THU559 3D Immuno-smFISH Reveals ER-responsive Cells At The Single-cell Level In ER+ Breast Cancer

Abstract

Abstract Disclosure: L.G. Kammel: None. R.N. Wine: None. J. Rodriguez: None. Abstract: The current guidelines by the American Society of Clinical Oncology/College of American Pathologists recommend that breast tumors with ≥1% nuclear estrogen receptor (ER) expression by immunohistochemistry be classified as ER+ based on evidence that these tumors are clinically responsive to endocrine therapy (1). However, this recommendation has been challenged by in vitro data from ER+ breast cancer cells showing that the cellular response to ER is not directly correlated to the expression of ER or the activation of ER per cell, but rather the allelic state of ER target genes among individual cells (2, 3). In parallel, single-cell RNA sequencing of ER+ breast cancer cells revealed that the steroid hormone receptor-induced response of individual genes varies widely from cell-to-cell and that in the average individual cell, less than a third of hormone-regulated genes show a transcriptional response (4). Together, these findings suggest that the clinical outcome of an ER+ breast tumor to endocrine therapy would be better predicted by a measure of the induced response to ER than ER status alone, but such a strategy does not exist. Here, we developed an approach based on single molecule fluorescent in situ hybridization (smFISH) combined with immunohistochemistry (immuno-smFISH) to simultaneously determine ER expression and the ER transcriptional response at single-cell resolution in the 3D context of a tissue or tumor. Briefly, we adapted a protocol for hybridization chain reaction to label multiple RNAs in thick fixed tissues followed by traditional immunohistochemistry. We then optically cleared tissues and imaged them by lightsheet microscopy. We quantified RNA and protein expression in the 3D space by integrating the open-source cell segmentation software Cell Pose with a fluorescent spot detection pipeline. Using this quantification, we calculated the RNA distributions of ER target genes in individual cells from a mouse mammary gland. We then used the colocalization of intronic and exonic RNA to identify active transcription sites and used this value to calculate the fraction of cells transcribing ER-target genes among the ER+ population. This fractional response represents the percent of cells expected to show a response to ER-targeted therapy, and applying this measurement in 3D allows for accurate analysis of complex tissues and tumors. In summary, we developed a comprehensive immuno-smFISH based-approach to characterize both ER expression and the cells responsive to ER at the single-cell level in 3D and we speculate that integrating these outcomes will improve the utility of ER as a predictive marker in breast cancer.