Abstract P4-04-06: Integrative analysis of single-cell transcriptomic and spatial profiles characterized distinct tumor microenvironment phenotypes in hormone receptor positive (HR+) breast cancer

Abstract

Abstract Background Clinical trials have shown minimal activity of immune checkpoint blockade (ICB) in HR+ breast cancer. While this is, in part, attributable to low tumor-infiltrating lymphocyte (TIL) levels in HR+ breast cancer, a thorough investigation of the immunologic aspects of the tumor microenvironment (TME) is required to fully understand the limited efficacy of ICB in HR+ disease. We sought to characterize the TME of HR+ breast cancer at single cell resolution using single-cell RNA-seq (scRNA-seq), paired with tissue-based cyclic immunofluorescence (t-CyCIF), a multiplexed imaging technology for formalin-fixed paraffin-embedded (FFPE) tissue sections across up to 30 distinct antigen channels. We hypothesized that integrative analysis of clinical samples using these technologies would elucidate inter-patient heterogeneity of the TME of HR+ breast cancer. Methods Surgically excised primary tumor samples were obtained from five, treatment-naïve, HR+ breast cancer patients. Each tumor was divided with half preserved as a FFPE block for t-CyCIF and bulk RNA-seq, and half frozen and later dissociated into single cells for scRNA-seq. For each t-CyCIF cycle, slides were stained with 3 antibodies and a DNA dye. The resultant image was captured with a CyteFinder (RareCyte) slide scanning fluorescence microscope. Each slide was stained for 3-7 cycles (9-21 antibodies), depending on tissue stability. Acquired images were processed (illumination correction, alignment, stitching, and segmentation) using mcmicro (https://mcmicro.org/) and the fluorescence intensity per cell per antibody was determined. After identifying regions where cells were appropriately stained throughout the cycles, the tumor, stromal, and immune cell types were determined based on expressed antigens. Samples were further characterized based on the spatial distribution of T cells within the tumor and categorized into three classes: immunologically cold (lacking T cells), immune stromal-restricted (T cells located predominantly in the stromal) and highly inflamed (T cells located throughout entire tumor). CyCIF data was compared to bulk RNA-seq data for each sample. For scRNA-seq, two samples with high cellularity were identified. The single-cell barcoded library was constructed with 10X Chromium and sequenced with NextSeq (Illumina). The cell types were compared as were gene expressions of the corresponding T cell populations between the two samples. For all experiments, downstream analysis was performed in R. Results CyCIF analysis revealed diverse compositions of immune cell types. In particular, the spatial distributions of T cells revealed that the five samples were classified into the three TME classes: 2 immune cold, 1 immune stromal-restricted, and 2 highly inflamed. This trend was consistent in the corresponding bulk RNA-seq data, in which we found higher expression of an interferon-gamma signaling signature in the two highly-inflamed tumors. For the two scRNA-seq samples, one corresponds to the immunologically cold class in CyCIF and the other to the immune stromal-restricted class. Consistent with CyCIF analysis, scRNA-seq revealed that the immune stromal-restricted tumor had higher T cell infiltration than the immune cold tumor, particularly with a higher abundance of naïve T cells, Th17 cells, and helper T cells. Conclusions While HR+ breast cancer is generally considered immunologically cold, our pilot analysis integrating CyCIF, scRNA-seq, and bulk RNA-seq revealed considerable diversity between the five HR+ samples, including the distribution of T cells relative to tumor cells. Additional samples are being collected to allow for more detailed exploration of the TME heterogeneity of HR+ breast cancer. Citation Format: Kenichi Shimada, Yvonne X Cui, Jonathan S Goldberg, Ricardo Pastorello, Janae Davis, Tuulia Vallius, Lukas Kania, Ashka Patel, Mckenna Moore, Esther R Ogayo, Deborah Dillon, Peter K Sorger, Jennifer L Guerriero, Elizabeth A Mittendorf. Integrative analysis of single-cell transcriptomic and spatial profiles characterized distinct tumor microenvironment phenotypes in hormone receptor positive (HR+) breast cancer [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P4-04-06.