Myeloid activation clears ascites and reveals IL27-dependent regression of metastatic ovarian cancer

Abstract

Patients with metastatic ovarian cancer (OvCa) have a 5-year survival rate of <30% due to the persisting dissemination of chemoresistant cells in the peritoneal fluid and the immunosuppressive microenvironment in the peritoneal cavity. Here, we report that intraperitoneal administration of β-glucan and IFNγ (BI) induced robust tumor regression in clinically relevant models of metastatic OvCa. BI induced tumor regression by controlling fluid tumor burden and activating localized antitumor immunity. β-glucan alone cleared ascites and eliminated fluid tumor cells by inducing intraperitoneal clotting in the fluid and Dectin-1-Syk-dependent NETosis in the omentum. In omentum tumors, BI expanded a novel subset of immunostimulatory IL27+ macrophages and neutralizing IL27 impaired BI efficacy in vivo. Moreover, BI directly induced IL27 secretion in macrophages where single agent treatment did not. Finally, BI extended mouse survival in a chemoresistant model and significantly improved chemotherapy response in a chemo-sensitive model. In summary, we propose a new therapeutic strategy for the treatment of metastatic OvCa.

Figure S1.

KPCA tumors model stage III OvCa and β-glucan treatment clears ascites. (A and B) Treatment timelines of (A) ID8 and (B) KPCA tumors treated with β-glucan. (C) Representative flow plot identifying GFP+CD45⁻ KPCA cells in the peritoneal lavage of mice 1 wk after tumor seeding. (D) Representative image and quantification of compartmental bioluminescent imaging. The omentum is removed from the cavity; signals (red circle) are obtained separately from the rest of the peritoneal cavity (non-omentum signal). (E) Representative fluorescent images of the omentum and KPCA numbers in the omentum of mice 1 wk after KPCA cell injection. Scale bar is 2.5 mm. (F) Representative images of ascites and calculated changes in ascites volumes from PBS- or β-glucan–treated mice. Student’s t test was used. ****P < 0.0001. Error bars are standard errors of the mean.

Figure 1.

β-glucan significantly reduces OvCa fluid tumor burden. (A) Representative bioluminescence images and quantification of bioluminescence signals in PBS- and β-glucan–treated ID8 tumor-bearing mice 42 days after tumor seeding. (B and C) (B) Representative pictures of peritoneal lavage and (C) quantification of GFP⁺ ID8 OvCa cells in the peritoneal lavage of PBS- and β-glucan–treated mice. (D) Tissue of origin and mutation status of Trp53, Kras, Ccne1, and Akt2 in ID8 and KPCA OvCa cell lines. (E) Quantification of KPCA cells in the peritoneal lavage 1 wk after KPCA seeding. (F) Quantification of omentum bioluminescence signals 1 wk after KPCA seeding. (G) Representative bioluminescence images and quantification of bioluminescence signals in PBS- and β-glucan–treated KPCA tumor-bearing mice. (H and I) (H) Representative pictures of peritoneal lavage and (I) quantification of GFP⁺ KPCA OvCa cells in the peritoneal lavage in PBS- and β-glucan–treated mice. Data from three independent runs combined and presented in bar graphs. Each dot represents one mouse. Student’s t test was used *P < 0.05; **P < 0.01; ****P < 0.0001. Error bars are standard errors of the mean.

Figure S2.

β-glucan captures OvCa cells into solid nodular structures via intraperitoneal clotting and Dectin-1-Syk–dependent NETosis in the omentum. (A) Acute cancer cell capture timeline. (B) Quantification of ID8 cells in the peritoneal lavage of mice 5 h after β-glucan treatment. (C) Representative flow plots of GFP+ KPCA cells disappearing from the peritoneal lavage 5 h after intraperitoneal β-glucan administration. (D) Representative in situ images of peritoneal clots formed in the peritoneal cavity after β-glucan treatment. These clots contain GFP+ KPCA cancer cells. Scale bars are 1 mm. (E) Quantification of macrophages and neutrophils in the peritoneal fluid following CLL administration as analyzed by flow cytometry. (F) Quantification of KPCA cells in peritoneal lavage as determined by flow cytometry in control or CLL-pretreated mice 5 h after intraperitoneal β-glucan administration. (G) MFI of TUNEL staining in KPCA cells in the clots β-glucan–treated mice and peritoneal lavage from PBS-treated mice, which do not form clots. (H) Quantification of PRMΦs in the peritoneal lavage of control or OMX mice after β-glucan administration. (I) Quantification by flow of GFP+ KPCA cells in the omentum of Syk^WT and Syk^Mye∆ mice treated with β-glucan. (J) Quantification of KPCA cells in peritoneal lavage in Syk^WT and Syk^Mye∆ mice 5 h after indicated treatment with PBS, β-glucan, or CLL. (K) Graphical representation of two mechanisms of cancer cell capture following intraperitoneal injection of β-glucan. Illustration created with https://biorender.com. Graphs are represented by combined data of three or more independent runs. One-way ANOVA and Student’s t test were used. *P < 0.05; **P < 0.01; ***P < 0.001. Error bars are standard errors of the mean.

