells that suppress inflammation has recently been demonstrated, known as regulatory B cells (Bregs), which are characterized by the production of anti-inflammatory cytokines such as IL-1011,12,13. Bregs contribute to tumor progression because of their ability to induce immune evasion with an increased number and activity of other immunosuppressive cells. IL-10-producing B cell subsets are more abundant in the peritoneal cavity (7–8%) than in the spleen (1–2%), blood (0.6%), or lymph nodes (0.2%)14. Peritoneal Bregs have been reported to regulate immune homeostasis within the gut tissues by modulating CD4+ T cell function and neutrophil infiltration in induced colitis models15. Although considerable work has been done to elucidate the role of Bregs in tumor immunity, little effort has been made to examine its potential function in PM of GC.

Recent studies have shown that the phosphatidylinositol 3-kinase (PI3K)-Akt pathway was shown to be an important downstream effector of B cell antigen receptor signaling and to promote IL-10 production in B cells16. Since the phosphatase and tensin homolog (PTEN) is an inhibitor of Akt activity, knockdown of PTEN enhances IL-10 production16. Matsushita et al. confirmed that Bregs were expanded by inactivated PTEN specifically in B cells using a Cre-loxP system. Conditional deletion of PTEN in B cells was achieved by crossing mice expressing loxP-flanked Pten alleles with Cd19Cre mice. Bregs were significantly increased in the peripheral blood, spleen, and peritoneum of B cell-specific PTEN-deficient mice (Cd19Cre+/−PtenloxP/loxP mice)16. Thus, B cells specific PTEN deficient mice are a good model for investigating Bregs function in tumor immunity17.

We hypothesized that Bregs, via IL-10 production, suppress CD8+ T cells and promote polarization into M2 macrophages, thereby inducing immune evasion, leading to the progression of PM. The present study investigated the distribution of Bregs in patients with GC, and analyzed the interaction between Bregs and immune cells in the peritoneal TME using B cell-specific PTEN-deficient mice PM models.

Materials and methods

Patients

We performed immunostaining of paired primary lesions and peritoneal metastases from 60 patients with gastric cancer treated at Kanazawa University Hospital from 2000 to 2021, for whom clinicopathological data were available. All eligible cases during this period that met the criteria were included. Among them, we also evaluated lymph node metastases in 36 patients who underwent concomitant resection of the PM and primary tumors. Additionally, liver and ovarian metastases were evaluated in 10 patients. Blood and ascites samples were collected from 10 patients in each group of early GC, advanced GC, PM, and healthy subjects, which were analyzed using flow cytometry. Abdominal lavage fluid was used for patients with early and advanced GC.

Written informed consent was obtained from all patients prior to the study. This study was approved by the Institutional Review Board of Kanazawa University Graduate School of Medical Science (study permission number: 114264-1). All methods were applied in accordance with relevant guidelines and regulations.

Cells and materials

YTN16 is a GC cell line transplanted into C57BL/6 mice. YTN16 cells were established from subcutaneous tumors by injecting primary cultured cells derived from mouse gastric adenocarcinomas. Gastric tumors were established in p53 heterozygous knockout C56BL/6 mice by adding N-methyl-N-nitrosourea (MNU) to drinking water18. The acquired tumor cells were cultured in high-glucose Dulbecco’s modified Eagle medium (DMEM, Sigma–Aldrich Japan, Tokyo, Japan) containing 1.0 mlL/L MITO (Coning Japan, Tokyo), 10 mL/L L-Glutamine, 10 mL/L Penicillin/Streptomycin, and 10% fetal bovine serum (FBS), on plastic dishes coated with type I collagen solution (0.5% Atelocollagen Acidic Solution IPC-50; Koken Co., Ltd., Japan).

Chemicals

Wortmannin was obtained from FUJIFILM (Japan), diluted to the required concentrations in RPMI1640 medium for the in vitro studies, and dissolved in dimethyl sulfoxide (DMSO) for the in vivo studies.

