DOP110Spatial mapping of tumor-stromal cell interactions defines mesenchymal phenotypes in colitis-associated colorectal cancer

Abstract Background Colitis-associated colorectal cancer (C-CRC) arises in the context of chronic inflammation, where tumor microenvironment interactions strongly influence progression. Spatial transcriptomics enables characterization of these interactions by retaining tissue architecture and cellular context. We aimed to characterize the cellular and microenvironmental features that underlie tumor development in C-CRC. Methods Seven tissue samples from three C-CRC patients were analyzed using consecutive sections processed by 10x Xenium and 10x Visium. Spatial coordinates of Xenium cells were mapped onto Visium spots to enable integrated spatial transcriptomic profiling (Figure 1A). Results Xenium analysis identified 3,506,705 cells with diverse epithelial, fibroblast, and immune cell populations (Figure 1B-C). Tumor cell subclustering revealed two functionally segregated populations: (1) proliferating tumor cells characterized by cell-cycle genes (MKI67, TOP2A) and (2) invasive tumor cells with elevated epithelial mesenchymal transition (EMT)-associated genes (CD44, CXCL1). Cellular neighborhood analysis based on distances between cells identified a distinct microenvironmental niche unique to invasive tumor cells, characterized by spatial proximity of cancer-associated fibroblasts (CAFs) and SPP1+ macrophages (Figure 1D). Pathway enrichment analysis showed selective up-regulation of angiogenesis, hypoxia, and inflammatory pathways in invasive versus proliferating tumor regions (Figure 1E). Receptor–ligand analysis identified TNC–ITGB6 interactions between CAFs and invasive tumor cells (Figure 2A). A differentially expressed gene signature distinguishing invasive from proliferating tumor cells stratified TCGA sporadic CRC samples into invasive-tumor-like and proliferating-tumor-like groups. Invasive-tumor-like tumors predominantly displayed consensus molecular subtype (CMS) 4 characteristics, consistent with the known enrichment of CMS4 in UC-CRC and supporting a role for CAF-mediated interactions in mesenchymal phenotype development (Figure 2B). Conclusion This integrated spatial transcriptomics study identified two distinct tumor populations in C-CRC with segregated microenvironmental contexts. The invasive tumor phenotype is characterized by CAF proximity, SPP1+ macrophages, and TNC–ITGB6-mediated signaling. These findings reveal that tumor-stromal composition and architecture are linked to mesenchymal phenotype development, suggesting the microenvironment as a potential target for biomarker and therapeutic development in C-CRC. Conflict of interest: Lee, Yoonho: No conflict of interest Kim, Yongjae: No conflict of interest Baek, Jiwon: No conflict of interest Hwang, Sung Wook: no conflicts Lee, Jong Lyul: None. Park, Sang Hyoung: No conflicts Yang, Dong-Hoon: Nothing to declare Kim, Jihun: No conflict of interest Yang, Suk-Kyun: No conflict of interest Han, Buhm: No conflict of interest Song, Kyuyoung: No conflict of interest Yoon, Yong Sik: No conflict of interest Ye, Byong Duk: Byong Duk Ye reports consulting fees from AbbVie Korea, BMS Pharmaceutical Korea Ltd., Celltrion, Chong Kun Dang Pharm, CJ Red BIO, Curacle, Daewoong Pharm, Dong-A ST, Ferring Korea, Hanmi Pharmaceutical, Imscout, IQVIA, Johnson & Johnson, Johnson & Johnson Korea, Jeil Pharmaceutical Co., Kangstem Biotech, Korea Otsuka Pharm, Korea United Pharm, Lilly Korea, Medtronic Korea, NanoEntek, ORGANOIDSCIENCES Ltd., Pfizer Korea, Samsung Bioepis, Takeda, Takeda Korea and Yuhan speaker fees from AbbVie Korea, BMS Pharmaceutical Korea Ltd., Celltrion, Cornerstones Health, Curacle, Daewoong Pharm, Eisai Korea, Ferring Korea, IQVIA, Johnson & Johnson Korea, Pfizer Korea, Samsung Bioepis, and Takeda Korea and research support from Celltrion and Pfizer Korea. Lee, Ho-Su: No conflict of interest