The dynamic immune behavior of primary and metastatic ovarian carcinoma

Abstract

Patients with high-grade serous ovarian carcinoma (HGSC) are usually diagnosed with advanced-stage disease, and the tumors often have immunosuppressive characteristics. Together, these factors are important for disease progression, drug resistance, and mortality. In this study, we used a combination of single-cell sequencing and spatial transcriptomics to identify the molecular mechanisms that lead to immunosuppression in HGSC. Primary tumors consistently showed a more active immune microenvironment than did omental tumors. In addition, we found that untreated primary tumors were mostly populated by dysfunctional CD4 and CD8 T cells in later stages of differentiation; this, in turn, was correlated with expression changes in the interferon α and γ pathways in epithelial cells, showing that cross-communication between the epithelial and immune compartments is important for immune suppression in HGSC. These findings could have implications for the design of clinical trials with immune-modulating drugs.

Fig. 1. Schematic workflow and single-cell analysis of major cell types across HGSC specimens.

A Schematic workflow. A total of 19 freshly resected HGSC specimens were collected. Samples were processed using 10X genomic Chromium Single cell 3' v3. Created with BioRender.com using data contained elsewhere in the paper. B UMAP visualization of major cell types that were used for the subsequent analysis by cell type, tissue type, treatment, and tissue sample (patients). C UMAP visualization of the lineage canonical marker. D Cell count (up) and cell percentage (down) of the major cell types for each sample analyzed.

Fig. 2. Characterization of epithelial cell heterogeneity across patients, tissue type, and treatments.

A UMAP visualization of epithelial cells by clusters and samples (patients). B CNV profiles using InferCNV to distinguish cancer cells from normal epithelial cells. The color-coded bars indicate sample, cluster, treatment, tissue type, and InferCNV prediction. C CNV profile of patient Pt_10 using InferCNV to distinguish subclusters of epithelial cells on the basis of their CNV characteristics. The color-coded bars indicate cluster (C1 and C2), treatment (untreated), and tissue type (primary). D UMAP visualization of C1 and C2 shows a clear distinction on the basis of CNV inference. E UMAP visualization of the CytoTrace score of patient Pt_10, showing C2 as a more differentiated cluster. F Pseudotime trajectory of the cancer cells from patient Pt_10. Pseudotime was colored in a gradient from dark to light blue. The start of pseudotime is indicated by dark blue and the end by light blue. ssGSEA of (G) omentum versus primary tumor tissues, H omentum versus primary treated tumor tissues, I primary untreated tumors versus primary NACT-treated tumors, and J NACT-treated versus untreated omentum tumors.

Fig. 3. Immune cell composition, distribution, and correlation with survival across HGSC patient groups.

A UMAP of immune cells from all patients, distributed by cell type: B cells, T cells, natural killer cells (lymphoid cells), and myeloid cells. B Immune cell distribution across all patients. Top: immune cell type by group; Bottom: tissue distribution of immune cells by Ro/e analysis in different groups. C Immune cell proportion of each group of patients showed statistically significantly different proportions between groups. D Correlation between immune cell proportion and overall survival of patients with HGSC. E Ecotypes (relative abundance) of immune cell types by cancer type and treatment.

Fig. 4. Differentiation trajectory and functional state of CD8 and CD4 T cells by treatment and tissue type.

Differentiation trajectory analysis of A CD8 T cells and B CD4 T cells projecting the expression of multiple cell types representing different stages of differentiation. Inference of T cell functionality based on gene expression modules for C T cell cytotoxicity, D T cell dysfunction, and E T-reg scores, showing clustering by tissue type and treatment.

Fig. 5. Co-mapping of CD8 T cells, epithelial areas, and IFNα and γ pathways using spatial transcriptomics (Visium, 10X Genomics).

Representative tissue sections from three HGSC patients (Pt_12_OM, Pt_19_P, and Pt_19_OM), showing co-localization of CD8 T cells, tumor areas, and interferon signaling pathways (IFNα and IFNγ). Red areas indicate regions of high expression. Each row corresponds to a different patient sample, and each column displays the spatial distribution of one feature, demonstrating spatial relationships between immune infiltration, tumor regions, and interferon pathway activation.