Immunohistochemistry

Tumor specimens were fixed in 10% neutral buffered formalin and embedded in paraffin. For double immunohistochemical staining of CD19 and IL-10, sections were incubated with primary antibodies (CD19 antibody [NBP2-52716, mouse, diluted 1:300; Novus Biologicals] and IL-10 antibody [ab34843, rabbit, diluted 1:200; abcam]) at 4℃ overnight. The sections were then incubated with ImmPRESS® Duet Reagent (HRP Horse Anti-Mouse IgG/AP Horse Anti-Rabbit IgG) for 10 min After washing in phosphate-buffered saline (PBS), immunoreactivity was visualized using the ImmPACT DAB EqV Substrate, which was visible as a brown color, and the ImmPACT Vector Red Substrate, which was visible as a red color, and counterstained with hematoxylin. In addition, sections were stained with hematoxylin and eosin, followed by immunostaining for CD4 (NCL-L-CD4-368, mouse, diluted 1:100; Leica), CD8 (NCL-L-CD8-4B11, mouse, diluted 1:50; Leica), CD68 (ab955, mouse, diluted 1:3000; Abcam) and CD163 (ab182422, rabbit, diluted 1:200; Abcam). CD19+IL-10+ cells were defined as Bregs and CD19+ cells as B cells. Each cell was counted, and the percentage of Bregs among B cells was analyzed at 200 × magnification. All sections were examined using a fluorescence microscope (Olympus, Tokyo, Japan). Positively stained cells in five non-overlapping intratumoral fields were observed at × 100 and 200 magnifications, counted manually, and interpreted independently by two researchers.

Mouse peritoneal metastatic and subcutaneous tumor model

The animal use proposal and experimental protocol (AP-224333) were reviewed and approved by the Animal Care and Use Committee of the Kanazawa University. All animal experiments were performed according to the standard guidelines of Kanazawa University. C57BL/6J mice (20 g, 6–8 weeks old) were purchased from Charles River Laboratories, Inc. (Yokohama, Japan). B cell-specific Pten-deficient mice were obtained using the Cre-loxP system from Prof. Matsushita of Kanazawa University, Department of Dermatology and Molecular Genetics. All the procedure of the study is followed by the ARRIVE guideline.

For the mouse allograft PM model using wild-type C57BL/6J mice, Pten control mice (Cd19Cre+/− mice), and B cell-specific Pten-deficient mice (Cd19Cre+/−PtenloxP/loxP mice), YTN16 cells were cultured in high-glucose DMEM. Cells (5 × 106) were injected intraperitoneally, and 2 × 106 cells were subcutaneously injected. After 12 days, the mice were sacrificed and the body weight, ascites volume, and total weight of the peritoneal tumors were examined. We collected ascites and resected the spleen, peritoneal, and skin tumors. The specimens were analyzed using flow cytometry.

To investigate the effects of inhibiting the PI3K pathway, we treated mice intraperitoneally with wortmannin (PI3K inhibitor) at 0.5 [mg/kg]/day or an equal volume PBS daily following injection of YTN16 cells. On day 12 after inoculation, the mice were sacrificed, and ascites, peritoneal tumors, and spleens were removed for weight calculation and flow cytometry.

Isolation of immune cells from tissue samples in mice

Ascites were collected using a syringe and centrifuged. ACK lysis buffer (Sigma–Aldrich Japan, Tokyo, Japan) was added and diluted with buffer. Abdominal lavage fluid was used for control mice. The spleen was cut and centrifuged, and ACK lysis buffer was added and diluted with buffer. Peritoneal and skin tumors were cut into small pieces and collagenase II and PBS were added. They were shaken in a water bath at 37℃ for 30 min. After adding ACK lysis buffer, the resulting cell suspensions were filtered through a 100 μm mesh filter.

