Abstract 5287: Decoding tumor-immune interactions in colorectal cancer liver metastases: Insights from single-cell analysis

Abstract Background: Liver is the most common site of colorectal cancer (CRC) metastasis. As liver microenvironment differs significantly from the colon microenvironment surrounding the tumor, we hypothesized that both cancer cells and liver microenvironment undergo changes that promote tumor growth and suppress the immune response. Here, we aimed to identify distinct cancer cell clusters in the liver and evaluate the upregulated interactions between immune cells in the liver. Method: We conducted single-cell RNA sequencing of primary and liver metastasis (CRLM) samples, along with paired normal samples from colon and liver, respectively. We conducted non-negative matrix factorization to define multiple epithelial cell clusters and evaluate proportional enrichment and interactions between immune cells. Results: To investigate colorectal epithelial cells and identify those contributing to liver metastasis, we analyzed a total of 273, 711 cells derived from 78 CRC samples (primary tumor from localized CRC, N = 44; primary tumor from CRLM, N = 21; metastatic tumor from CRLM, N = 13), and 33 normal samples (25 from the colon and 8 from the liver). Excluding five epithelial cell clusters that clustered with normal tissues, we defined 14 tumor-associated epithelial cell clusters. Three dominant clusters in primary colorectal cancers (EPI_PC) showed upregulation of hypoxia, stress, and MHC pathways. On the other hand, two dominant clusters in CRLM (EPI_LM) exhibited enriched hyperkeratosis and cell junction pathways, and these cells were found to be in a later stage of differentiation in trajectory analysis. Furthermore, these clusters exhibited an increase in the activation of angiogenesis and epithelial-mesenchymal transition (EMT) pathways, from localized CRC (without metastasis) to primary tumor from CRLM, and metastatic tumor from CRLM. To examine differences in interactions between epithelial cells and the tumor microenvironment in the colon and liver, we performed cell-to-cell interaction analysis focusing on CRLM and matched primary CRC samples. The analysis revealed upregulation of semaphorin 6, semaphorin 7, CD137, and TRAIL pathways between these epithelial cells and C1QC+ macrophage and CD8+ effector T cells. Notably, immune checkpoint signaling, including TIGIT and PVR pathways, was upregulated between epithelial cells and CD8+ exhausted T cells in the CRLM. Spatial transcriptomics showed distinct showed distinct distributions of EPI_PC and EPI_LM, presenting specific immune-cell interactions in some CRLM samples. Conclusions: Our study highlights that the CRLM-associated clusters were featured in a later stage of differentiation and high expression of angiogenesis and EMT pathways. Considering the upregulated pathway between EPI_LM and immune cells, treatment strategies targeting these interactions may be helpful for CRC patients with liver metastasis. Citation Format: Ji Hye Choi, Yeong Hak Bang, Kyung Yeon Han, Woong-Yang Park. Decoding tumor-immune interactions in colorectal cancer liver metastases: Insights from single-cell analysis [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 5287.