Figure 2.

β-glucan captures OvCa cells into solid nodular structures via intraperitoneal clotting and Dectin-1-Syk–dependent NETosis in the omentum. (A and B) Quantification of (A) PRMΦ and (B) KPCA cells in the peritoneal lavage of mice 5 h following PBS or β-glucan treatment. (C and D) (C) Representative image and (D) flow plot of peritoneal clots formed in the peritoneal fluid following β-glucan treatment containing GFP+CD45⁻ KPCA cells. (E) Quantification of PRMΦ and KPCA cells in the peritoneal lavage 5 h after PBS, β-glucan, and β-glucan+heparin treatment. (F and G) Quantification of (F) KPCA and (G) PRMΦ in the peritoneal lavage of Syk^WT and Syk^Mye∆ mice treated with PBS or β-glucan. (H) Representative images of omentum in mice 5 h after PBS and β-glucan treatment. Omentum was stretched over the liver for better imaging. (I) Quantification of KPCA cells in the peritoneal lavage in intact and OMX mice treated as indicated with PBS, β-glucan, and heparin. (J) Representative images of omentum in Syk^WT and Syk^Mye∆ mice 5 h after β-glucan treatment. (K and L) Quantification of KPCA in the peritoneal lavage of (K) Syk^Mye∆ and (L) Dectin-1 KO mice 5 h following treatment as indicated with PBS, β-glucan, and heparin. (M) representative confocal images of omentum of Syk^WT and Syk^Mye∆ mice 5 h after β-glucan treatment. Positive cells were stained blue (S100A9; neutrophil), green (GFP; cancer cell), and white (cHH3; NETs). (N and O) (N) Representative images of omentum, and (O) quantification of KPCA cells in peritoneal lavage in WT and PAD 4KO mice 5 h after indicated treatment. Data from three or more independent runs combined and presented. Each dot represents one mouse. One-way ANOVA and Student’s t test were used. *P < 0.05; **P < 0.01; ****P < 0.0001. Error bars are standard errors of the mean. Relative cell number is reported as a fold change to the average of the control and was used when experimental replicates were combined. Scale bar is 1 mm for panel C, 2 mm for panels H, J, and M, and 50 μM for panel M.

Figure 3.

Combining β-glucan with IFNγ reduces KPCA tumor burden through host immunity. (A) BI timeline in KPCA model of OvCa. (B) Representative bioluminescence images and quantification of bioluminescent signal in mice treated with PBS, IFNγ, β-glucan, or BI. (C and D) Representative bioluminescence images and quantification of bioluminescent signal in (C) IFNγR KO mice and (D) T cell–depleted mice treated with PBS or BI. (E and F) Representative bioluminescence images and (F) quantification of bioluminescent signal of PBS and BI-treated mice on days 8, 10, 14, and 21. (G and H) (G) Omentum and (H) non-omentum body cavity bioluminescent signal in PBS-, IFNγ-, β-glucan–, or BI-treated mice. Data from two or more independent runs combined and presented in bar graphs. Each dot represents one mouse. One-way ANOVA and Student’s t test were used. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Error bars are standard errors of the mean.

Figure S3.

Analyses of tumor burden and toxicity after BI treatments. (A and B) (A) KPCA cell numbers in the omentum evaluated by flow cytometry and (B) omentum tumor weight in PBS- or BI-treated mice. (C) Quantification of KPCA numbers 48 h following PBS and BI treatment in vitro. Student’s t test was used. (D) Treatment and longitudinal imaging timeline in PBS-, IFNγ-, β-glucan–, and BI-treated mice. (E) Quantification of bioluminescence signals in mice tracked longitudinally from day 8 to day 21 after tumor seeding. (F and G) (F) Representative images of ascites accumulation and (G) quantification of KPCA cells and CD45⁺ cells and ascites volumes based on the peritoneal lavage of mice treated with PBS, IFNγ, β-glucan, and BI 21 days after tumor seeding. (H) IDEXX clinical chemistry analyses of sera from PBS- or BI-treated mice. (I) Body weight of PBS- or BI-treated mice 18 days after tumor cell seeding. Graphs are represented by combined data of three or more independent runs. Student’s t test and one-way ANOVA were used *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Error bars are standard errors of the mean.

Figure 4.