Flow cytometry

Analyses were performed using FlowJo v 9.6.1 software (FlowJo, Ashland, OR, USA). In mice, B cells (CD45+CD19+), Bregs (CD45+CD19+CD1d+CD5), CD4+ T cells (CD3e+CD4+), CD8+ T cells (CD3e+CD8+), and M2 macrophages (CD11b+CD163+) were stained in the spleen, ascites, PM, and skin tumors with the following antibodies: anti-mouse CD19 (eBioscience, clone 1D3), anti-mouse CD45 (eBioscience, clone 30-F11), anti-mouse CD5 (Biolegend, clone 53 − 7.3), anti-mouse CD1d (BD, clone 1B1), anti-mouse CD3e (Invitrogen, clone 500A2), anti-mouse CD4 (Invitrogen, clone GK1.5), anti-mouse CD8 (Invitrogen, clone 53 − 6.7), anti-mouse CD11b (Biolegend, clone M1/70), and anti-mouse CD163 (Biolegend, clone S15049F). In humans, B cells (CD19+) and Bregs (CD19+CD24+CD27+) from the blood and ascites were stained with the following antibodies: anti-human CD19 (BioLegend, clone HIB19), anti-human CD24 (BioLegend, clone ML5), anti-human CD27 (BioLegend, clone M-T271), and anti-human CD38 (BioLegend, clone HB-7). Proliferating cells were analyzed using a CellTrace CFSE Cell Proliferation Kit (Thermo Fisher Scientific).

MTT assay

The effect of wortmannin on the proliferative capacities of YTN16 cells was quantified using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. YTN16 cells were seeded in 96-well plates at 4 × 103 cells per well. Following incubation, supernatant was discarded and replaced with fresh serum-free medium and various concentrations of wortmannin (0.1, 0.5, 1, 5, 10µM) were added. At 48 h post-treatment, supernatant was discarded, MTT solution and DMSO (Wako, Japan) were added. The absorbance of the solution in each well was measured at 535 nm using a microplate reader (model 550; Bio-Rad, Japan). Cell viability was calculated as viability = (absorbance of experimental wells)/(absorbance of control wells).

ELISA

The mouse Il-10 DuoSet ELISA (R&D Systems) was used to measure Il-10 levels in the culture media of Bregs and other B cells, according to the manufacturer’s instructions. Bregs and other B cells were cultured with or without wortmannin in each 96 well plates, and after 72 h, IL-10 levels were quantified at 450 and 570 nm using a microplate reader (Multiskan GO, Thermo Scientific, USA).

Cell co-culture

Bregs and CD8+ T cells were isolated from splenocytes by magnetic-activated cell sorting (MACS) technology with Regulatory B cell Isolation Kit and CD8+ T cell Isolation Kit (Miltenyi Biotech, Auburn, CA, USA).

Bone marrow-derived macrophages (BMDMs) were differentiated from bone marrow cells after incubation with macrophage colony stimulating factor (40 ng/mL) for 5 days in RPMI medium containing 10% FBS .

For differentiation into M2 macrophages, BMDMs seeded at 2 × 105 cells per well in 96 well plates were incubated alone or with Bregs or other B cells at a ratio of 1:1 for 72 h. Lipopolysaccharide (LPS, 10 µg/mL), phorbol 12-myristate 13-acetate (PMA, 50 ng/mL), and ionomycin, (500 ng/mL) were also added into each plate with or without wortmannin (0.5 µM).

For T cell differentiation, CFSE-labeled CD8+ T cells were co-cultured with Bregs or other B cells at a ratio of 1:1 in RPMI containing 10% FBS on 96 well plates (2 × 105cells/ well). in presence of anti-CD3 (3 µg/mL), anti-CD28 (5 µg/mL), LPS, PMA, and ionomycin for 5 days. Each cell line was analyzed by flow cytometry, as described above.

Statistical analysis

Statistical analyses were conducted using GraphPad Prism (software version 10.4.2, https://www.graphpad.com/). Analyses were performed using the Student’s t-test or repeated-measures ANOVA. p < 0.05 indicated statistical significance. The overall survival (OS) was calculated based on the date of peritoneal dissemination or recurrence. The Cox hazard model was used for univariate and multivariate analyses, and statistical significance was set at p < 0.05.

