Single Cell RNASeq Data in Colorectal Cancer
Description
Colon adenocarcinoma (CA) is a prevalent gastrointestinal malignancy, with significantly higher incidence than small intestine (SI) adenocarcinoma. However, it remains lack of comprehensive understanding about whether the distinct immune microenvironment between SI and colon is related to CA. Here, we performed bulk and single-cell analyses on immune cells of multiple matched tissues of CA patients and health donors, identifying the ileum-likeness pattern within immune microenvironment as a key feature of CA patients with favorable prognosis. The ileum-likeness pattern is primarily characterized by the less TGF-β+ long-lived plasma cells and the abundant cytotoxic and unexhausted CD160+ T cells. Tracking T cell clones confirmed that CD160+ T cells are highly cytotoxic and less prone to exhaustion. CD160 overexpression on CD8+ T cells significantly enhances anti-tumor immunity and efficiently inhibits tumor growth upon adoptive transfer to tumor-bearing mice. Our study provides comprehensive immune landscape within intestinal tissues and novel insights into intraepithelial T cells in CA, underscores potential targets for CA treatment.
Files
Steps to reproduce
We obtained bulk RNA and single-cell RNA/TCR/BCR sequencing data from multiple regional sites in both healthy donors and CA patients, including PBMC, lymph nodes, ileum, colon and tumor. Library size normalization was performed using NormalizeData function in Seurat (v4.1.3) on the filtered gene-barcode matrices. FindVariableFeatures of Seurat was applied to the normalized data to identify variable genes for unsupervised clustering. TCR/BCR variable genes were masked when performing clustering analysis. Principal component analysis (PCA) was performed on the top 2500 variable genes for dimension reduction. And Harmony algorithm was used for batch effect correction in the reduced dimension space of PCA with default parameters. The function FindNeighbors of Seurat was used to construct the shared nearest neighbor graph, based on which the function FindClusters of Seurat was used for unsupervised clustering. Different resolution parameters were examined in order to determine the optimal number of clusters. For visualization, dimensionality reduction and 2-D visualization of the single cell clusters was performed using Uniform Manifold Approximation and Projection (UMAP) with Seurat function RunUMAP. Multiple layers of information including the cluster distribution, cluster specific genes, canonical immune cell markers and transcription factors were integrated and carefully reviewed to define cell subtypes.
Institutions
- Harbin Medical University