Abstract 2199: Identifying genomic heterogeneity at single-cell resolution in endometrial cancer

Introduction: Endometrial cancer accounts for ~76,000 deaths amongst women worldwide. In the United States, it is estimated 65,620 new cases and 12,590 deaths occurring in 2020, being the fourth most common gynecologic malignancy in the country. It has been reported that specific molecular characteristics in tissues are associated with poorer prognosis. Publicly available data generated from bulk molecular profiling of tumors has provided a molecular taxonomy of endometrial cancers with common alterations occurring in a number of known oncogenic pathways. However, little is known about the degree of heterogeneity in endometrial cancer. Thus, we have set out to utilize single cell technologies to assess tumor heterogeneity in endometrial cancers. Methods: Our preliminary study includes serous and mixed lineage endometrial tumors from five (3 African Americans, 1 Hispanic/Latino, 1 Caucasian) cases. In addition to bulk whole exome sequencing, we utilized single cell whole genome sequencing to assess global copy number heterogeneity. OCT-embedded frozen tumor sections were dissociated to collect nuclei, which were subjected to single nuclei sequencing using the 10x ChromiumTM single cell copy number variation (scCNV) assay targeting 500 cells per sample. Single cell libraries were quality assessed and sequenced using the Illumina NovaSeq6000 system. Data were processed using the 10X Genomics CellRanger pipeline for assessment of copy number heterogeneity. Results: We generated scCNV data on a total of 1,546 cells across five tumors. Hierarchical clustering and visual inspection of scCNV data reveals obvious somatic copy number tumor cell heterogeneity in all samples, including single and mixed histology tumors. Heterogeneity was mostly distinguished by whole chromosome or chromosome arm level gains or losses. Clonal focal amplifications were detected at 5p, 8p, 8q, and 17q encompassing known oncogenes. As we are able to identify clustered diploid populations of normal cells, we are using these data to perform segregation analysis for CGH analysis and the identification of somatic mutations in clonal populations of cells. These data are being compared to bulk exome sequencing data to determine the power of scCNV for detecting clonal populations of tumor cells. Conclusions: This project used high resolution single cell sequencing in five endometrial cancers with varying histologies. scCNV analysis provided clear evidence of heterogeneity in all tumors that were