Results

Increased frequency of Bregs in the peripheral blood and Ascites of patients with GC and PM

We analyzed Breg frequency using flow cytometry in peripheral blood and ascites samples from healthy control subjects and patients with early GC, advanced GC, and PM (n = 10 for each group). The Breg frequency is represented by the ratio of the number of CD19+CD24hiCD27+ B cells to the number of CD19+ B cells. Abdominal lavage fluid without ascites was used for patients with early and advanced GC without PM. The frequency of Bregs in the peripheral blood of PM patients was higher than that of healthy control subjects and early GC patients, and was higher than in advanced GC patients, although there was no significant difference (control vs. PM; 10.1 ± 1.6% vs. 19.4 ± 3.9%, p = 0.004, early GC vs. peritoneal metastasis; 11.2 ± 2.0% vs. 19.4 ± 3.9%, p = 0.0143) (Fig. 1A). The frequency of Bregs in ascites of PM patients was also higher than that of early GC and advanced GC patients (early GC vs. PM; 2.5 ± 2.4% vs. 17.6 ± 9.0%, p = 0.02, advanced GC vs. peritoneal metastasis; 5.5 ± 2.3% vs. 17.6 ± 9.0%, p = 0.02) (Fig. 1B).

Fig. 1
figure 1

Regulatory B cell (Breg) frequency in the peripheral blood and ascites of patients with gastric cancer (GC). (A) Bregs among B cells in the peripheral blood samples from healthy control subjects, patients with early GC, advanced GC, and peritoneal metastases (PM) (n = 10 for each group).The Bregs ratio in the peripheral blood of patients with PM was higher than other groups. (B) Bregs among B cells in ascites from patients with early GC, advanced GC, and PM (n = 10 for each group). The Bregs ratio in ascites of patients with PM was higher than that of patients with early and advanced GC (*p < 0.05, **p < 0.01). The results are provided in terms of means ± standard deviation values.

Full size image

Breg frequency is elevated in PM tissue and is associated with CD8+ T cell and M2 macrophage infiltration

In double immunostaining, CD19+IL-10+ cells were defined as Bregs, and CD19+ cells as B cells. The frequency of Bregs, represented by the ratio of the number of double-positive cells to the number of CD19+ cells, was evaluated in the primary tumor, PM, and lymph node metastases in paired patients with PM. Breg frequency increased in PM samples compared with their primary tumor (p < 0.001). The frequency of Bregs in PM tissues was also higher than that in the lymph nodes, liver, and ovarian metastatic tissues (p < 0.001) (Fig. 2A, B).

Fig. 2
figure 2

Association between regulatory B cells (Bregs) and immune cell infiltration in peritoneal metastases (PM) of patients with gastric cancer (GC). (A) Representative microscopic images of Bregs expression in primary tumors (n = 60), and peritoneal (n = 60), lymph mode (n = 36), liver (n = 10), and ovary (n = 10) metastases. B cells were stained brown and IL-10 was stained pink. Each cell was counted and the percentage of Bregs (yellow triangle) among the B cells were analyzed under 200 × magnification. (B) The Breg frequency in PM was higher than in primary tumors and other metastases (***p < 0.001) (C) Representative microscopic images of CD8, CD163 positive cells in PM in GC (magnification ×100). The number of CD8, CD163 positive cells was measured and shown as the average count in five non-overlapping intratumoral fields. (***p < 0.001). The results are provided in terms of means ± standard deviation values.

