Decoding epithelial-T cell interactions in colorectal cancer through single-cell and spatial transcriptomics

<h4>Background</h4>Colorectal cancer (CRC) is a heterogeneous malignancy shaped by genetic alterations and immune microenvironmental interactions. While single-cell RNA sequencing (scRNA-seq) has uncovered diverse epithelial and T cell subsets, the spatial architecture of their crosstalk remains poorly understood.<h4>Methods</h4>We integrated scRNA-seq and spatial transcriptomics data from 36 CRC patients to comprehensively characterize epithelial and T cell heterogeneity, differentiation dynamics, and intercellular communication networks. Data preprocessing and integration were conducted using Seurat and Harmony. Cell trajectory inference was performed via CytoTRACE and Slingshot. Ligand-receptor signaling was assessed using CellChat, while spatial mapping was achieved using CellTrek.<h4>Results</h4>Nine epithelial and eight T cell subpopulations were identified, each exhibiting distinct transcriptional states, CNV burdens, and pseudotime trajectories. CD8⁺ T_GZMK cells displayed tumor-specific activation signatures and engaged in enriched communication with epithelial subsets through CEACAM, APP, and MIF signaling pathways. Spatial transcriptomics confirmed the in situ colocalization of CD8⁺ T_GZMK cells with PTP4A3⁺ epithelial cells in tumor regions, revealing spatially organized epithelial-immune niches. These interactions suggest potential mechanisms of immune modulation and tumor progression.<h4>Conclusion</h4>This study provides an integrative single-cell and spatial atlas of CRC, revealing structured epithelial-T cell communication and spatial architecture within the tumor microenvironment. Our findings offer novel insights into immune-epithelial crosstalk and identify signaling pathways that may serve as therapeutic targets or biomarkers for CRC precision treatment.