Molecular Mechanisms of Chinese Medicine in Regulating Colorectal Cancer Immune Microenvironment: Insights from Single-Cell Transcriptomics and Network Pharmacology

Abstract

BackgroundThe molecular mechanisms underlying the efficacy of Traditional Chinese Medicine (TCM) in colorectal cancer treatment remain largely unexplored. We developed a computational systems biology approach integrating single-cell transcriptomics with network pharmacology to elucidate the potential mechanisms of TCM in modulating colorectal cancer progression. MethodsWe developed an integrated computational pipeline for multi-omics data analysis combining single-cell transcriptomics with network pharmacology. Raw single-cell RNA-seq data from 3 normal tissues and 3 colorectal tumors were obtained from GEO database and processed using a customized workflow in R. Quality control, normalization, and dimensionality reduction were performed using the Seurat v4.0 algorithm, followed by unsupervised clustering to identify cell subpopulations. Differentially expressed genes (DEGs) were identified using MAST algorithm with adjusted p-value < 0.05 and |log2FC| > 1.0. These computationally identified DEGs were subsequently mapped to a comprehensive Traditional Chinese Medicine (TCM) database using a network pharmacology approach to predict herb-target interactions. In parallel, we integrated TCGA RNA-seq data (STAR-counts) with clinical information, applying log2(TPM+1) transformation for normalization. We then implemented a machine learning-based correlation analysis to construct gene-cell-immunity-pathway networks, using weighted gene co-expression network analysis (WGCNA) to identify key regulatory modules. ResultsOur computational analysis of single-cell RNA-seq data identified 109 differentially expressed genes (DEGs) that define the molecular signature of colorectal cancer microenvironment. Clustering algorithms revealed 14 distinct cell subpopulations, with predominant immune cell infiltration, particularly B and T lymphocytes, suggesting a complex immune regulatory network. Network pharmacology analysis mapped these DEGs to potential therapeutic targets, computationally predicting interactions with 140 traditional Chinese herbs. These herbs were classified into 8 functional categories. Through integrative multi-omics analysis and pathway enrichment algorithms, we identified core regulatory networks comprising 23 genes and 39 significantly enriched signaling pathways (FDR < 0.01) that orchestrate immune cell function in the tumor microenvironment. Notably, our analysis in silico revealed previously uncharacterized gene-pathway interactions that may explain the immunomodulatory effects of specific herbal compounds. ConclusionsOur systems biology and computational analysis revealed a potential mechanism by which 8 categories of Chinese herbal medicines and 23 genes across 39 signaling pathways may regulate colorectal cancer progression through modulation of specific gene regulatory networks and immune cell functions. These findings demonstrate the value of integrative computational approaches in elucidating complex biological mechanisms of traditional medicines