Full size image

In addition, we evaluated the association between Breg frequency and CD8+ T cell and CD163+ cell (M2 macrophage) infiltration in PM (Fig. 2C). The median value determined the cut-off point for the Breg high /low frequency groups in PM, which was set at 17.2%. The high-frequency group had 29 patients and the low-frequency group had 31 patients. On the other hand, the CD8+ T cell count in the Breg high frequency group was lower than that in the Breg low frequency group (p < 0.001). In contrast, the M2 macrophage count in the Breg high frequency group was higher than that in the Breg low frequency group (p < 0.001). However, there was no significant difference in the correlation between Bregs frequency and CD4+ T cell and CD68+ cell (M1 macrophage) infiltration. As Bregs in the cancer microenvironment may affect patient prognosis, we investigated the relationship between Breg frequency and clinicopathological factors in patients with PM.

Bregs infiltration is an independent prognostic factor in patients with PM of GC

We analyzed the relationships between Breg frequency in PM and clinicopathological factors in patients with PM, and found no differences. There was a tendency for the Breg frequency of peritoneal tumors to correlate with the Breg frequency of the primary tumor, but the difference was not significant (Table 1). The median value for the Breg high /low frequency groups in primary tumor was 10.2% and was set as the cut-off point. The median survival time (MST) of patients in the Breg high frequency group was 10.8 months, and their OS was significantly lower than that of patients in the Breg low frequency group, whose MST was 25.2 months. (p = 0.005) (Fig. 3). The univariate analysis revealed differences in age, performance status (PS), history of chemotherapy, carcinoembryonic antigen (CEA) level, and Breg frequency. In the multivariate analysis, Breg frequency and chemotherapy were identified as independent prognostic factors (hazard ratio (HR): 2.314, 95% confidence interval (CI): 1.298–4.124, p = 0.004; HR: 0.158, 95% CI: 0.034–0.738, p = 0.02, respectively), as shown in Table 2.

Table 1 Relationship between Breg frequency and clinicopathological factors in patients with peritoneal metastasis.
Full size table
Fig. 3
figure 3

Kaplan–Meier survival curve for patients with peritoneal metastases according to regulatory B cell (Breg) frequency (p = 0.005).

Full size image
Table 2 Univariate and multivariate analysis for prognostic factors in patients with peritoneal metastasis.
Full size table

Bregs are frequently localized in Ascites and peritoneal tumors in an immunocompetent mouse model of PM

We evaluated Breg infiltration in the ascites, spleen, subcutaneous tumors, and peritoneal tumors in an immunocompetent mouse model of PM using flow cytometry. The total number of cells in the abdominal lavage of control mice was significantly lower than that in the ascites of the PM model. The number of Bregs was significantly higher in the PM model than in the control mice (p < 0.001), although there was no difference in the Bregs ratio (Fig. 4A). We then simultaneously created peritoneal and subcutaneous tumors in mice and compared Breg frequency, represented by the ratio of the number of CD19+CD1dhiCD5+ cells Bthan that in subcutaneous tumors (6.3 ± 2.5% vs. 1.4 ± 1.3%, p = 0.003) (Fig. 4B).

Fig. 4
figure 4

Analysis of the localization of regulatory B cells (Bregs) in a mouse peritoneal metastasis (PM) model. (A) Flow cytometry analysis of Bregs from the spleen and ascites in the control mouse and PM mouse model (n = 6 for each group). (B) Flow cytometry analysis of Breg frequency in peritoneal tumors and subcutaneous tumors (n = 6). (**p < 0.01, ***p < 0.001). The results are provided in terms of means ± standard deviation values.

Full size image

Analysis of Bregs and immune cells in a B cell-specific Pten-deficient mouse model of PM

We compared a PM model using B cell-specific Pten-deficient mice (Cd19Cre+/− PtenloxP/loxP mice) with a PM model using control mice (Cd19Cre+/− mice) and found an increased ascites volume and peritoneal tumor weight in B cell-specific Pten-deficient mice (Fig. 5A). The Breg frequency in the ascites and peritoneal tumor in the B cell-specific Pten-deficient mice were significantly higher than that in control mice (p = 0.001) (Fig. 5B). CD8+ T cell infiltration in ascites and peritoneal tumors was decreased in B cell-specific Pten-deficient mice (p = 0.005 and p = 0.03, respectively), while M2 macrophage infiltration increased compared to that in control mice (p = 0.005 and p = 0.003, respectively) (Fig. 5C).

