Abstract 6167: Spatial transcriptomic analysis identifies pancreatic cancer cell surface markers for imaging and therapy

Abstract Pancreatic cysts and precancer lesions have been reported in up to 50% of adults; however, few of the precancer lesions develop into aggressive pancreatic ductal adenocarcinoma (PDAC). Strategies for the differentiation of precancers and PDAC are needed. Recently proposed molecular imaging targets to accomplish this goal include integrin αvβ6 and fibroblast activation protein (FAP). To discover and quantitatively evaluate potential imaging and therapeutic targets, we performed Visium spatial transcriptomic (ST) sequencing on human surgical sections of normal tissue adjacent to precancers, intraductal papillary mucinous neoplasms and PDAC. All human tissues were collected in accordance with the Stanford Institutional Review Board. The resulting data were correlated with immunohistochemistry (IHC) and combined with analysis with publicly available spatial and single cell PDAC datasets to create a quantitative PDAC cancer surfaceome score. Pseudotime and receiver operating characteristic (ROC) analyses were also performed. We found that clusters with high cancer surfacesome scores were spatially correlated with a cellular morphology consistent with cancer. In the cancer cluster, the expression of surface markers CLDN4, GPRC5A, TSPAN8, and CEACAM5 was enhanced more than 13-fold compared with the normal pancreas, while FAP and ITGB6 were differentially expressed by less than 4-fold. Within cancer clusters, the fraction of spots overexpressing CLDN4, S100P, ITGB6, MUC5AC, TFF1, CEACAM5, and FAP was 88%, 87%, 78%, 66%, 59%, 47%, and 38%, respectively. The Pearson’s correlation across space and expression intensity between CLDN4, S100P, GPRC5A, TSPAN8, MUC1 and TFF1 was greater than 0. 6 for all combinations. The correlation of mRNA encoding ITGB6 and FAP with cancer surfaceome markers was smaller, e. g. 0. 37 and 0. 01 versus S100P, respectively. The area under the curve (AUC) for the differentiation of PDAC from normal or precancerous tissue for CLDN4, S100P, MUC5AC, TFF1, CEACAM5, and TSPAN8 was 0. 9, 0. 5, 0. 3, 0. 5, 0. 8, and 0. 4, respectively. From pseudotime trajectory analysis, we found that MUC5AC and TFF1 increased at a precancerous pseudotime whereas CLDN4 increased in time and space in spots corresponding to a PDAC-specific surfaceome score. This work demonstrates the feasibility of using spatial transcriptomics and IHC in human pancreatic cancer samples to discover and compare the expression of cell surface markers that can be employed in imaging and therapeutic protocols. Citation Format: James Wang, Aris J. Kare, Martn K. Schneider, Jai Woong Seo, Andrei Iagaru, Gregory W. Charville, Walter G. Park, Katherine W. Ferrara. Spatial transcriptomic analysis identifies pancreatic cancer cell surface markers for imaging and therapy abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts) ; 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84 (6Suppl): Abstract nr 6167.