Abstract 503: Tracking evolution of transcriptional landscapes during colorectal cancer development and progression

Abstract Colorectal cancer (CRC) has been classified into molecular subtypes using gene-level data from both bulk and single-cell RNA sequencing (scRNA-seq), leading to the development of the consensus molecular subtypes (CMS) and intrinsic CMS (iCMS). In addition to CRC studies, scRNA-seq and spatial transcriptomics characterization of human polyps, aligned with the conventional adenoma and serrated polyps, identified distinct driver mutations, cells-of-origin and iCMS traits associated with each polyp type. These gene-level transcriptional classification approaches were both strongly aligned to mutations underpinning the Vogelstein model of CRC development. However, when viewed using pathway-level information, the most dominant mechanistic signaling cascades in CRC were completely independent of these same mutational drivers. Therefore, to deliver a holistic landscape-view of the tumor ecosystem, we proposed that pathway-level subtyping could offer a broader understanding of non-genetic phenotypic signaling in CRC, leading to more informative stratification and therapeutic decision-making. Using this new approach in combination with bulk transcriptional data from n=1200+ tumors, alongside n=29, 452 scRNA-seq profiles derived from murine organoids, and n=49, 155 epithelial cells derived from n=63 human tumor samples, we identified three pathway-derived subtypes (PDS1-3), revealing previously overlooked changes in epithelial cell identities and phenotypic states. PDS2 represents tumors with stroma/immune-dense, regenerative stem-like features. PDS1 were associated with elevated MYC signaling, proliferation and canonical stem-cells with good prognosis, while PDS3 displayed a differentiated-like phenotype associated with reduced polycomb repressive complex 2 (PRC)-mediated gene repression and poorer prognosis. The inverse relationship between MYC/PRC targets within PDS1/PDS3 tumors supported the development of stem maturation index, which was directly related to cellular entropy and differentiation status of colonic cells allowing our tumor data to be projected onto a Waddington’s landscape. Given these associations with developmental patterns in established tumors, we next utilized the same pathway-level approach in combination with transcriptional data derived from n=2641 human polyps, where we observe that PDS traits can be established early in precancer development. Using a combination of trajectory analyzes, lineage identities and cellular phenotyping, these approaches have again revealed that, even in very early precancers, there is no association between Vogelstein-related mutational drivers and prominent phenotypic signaling cascades. While work is ongoing, the use of pathway-level information provides a unique insight into the dynamic nature of phenotypic evolution across early-to-late stages of colorectal precancers and tumors. Citation Format: Sudhir B. Malla, Natalie Fisher, Raheleh Amirkhah, Joanne Edwards, Owen Sansom, Philip D. Dunne. Tracking evolution of transcriptional landscapes during colorectal cancer development and progression [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 503.