Fig. 5
figure 5

Analysis of regulatory B cells (Bregs) and immune cells in a B cell-specific Pten-deficient mouse model of peritoneal metastasis (PM). (A) Comparison of ascites volume and peritoneal tumor weight between Pten-deficient mice (n = 6) and control mice (n = 6) (**p < 0.01, ***p < 0.001). (B) Flow cytometry analysis of Breg frequency from the spleen, ascites, and peritoneal tumors in Pten-deficient mice and control mice (**p < 0.01, ***p < 0.001). Each group consisted of 6 mice. (C) Comparison of CD8+ T cells and M2 macrophages from the spleen, ascites, and peritoneal tumors between Pten-deficient mice and control mice (*p < 0.05, **p < 0.01). Each group consisted of 6 mice. The results are provided in terms of means ± standard deviation values.

Full size image

Bregs promote M2 macrophage polarization and inhibit CD8+ T cell proliferation through the production of IL-10

The PI3K inhibitor, wortmannin attenuated the proliferation of YTN16 cells in a dose-dependent manner, with no significant effect on growth at a dose of 0.5µM as shown in Fig. 6A. Based on these results, 0.5µM wortmannin was used in subsequent in vitro experiments.

Fig. 6
figure 6

Effect of regulatory B cells (Bregs) on CD8+ T cell proliferation and M2 macrophages polarization in vitro. (A) Antiproliferative effects of wartmannin in YTN16 cells. Cell viability was assessed by MTT assay after 48-h exposure to a single-dose of wartmannin (0–10 µM) (*p < 0.05, **p < 0.01). (B) IL-10 concentration in the supernatant of Bregs and other B cell subsets excluding Bregs cultured with or without 0.5µM wortmannin measured by ELISA (**p < 0.01, ***p < 0.001). (C) Flow cytometry analysis of M2 macrophage polarization by co-culturing bone marrow-derived macrophages (BMDMs) and Bregs with or without 0.5µM wortmannin (**p < 0.01). (D) The proliferation of CD8+ T cells cultured with Bregs and other B cell subsets with or without 0.5µM wortmannin, assessed using the CFSE assay (*p < 0.05). The results are provided in terms of means ± standard deviation values.

Full size image

We evaluated the concentration of IL-10 in the supernatants of Bregs and other B cell subsets, excluding Bregs cultured with or without wortmannin. Bregs produced significantly more IL-10 than other B cell subsets, which were reduced by the addition of wortmannin (Fig. 6B). We examined the effect of Bregs on M2 macrophage polarization by co-culturing BMDMs collected from femurs with Bregs for 72 h in the presence of LPS, PMA, and ionomycin, followed by analysis with flow cytometry. We found that the Bregs co-culture group showed increased M2 macrophage differentiation compared to the other group, which was suppressed by 0.5µM wartmannin. (Fig. 6C). CD8+ T cells labeled with CFSE were cultured with Bregs and other B cells, and CD8+ T cell proliferation was analyzed using flow cytometry. CD8+ T cells co-cultured with Bregs showed a decrease in the suppression of proliferation, while wartmannin attenuated the effects of Bregs on CD8+ T cells. (Fig. 6D).

The PI3K inhibitor Wortmannin modifies tumor immunity via regulation of bregs, suppressing peritoneal tumor growth

