Identifying Cell Vulnerability States in PDAC Tumors

Sarah Siemers

Mentor: Dr. Robert Suter, Department of Oncology, Lombardi Comprehensive Cancer Center, Georgetown University Medical Center.

Date/Time: August 22nd, 2024 at 3:40pm.

Abstract: Pancreatic Ductal Adenocarcinoma (PDAC) is one of the deadliest and most common forms of cancer, yet effective treatment options remain limited (Rahib et al., 2014). PDAC tumors are composed of epithelial cells in distinct transcriptional states that interact with diverse cell types in the tumor microenvironment (TME) (Poh & Ernst, 2021). The heterogeneity of PDAC tumors and the complexity of cell-cell interactions within them present significant challenges when developing effective treatments. To address these challenges, a systems-level discovery approach was employed to target PDAC tumor epithelial cells. Single-cell RNA sequencing (scRNA-seq) data from an established PDAC atlas (Steele et al., 2016; Peng et al., 2019) was integrated with genome-scale CRISPR screen data from the BROAD Institute’s Cancer Dependency Map (DepMap) to identify the most critical cells to target within a tumor. Tumor epithelial cells were parsed from the total epithelial cell population using copy-number inference analysis, InferCNV (Tickle et al., 2019). Sc0Day (Zero-day) (D’Antuono et al., In review at Cell Reports; DepMap B, 2023; Dempster et al., 2021) a novel framework developed in the lab, was used to identify PDAC gene dependency signatures and tumor epithelial cell vulnerability states. Our analysis identified 115 PDAC lineage-specific genes essential for PDAC cell viability, which were used for vulnerability state (VS) identification. Similarity scoring between essential gene profiles in scRNA-seq expression data from individual patients, coupled with hierarchical clustering of resulting signatures, uncovered four distinct conserved tumor epithelial cell VSs. Differentially expressed genes across all VSs were identified using MAST (Model-based Analysis of Single-cell Transcriptomics) (McDavid et al., 2024); among significant DEGs (p-adj < 0.05), 103 of the 115 PDAC essential genes were present, suggesting distinct vulnerabilities in heterogeneous tumor cell populations. Enrichment analysis revealed that each VS has unique genomic and enrichment profiles. These findings lay a basis for the further characterization and analysis of these VSs to develop novel heterogeneity-aware precision medicine approaches for this thus far incurable cancer.