High-grade serous ovarian cancer development and anti-PD-1 resistance is driven by IRE1α activity in neutrophils

Abstract

High-grade serious ovarian cancer (HGSOC) is an aggressive malignancy that remains refractory to current immunotherapies. While advanced stage disease has been extensively studied, the cellular and molecular mechanisms that promote early immune escape in HGSOC remain largely unexplored. Here, we report that primary HGSO tumors program neutrophils to inhibit T cell anti-tumor function by activating the endoplasmic reticulum (ER) stress sensor IRE1α. We found that intratumoral neutrophils exhibited overactivation of ER stress response markers compared with their counterparts at non-tumor sites. Selective deletion of IRE1α in neutrophils delayed primary ovarian tumor growth and extended the survival of mice with HGSOC by enabling early T cell-mediated tumor control. Notably, loss of IRE1α in neutrophils sensitized tumor-bearing mice to PD-1 blockade, inducing HGSOC regression and long-term survival in ~ 50% of the treated hosts. Hence, neutrophil-intrinsic IRE1α facilitates early adaptive immune escape in HGSOC and targeting this ER stress sensor might be used to unleash endogenous and immunotherapy-elicited immunity that controls metastatic disease.

Keywords: ER stress; IRE1; PD-1 blockade; immunotherapy; neutrophils; ovarian cancer.

Figure 1.

Neutrophils facilitate tumor progression in an autochthonous model of HGSOC. (a) Schematic of the experimental system to generate autochthonous HGSOC. Plasmid vectors enabling myc overexpression and CRISPR-Cas9-mediated targeting of Trp53 are co-delivered into the ovary by in vivo electroporation. (b-h) correlation plot between the mass of the primary ovarian tumor of C57BL/6J mice and infiltrating immune cell populations based on FACS analysis. (b) Neutrophils (CD45⁺CD11b⁺Ly6G^HiLy6C^Lo), (c) macrophages (CD45⁺CD11b⁺F4/80⁺), (d) monocytes (CD45⁺CD11b⁺Ly6G⁻Ly6C⁺), (e) dendritic cells (CD45⁺CD11c⁺MHC-II⁺), (f) T cells (CD45⁺CD3⁺), (g) B cells (CD45⁺CD19⁺), (h) NK cells (CD45⁺NK1.1⁺). Correlation plots between mass of the primary tumor and the proportional ratio of (i) T cell to neutrophils and (j) T cells to macrophages. Human HGSOC patient survival analysis based on bulk RNA-seq TCGA data of the indicated genes. Patients are assessed based on relative ratio of (k) intratumoral T cells (CD3G) to neutrophils (MPO) and (l) intratumoral T cells (CD3G) to macrophages (CD163). Patients with low ratio of gene expression are compared to those with a high ratio score. (m) In vivo antibody-mediated depletion scheme. Ten days after autochthonous HGSOC induction, C57BL/6J mice were treated as indicated in the figure. (n) Flow cytometry quantification of neutrophils in peripheral blood. (o) Overall survival of ovarian tumor-bearing C57BL/6J mice that were i.p. injected with neutrophil-depleting antibody (anti-Ly6G) or an isotype control (IgG2a). (b-j) simple linear regression, Pearson correlation coefficient (r²) and P-value provided (n = 15 mice). (k-l) cox proportional hazards model, results were adjusted for false discovery rate (FDR). (n) Unpaired t-test (n = 6). (o) Log-rank (mantel cox) test (isotype n = 13, anti-Ly6G n = 14) exact P-values are provided.

Figure 2.

Enhanced IRE1α-XBP1s activation in intratumoral neutrophils. C57BL/6J mice were induced with autochthonous HGSOC as described in figure 1a. (a-d) CD45⁺CD11c⁻CD19⁻F4/80⁻ CD11b⁺Ly6G^HiLy6C^Lo TANs were FACS sorted from tumor-bearing mice and expression of the indicated genes was analyzed by rt-qPCR. (e-g) correlation of ER stress response gene expression in neutrophils from tumor-bearing mice. (a-d) one-way ANOVA (n = 7 mice). (e-g) pearson correlation coefficient (r²) and P-value (n = 7 mice) provided.

Figure 3.

