Abstract C115: Genomic evolution of pancreatic cancer at single-cell resolution

Abstract

Abstract Late-stage disease consists of more than half of pancreatic ductal adenocarcinoma (PDAC) at diagnosis, but most samples have low tumor purity that limits bulk DNA sequencing. Additionally, as clonal evolution happens at single-cell level, bulk sequencing misses many insights such as subclones, colocalization of events etc. Thus, we aim to refine our understanding of PDAC clonal evolution and heterogeneity with single-cell DNA sequencing. Previously we developed a scalable, high-throughput single-nucleus DNA-seq (snDNA-seq) method (doi.org/10.1038/s41467-023-36344-z) for archival primary solid tumor samples. Here we used it to sequence over 100,000 single nuclei from 70 archival primary samples of 18 PDAC patients, with a targeted panel covering 254 most relevant genes for PDAC. The cohort included 5 early- and 10 late-stage diagnoses whose primary tumors and metastases were multiregionally sampled, as well as 3 longitudinal cases whose tumors were collected before and after treatment. We also developed a set of computational tools to call short variant (SNV) and copy number variation (CNV) events and use both signals to infer phylogenies from this dataset. Owing to snDNA-seq’s sensitivity to detect subclonal, small-scale genomic events, we found that alterations to SMAD4 and CDKN2A were more frequent than previously shown and enriched in late-stage disease. Many manifested as focal (<1kbp) homozygous deletions difficult to detect by bulk. TGF-β signaling pathway was convergently targeted for inactivation two late-stage cases, which were wildtype for SMAD4, but had homozygous deletions of TGFBR2 and ACVR1B and a nonsense mutation to ARID2 respectively. With single cell phylogenies, we observed that PDAC generally displayed linear rather than bifurcating clonal evolution, which could be linked to its aggressiveness that does not permit time for significant clonal branching. Driver SNVs were mostly truncal (present in all tumor cells) irrespective of stage. High spatial homogeneity with respect to all (driver and passenger) SNVs was seen across both primary tumors and metastases, suggesting that PDAC invasion and metastasis were late events in its course of genetic evolution. The major source of spatial and temporal heterogeneity was CNV. Although most were large-scale and appeared random, some focally targeted important genes, such as step-wise deletion of the CDKN2A and SMAD4 gene and focal amplification of MTOR and GATA6 genes, suggesting continuous evolution towards the most fit genotype. In sum, snDNA-seq revealed more PDAC driver gene alterations than previously estimated. Rather than SNVs, CNVs caused by increased chromosomal instability generated the genetic diversity for selection to operate on as disease progress. Equally important is we demonstrate a single-cell analysis framework that can reveal high-resolution clonal evolution patterns which can be useful for many further studies of cancer evolution. Citation Format: Haochen Zhang, Palash Sashittal, Elias-Ramzey Karnoub, Shigeaki Umeda, Nicolas Lecomte, Jungeui Hong, Akhil Jakatdar, Katelyn Mullen, Akimasa Hayashi, Caitlin A. McIntyre, Benjamin J. Raphael, Christine A. Iacobuzio-Donahue. Genomic evolution of pancreatic cancer at single-cell resolution [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Pancreatic Cancer; 2023 Sep 27-30; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(2 Suppl):Abstract nr C115.