We used an immunocompetent mouse model of PM to analyze the effect of wortmannin, on Bregs and other immune cell infiltrates in the spleen, ascites, peritoneal tumors, and on peritoneal tumor growth. Wortmannin was administered intraperitoneally at the same dose (0.5 [mg/kg]/day, daily); no body weight loss was observed compared with the controls without wortmannin (n = 6 for each group). The wortmannin-treated group showed a significant reduction in ascites volume and peritoneal tumor weight compared with the control group (Fig. 7A). In the wortmannin-treated group, Breg frequency in ascites and peritoneal tumors was significantly lower than that in control mice (2.8 ± 1.5% vs. 8.5 ± 2.5%, p = 0.002, 3.5 ± 1.3% vs. 6.5 ± 2.8%, p = 0.03, respectively) (Fig. 7B). Wortmannin treatment significantly increased CD8+ T cell infiltration in the spleen, ascites, and peritoneal tumors (p = 0.02, p = 0.04, and p = 0.002, respectively) and significantly decreased M2 macrophage infiltration, which was particularly pronounced in peritoneal tumors (p < 0.001). (Fig. 7C).

Fig. 7
figure 7

Effect of the PI3K inhibitor wortmannin on regulatory B cells (Bregs), tumor-infiltrating immune cells, and peritoneal tumor growth. (A) Comparison of ascites volume and peritoneal tumor weight between the control and wortmannin-treated model (**p < 0.01, ***p < 0.001). Each group consisted of 6 mice. (B) Flow cytometry analysis of Bregs from the spleen, ascites, and peritoneal tumors in the control and wortmannin-treated model (*p < 0.05, **p < 0.01). (C) Infiltration of CD8+ T cells and M2 macrophages in the spleen, ascites, and peritoneal tumors in control and wortmannin-treated models (*p < 0.05, **p < 0.01, ***p < 0.001). The results are provided in terms of means ± standard deviation values.

Full size image

Discussion

The interaction between Bregs and immune cells in the TME promotes tumor progression by blocking antitumor responses and helping tumor cells evade immune surveillance10. The human peritoneal cavity has markedly distinct immunological features compared to systemic immunity19,20,21,22. Therefore, it is important to elucidate the distribution of Bregs and their role in the peritoneal immune environment from a therapeutic perspective. Our study is the first to show that Bregs are more frequently localized in cancerous ascites and peritoneal tumors in GC patients, as well as in mouse models of PM, compared to primary tumors and other metastatic forms. Furthermore, our results revealed that Breg infiltration in PM, but not in primary tumors, is an independent prognostic factor in patients with GC with PM. Takahashi et al. recently reported that the percentage of CD19+ B cells in the ascites fluid was significantly increased in GC patients with PM and that this percentage was positively correlated with the peritoneal carcinomatosis index (PCI) score23. They speculated that B cells in the peritoneal cavity mostly shared functionality with Bregs, and our results seem to support their speculation that Bregs mobilized from the peritoneal cavity are involved in PM progression.

The present study also demonstrates that Bregs are involved in immune evasion in PM possibly through the production of IL-10. IL-10 is a representative anti-inflammatory cytokine that suppresses immune cells including T cells, macrophages, and dendritic cells10. In the present study, B cell-specific Pten-deficient mice showed a significant increase in Breg frequency via activation of the PI3k-Akt pathway in B cells. In conjunction with the activation of Bregs, a decrease in CD8+ T cells and an increase in M2 macrophages within the ascites and peritoneal tumors were observed, further promoting ascites accumulation and peritoneal tumor growth in B-cell-specific Pten-deficient mice. In melanoma, Bregs have been reported to suppress tumor immunity by inhibiting the production of IFN-γ and TNF-α by CD8+ T cells17. Similarly, our in vitro data indicated that Bregs reduced the proliferative capacity of CD8+ T cells in the co-culture. In contrast, the differentiation of monocytes to M2 macrophages was promoted by co-culture with Bregs. These changes were not observed in other B cells, except for Bregs, suggesting that they are associated with regression-specific cell–cell interactions. Taken together, these results suggest that Bregs induce an immune evasion of PM through suppressing CD8+ T cells and enhancing M2 macrophage by IL-10 production.