Loss of IRE1α in neutrophils blunts primary HGSOC progression. (a-f) analysis of Ern1^f/f and Ern1^f/f Mrp8^Cre mice bearing autochthonous HGSOC. (a) RT-qPCR analysis of neutrophils sorted from the bone marrow and primary ovarian tumor of mice with autochthonous HGSOC. (b) Violin plots showing the proportion of neutrophils in the primary ovarian tumor, ascites, and omental metastases. (c) Correlation between primary ovarian tumor mass and neutrophil infiltration by flow cytometry. (d-f) tumor bearing mice were palpated bi-weekly from day 21 post-surgery to determine (d) median time to develop ovarian tumors, (e) Kaplan-Meier survival curve of overall survival rates, and (f) median time of the progression of HGSOC tumor growth. (g-h) Ern1^f/f or Ern1^f/f Mrp8^Cre mice implanted i.P. With the transplantable MP cell line. (g) Metastatic progression of tumor-bearing mice. left: Representative BLI of tumor-bearing mice at day 24 post-implantation. right: quantification of peritoneal tumor burden over time. (h) Kaplan-meier survival curve of mp-bearing mice. (a-b) 2-way ANOVA (n = 4-6 mice/group). (c) Simple linear regression, Pearson correlation coefficient (r²) and P-value provided (Ern1^f/f n = 21, Ern1^f/f Mrp8^Cre n = 27). (d, f) unpaired t-test (n = 17/group). (g) 2-way ANOVA (Ern1^f/f n = 10, Ern1^f/f Mrp8^Cre n = 10). (e, h) log-rank (mantel cox) test. Group numbers and P-values provided.

Figure 4.

Neutrophil-intrinsic IRE1α curtails T cell-dependent control of primary HGSOC: (a) flow cytometry-based analysis of T cell populations found in peripheral blood 48 hours after i.P. administration of T cell-depleting antibodies (anti-CD4/CD8) or an isotype (IgG2b) control. (b) Kaplan-Meier survival curves for the indicated treatment groups. (c-e) indicated tissue samples were harvested from autochthonous hgsoc-bearing mice (c) total splenocytes or bone marrow cells were seeded in a 96-well plate and challenged with the indicated lysates. left: Representative image of ifn-γ ELISpot assay and right: quantification of results. (d-e) CD4 and CD8 T cells (CD45⁺CD11b⁻CD11c⁻CD19⁻CD3⁺) analyzed by flow cytometry. (d) left Representative FACS plots and corresponding violin plots for the proportion of center: CD44^Hi CD62L^lo effector CD4⁺ T cells and right: effector CD8⁺ T cells. (e) left representative FACS plots and corresponding violin plots for the proportion of center: CD44+ Ki67+ CD4⁺ and right: CD8⁺ T cells. (a) 2-way ANOVA (Ern1^f/f n = 4, Ern1^f/f Mrp8^Cre n = 4). (b) Log-rank (mantel cox) test, P-value and number of mice per group provided. (c) 2-way ANOVA (n = 7/group). (d) 2-way ANOVA (n = 12/group). (e) 2-way ANOVA (Ern1^f/f n = 12, Ern1^f/f Mrp8^Cre n = 14). P-values provided.

Figure 5.

Ablation of IRE1α in neutrophils sensitizes HGSOC hosts to PD-1 blockade immunotherapy: female mice were induced with autochthonous HGSOC as described in figure 1A and the indicated tissue samples were harvested from tumor bearing mice. (a) left: Representative FACS plots of total CD4 T cells (CD45⁺CD11b⁻CD11c⁻CD19⁻CD3⁺CD8⁻CD4⁺) and right: the corresponding violin plots quantifying the proportion of CD44⁺ PD-1⁺ CD4 T cells. (b) left: Representative FACS plots of CD8 T cells (CD45⁺CD11b⁻CD11c⁻CD19⁻CD3⁺ CD8⁺CD4⁻) and right: the corresponding violin plots quantifying the proportion of CD44⁺ PD-1⁺ CD8 T cells. (c) left: Representative FACS plots of CD45⁺CD11b⁺ cells. center: histograms showing PD-L1 expression in CD11b⁺ cells. right: quantification of gMFI of PD-L1⁺ CD11b⁺ cells. (d) Kaplan Meier survival curve for indicated treatment group. i.P. injected with anti-PD-1 or an isotype control (IgG2a). (a-b) 2-way ANOVA (Ern1^f/f n = 14, Ern1^f/f Mrp8^Cre n = 14). (c) 2-way ANOVA (n = 6/group). (d) Log-rank (mantel cox) test, P-value and group numbers provided.