Abstract 7427: Genomic analysis of microenvironment composition by breast cancer molecular subtype using an integrated single-cell RNA sequencing cohort

Abstract Breast cancer is a heterogeneous disease clinically classified by expression of three key biomarkers, namely ER, PR, and HER2. Between clinical subtypes, the prognostic value of immune cell infiltration varies significantly. In triple-negative/basal-like tumors, tumor infiltrating lymphocytes (TILs) are positively prognostic, while in ER+/luminal tumors, TILs are less prognostic and sometimes associated with worse outcomes. This opposing, subtype-specific behavior suggests different roles for the immune system and is important to understand as immunomodulatory treatments are increasingly tested and utilized across breast cancer subtypes. Previous work using TCGA Breast Cancer bulk RNA sequencing data determined that among highly immune infiltrated tumors, there is significant variation in the immune microenvironment according to PAM50 molecular subtype. High-TIL luminal tumors were enriched for innate immune features, while high-TIL HER2-enriched and basal-like tumors had elevated expression of B and T cell features, respectively. To extend these findings, we curated an integrated dataset of single-cell RNA sequencing from 115 human breast tumors and 17 normal tissue specimens. Samples spanning all clinical subtypes came from four publicly available datasets, and common bioinformatic approaches were used to guide cell type annotation. Cell type frequencies in each sample were calculated and compared by PAM50 subtype. Basal-like tumors were enriched for multiple effector T cell and macrophage populations, while HER2-enriched exhibited increased frequency of IgG-secreting plasma cells. Conversely, luminal A tumors were enriched for various stromal cell types, including fibroblasts, endothelial cells, and perivascular-like cells; interestingly, a macrophage subset expressing the immunosuppressive marker TREM2 was also more abundant in luminal A. We identified highly immune infiltrated tumors from our cohort and compared cell type frequencies in these tumors by PAM50, finding that many baseline differences in microenvironment composition by subtype persist into the immune-high setting. Finally, we analyzed a subset of tumors from Bassez et al. (PMID: 33958794) that included T cell expansion data as a proxy for anti-PD1 immunotherapy response. In pre-treatment basal-like tumors, increased proportions of plasma cells, effector T cells, and regulatory T cells were associated with T cell expansion following anti-PD1. Ongoing work is exploring tumor-microenvironment crosstalk using ligand-receptor interaction prediction, with a focus on stromal cells in luminal tumors and B cell-T cell crosstalk in basal-like; for the latter, CD40-CD40LG interaction has emerged in preliminary analysis as a potential target for clinical intervention to improve patient outcomes and is currently being tested in mouse model systems. Citation Format: Constandina (Dina) E. O'Connell, Alexander V. Lobanov, Kevin R. Mott, Daniel P. Hollern, Brian C. Miller, Charles M. Perou. Genomic analysis of microenvironment composition by breast cancer molecular subtype using an integrated single-cell RNA sequencing cohort abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 7427.