Recent evidence has revealed that Bregs mediate immunosuppression through several mechanisms beyond IL-10 production. B cell–derived IL-35 activates STAT3 in CD8⁺ T cells, suppressing the expression of chemokine receptors such as CXCR3 and CCR5 and the effector cytokine IFN-γ, thereby impairing CD8⁺ T cell tumor infiltration and cytotoxic function24,25. In parallel, Breg-derived TGF-β induces Foxp3 expression in CD4⁺ T cells, promoting their conversion into regulatory T cells and contributing to T cell anergy within the tumor microenvironment26,27. Moreover, PD-L1 expressed on Bregs can directly engage PD-1 on T cells, leading to T cell exhaustion and further dampening antitumor immunity26. These mechanisms are particularly relevant in the hypoxic and cytokine-rich peritoneal microenvironment, which may enhance Breg plasticity and suppressive function28,29. Future studies should investigate these alternative mechanisms in the context of PM.

We have confirmed that the PI3K inhibitor wortmannin decreased IL-10 production from Bregs and inhibits CD8+ T cell inactivation and macrophage M2 polarization by Bregs co-culture at low concentrations that do not inhibit YTN16 cell proliferation in vitro. In addition, we demonstrated that wortmannin suppressed the frequency of Bregs in PM mice models. Along with the suppression of Bregs, increased infiltration of CD8+ T cells and decreased infiltration of M2 macrophages and suppression of tumor growth were observed. These findings indicate that inhibition of the PI3K pathway in Bregs is effective in activating tumor immunity and reducing tumor growth of PM. However, since the PI3K-AKT pathway is critical in various cell types, including cancer cells, it remains to be elucidated whether the tumor shrinkage is due to suppression of Bregs.

While this study used B cell–specific PTEN-deficient mice to evaluate Breg expansion via the PI3K/Akt pathway, this model does not allow selective depletion of Bregs. Moreover, we did not conduct functional validation using IL-10 neutralizing antibodies or B cell-specific IL-10 knockout models, which limits our ability to draw causal conclusions regarding the immunosuppressive role of Bregs. Additionally, although wortmannin was effective in modulating tumor immunity, its broad inhibitory activity affects multiple cell types and limits specificity. In future studies, the use of B cell–targeted approaches, such as Il10flox/floxCD19-Cre mice30, B cell-specific PI3K inhibitors, or monoclonal antibodies against Breg-associated markers, will be essential for clarifying the precise role of Bregs in PM.

One of the most important limitation in research on Bregs is the lack of unique surface markers, which complicates the definition and classification of Bregs. Several reports have been published on this subset of Bregs10. Because the present study focused on IL-10-producing Bregs, which are potentially distributed in the peritoneal cavity, we investigated their phenotype, the CD19+CD1dhiCD5+ subset, as Bregs in mice14. Human studies have reported two Breg subsets: CD19+CD24hi CD27+ and CD19+CD24hi CD38hi31. CD19+CD24hi CD27+ Bregs were found at higher percentages in the primary tumor and peripheral blood of patients with GC than in those of healthy individuals32. In accordance with these findings, in the present study, we investigated the CD19+CD24hi CD27+ Breg subset in humans. Clarifying the dynamics of Bregs with different phenotypes in the PM of GC is an urgent issue for further research on the tumor immunity of B cells. In addition, several limitations of this study should be noted. First, all clinical specimens were obtained from a single institution, which may limit the generalizability of the findings. Second, external validation using independent cohorts was not performed. Third, although wortmannin modulated Bregs and immune cell infiltration, its non-specific inhibition of the PI3K pathway across multiple cell types prevents definitive attribution of effects to Bregs. Future studies involving multicenter cohorts and more selective Breg-targeted strategies are needed to validate and extend these findings toward clinical application.

In conclusion, we showed that Bregs are frequently localized in peritoneal tumors and are significantly correlated with the prognosis of patients with GC and PM. Bregs induce local immunosuppression and thereby may be involved in the progression of PM. The close correlation between Bregs and patient prognosis suggests that Bregs are attractive targets for PM treatment. Breg-targeted therapy is expected to modify tumor immunity and provide a breakthrough in immunotherapy for patients with GC and PM.