BI enriched IL27+ macrophages in omentum tumors. (A) Whole body bioluminescent signal of PBS or CLL-treated mice treated with BI and control mice. (B) A UMAP plot of monocyte/MΦ clusters in omentum tumors. (C) Top upregulated genes in all monocyte/MΦ clusters. (D) Frequencies of eight identified monocyte/MΦ clusters in omentum tumors. (E) Slingshot trajectory analysis from the origin (cluster 5) through three independent pathways (red arrows). (F) qPCR analysis of cluster genes in monocyte-derived MΦs treated with PBS or BI. (G) Top upregulated IPA cytokine regulators in clusters 1 and 2. (H)Il27 and Ebi3 co-expression heatmap of monocyte/MΦ clusters. (I)Il27 expression in monocyte-derived MΦs treated with PBS or BI analyzed by qPCR. (J and K) (J) Uniform manifold approximation and projection (UMAP) plot of monocyte/MΦ clusters and (K) IL27 and EBI3 co-expression in tumors from human OvCa patients. (L) Overall survival analysis in late-stage OvCa patients (stage III and IV) with high and low co-expression of IL27-EBI3. Data from two or more independent runs combined and presented in bar graphs. Each dot represents one mouse. Student’s t test was used. *P < 0.05; **P < 0.01; ***P < 0.001. Error bars are standard errors of the mean.

Figure S4.

Impact BI on MΦs, monocytes and myeloid progenitor cells and scRNA-seq of mouse and patient datasets. (A) Quantification of frequencies of Arginase⁺ MΦs, Tim4⁺ MΦs, and CD64⁺ MΦs in omentum tumors treated as indicated and determined by flow cytometry. (B) Number of progenitor cells and monocytes in the bone marrow of mice 1 wk after PBS or BI treatment. LSK, Lin-Sca1+cKit+ progenitor cells; LT-HSC, long-term hematopoietic stem cells; MPP, multipotent progenitor cells. (C) Expression of Il12a and Il12b in monocyte/MΦ clusters pooled from mice treated with PBS, β-glucan, IFNγ, or BI. (D and E) (D) Uniform manifold approximation and projection (UMAP) plot of immune cells and (E) co-expression of IL27-EBI3 in human OvCa patient tumors. (F) Expression of IL12A and IL12B in each myeloid cell subcluster from human OvCa tumors. Graphs are represented by combined data of three or more independent runs. Student’s t test was used. *P < 0.05. Error bars are standard errors of the mean.

Figure 5.

IL27 contributes to BI treatment by activating T cells and is specifically secreted by BI-stimulated MΦs. (A and B) (A) Bioluminescent signal of omentum tumors and (B) mesentery metastasis scores from mice injected with IgG or αIL27 treated with BI and untreated control mice. (C) MFI of IFNγ and TNF in CD8⁺ T cells in omentum tumors from control or BI-treated mice analyzed by flow cytometry. (D) ELISA quantification of IL27 heterodimer in supernatant from BMDM stimulated with PBS, IFNγ, β-glucan, or BI. (E) ELISA quantification of IL27 heterodimer in supernatant from WT-, Syk^Mye∆, Dectin-1 KO, and IFNγR KO-BMDM cultured with PBS or BI. (F) IFNγ and TNF MFI of CD8 T cells cocultured with MΦ pretreated with PBS or BI in the presence of αIL27 antibody or control IgG. In vivo data combined from three independent runs plotted where each dot represents one mouse. One representative run presented for in vitro data where each dot represents one replicate. Experiment replicated at least three times. Student’s t test and one-way ANOVA were used. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Error bars are standard errors of the mean.

Figure S5.

BI activates T cells in an IL27-dependent manner and kills KPCA cells. (A and B) (A) Quantification of frequencies and (B) activation of CD8⁺ T cells in omentum tumors from PBS- or BI-treated mice and flow cytometry plots of TNF- or IFNγ-stained samples, including fluorescence minus one (FMO) plots used to identify positive populations. (C) Quantification of frequencies and activation of CD4⁺ T cells in omentum tumors from PBS- or BI-treated mice. (D) ELISA quantification of IL30 (IL27p28) in supernatant from BMDM cultured with PBS, IFNγ, β-glucan, and BI. (E–G) Frequencies of (E) IFNγ+ or TNF+, (F) IFNγ+TNF+ CD8⁺ T cells, and (G) Granzyme B MFI of CD8⁺ T cells cocultured with MΦ pretreated with PBS or BI in the presence of αIL27 antibody or control IgG. (H) Representative TUNEL staining in β-glucan– or BI-induced clots. Scale bars are 500 μM. (I) FACS quantification TUNEL MFI in GFP+ KPCA cells. In vivo data combined from three independent runs plotted where each dot represents one mouse. One representative run presented for in vitro data where each dot represents one replicate. Student’s t test and one-way ANOVA were used. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Error bars are standard errors of the mean.

Figure 6.

BI extends overall survival in both chemoresistant and chemo-sensitive models and dramatically enhances chemotherapy response in the chemo-sensitive model. (A) Survival curves of KPCA tumor-bearing mice treated with BI and carboplatin as indicated. The number of mice is PBS n = 14, carboplatin n = 10, BI n = 20, BI + carboplatin n = 15. The graph is a combination of three independent experiments. (B) Survival curves of BPPNM tumor-bearing mice treated with BI and carboplatin as indicated. The number of mice is PBS n = 14, carboplatin n = 20, BI n = 15, BI + carboplatin n = 14. The graph is a combination of three independent experiments. Log-rank test was used.