Abstract P5-06-03: Single-cell and spatial approach to study endocrine therapy sensitivity in breast cancer models with heterogeneous estrogen receptor expression

Abstract Although most women with estrogen receptor (ER) positive breast tumors benefit from endocrine therapy (ET), up to 40% of these patients will experience relapse. The percentage of tumor cells expressing ER varies from 1 to 100%, and some studies suggest that higher expression of ER positively correlates with ET efficiency and a reduced rate of tumor relapse. Previously, we identified a cell population with active NFĸB and integrated stress response pathways that overcome selective pressure of ET and contribute to tumor relapse. However, we and others routinely use models that are enriched for ER-expressing cells to study the ET effect. Less is known about ET efficacy in breast cancers with low to moderate ER expression, which are considered to be more aggressive. In this study, we utilized patient-derived xenograft organoids (PDxOs) and single-cell RNA sequencing (scRNAseq) approach to examine how breast cancer models with low to moderate ER expression respond to ET. Alongside established PDxOs (HCI003 and HCI017), we developed two new models (UIC013 and UIC020), where 22%, 30%, 12%, and 7% of cells express ERα, respectively. Using scRNAseq and various bioinformatic tools, we found several cell populations that are common among all models, but vary in terms of prevalence, as well as some populations that were largely model-specific, indicating the models’ inter- and intra- tumors heterogeneity. To predict which cell populations are sensitive to ET and which may harbor aggressive features, we examined well-established gene signatures for ER activity, ET resistance, and metastatic relapse in each cell cluster using Functional Enrichment Analysis. However, the results were inconclusive because of inconsistency among gene signatures, suggesting that the variety of models, platforms, and experimental setups used to derive these gene signatures could account for such variability. Additionally, these gene signatures were generated from bulk-RNA-seq studies, which average the signals obtained from all cell populations within a tumor and may mask signals from smaller cell populations. To experimentally identify ET-sensitive and ET-resistant cell populations, we treated each PDxO with fulvestrant and 4-hydroxytamoxifen. scRNAseq revealed that some ER-active cells were lost following ET, indicating responsiveness, while others remained unchanged or were enriched, indicating resistance to ET. Moreover, comparative analysis of all ET-resistant cell populations among the models showed their high degree of heterogeneity, suggesting that each cell population might use different mechanisms to overcome the selective pressure of ET. To find alternative therapies for the ET-resistant cells, we performed DREEP analysis (DRug Estimation from single-cell Expression Profile), a tool predicting drug responses for 450 compounds using scRNAseq data. Multiple compounds were predicted to be effective in combination with ET, including vandetanib, a VEGFR2 inhibitor currently in clinical trials with fulvestrant for metastatic breast cancer. ET also altered ER-resistant cell transcriptomes, predicting sensitivity to several additional compounds. Overall, our findings suggest that stratifying breast cancer cell populations based on ET response, rather than ER expression or activity alone, can identify new cell populations and therapeutic strategies, thereby improving outcomes for women with breast tumors with heterogeneous ER expression. Citation Format: Svetlana Semina, Rosemary J. Huggins, Huiping Zhao, Debra Tonetti, Kent Hoskins, Geoffrey L. Greene, Jonna Frasor. Single-cell and spatial approach to study endocrine therapy sensitivity in breast cancer models with heterogeneous estrogen receptor expression [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P5-06-03.