Abstract 103: Spatially resolved transcriptomics identified distinct tumor-stroma crosstalk networks in long term ovarian cancer survivors

Abstract

Abstract The tumor microenvironment (TME), composed primarily of fibroblasts, endothelial cells, lymphocytic infiltrates and extracellular matrix proteins, can directly affect cancer cell growth, migration, differentiation as well as immune cell landscapes, thereby presenting a unique aspect of diagnosing and treating cancer. Cancer-associated fibroblasts (CAF) and immune cells are the main orchestrators of the ovarian cancer TME. We hypothesized that spatial transcriptomics (ST), which can provide a rich spatial context to gene expression, will identify crosstalk signaling networks among various cell types in the ovarian TME. Advanced stage high-grade serous ovarian cancer (HGSC) tissue frozen sections from two treatment naïve long-term survivors (LTS, overall survival ≥ 10 years) and two short-term survivors (STS, overall survival ≤ 2 years) were placed onto a ST expression slide pre-spotted with 1000 unique barcoded RNA-capturing probes. The ST slides were then stained with H&E and imaged. Tissue permeabilization and cDNA synthesis were then performed directly on the tissue section. The derived barcoded cDNA libraries from each sample were sequenced using Illumina NextSeq500 flow cells. After RNA-Seq analysis, supervised analysis was then performed by selecting different regions (based on tissue histology) in the tumor, tumor/stroma mixture, stroma close or far away from tumor areas to identify differentially expressed genes between STS and LTS. Stromal cells from different spatial locations had unique differentially expressed genes based on their proximity (11 genes in LTS; 11 genes in STS) or remoteness (18 genes in LTS; 10 genes in STS) to cancer cell compartments. Among the major modulated gene networks, higher levels of antigen-presenting molecules were found in the stroma in close proximity to the tumor cell nest than in stroma located far away in LTS tumor samples but not in the STS samples, suggesting an increased number of activated antigen-presenting cells in LTS tumors. In addition, higher levels of adhesion molecules and angiogenetic factors were found in the stroma in close proximity to the tumor cell nest than in stroma located far away in STS tumor samples but not in LTS samples, suggesting that adhesion molecules and angiogenetic factors produced by stromal cells may facilitate the invasiveness of the tumor cells, which subsequently leads to short term survival in patients with the disease. Our findings demonstrate for the first time that spatially resolved transcriptomics allows the identification of prognostic biomarkers associated with overall survival in HGSC patients. Further studies using deconvoluted ST data to further delineate the specific stromal cell subtypes in close proximity to tumors, and imaging mass cytometry (IMC) or multiplexed immunohistochemistry to validate the protein expression of the potential biomarkers are ongoing. Citation Format: Sammy Ferri-Borgogno, Jianting Sheng, Ying Zhu, Kwong K. Wong, Stephen T. Wong, Samuel C. Mok. Spatially resolved transcriptomics identified distinct tumor-stroma crosstalk networks in long term ovarian cancer survivors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 103.