Exploring the mechanism underlying the therapeutic effects of butein in colorectal cancer using network pharmacology and single‐cell RNA sequencing data

Abstract Background Butein has shown substantial potential as a cancer treatment, but its precise mechanism of action in colorectal cancer (CRC) remains unclear. This study aimed to uncover the underlying mechanisms through which butein operates in CRC and to identify potential biomarkers through a comprehensive investigation. Methods Target genes associated with butein were sourced from SwissTargetPrediction, CTD, BindingDB and TargetNet. Gene expression data from the GSE38026 dataset and the single‐cell dataset (GSE222300) were retrieved from the Gene Expression Omnibus database. The activation of disease‐related pathways was assessed using Kyoto Encyclopedia of Genes and Genomes, Gene Ontology and differential gene analysis. Disease‐associated genes were identified through differential analysis and weighted gene co‐expression network analysis (WGCNA). The protein–protein interaction network was utilized to pinpoint potential drug targets. Molecular complex detection (MCODE) analysis was employed to uncover relevant genes influenced by butein within key subgroup networks. Machine learning techniques were applied for the screening of potential biomarkers, with receiver operating characteristic curves used to evaluate their clinical significance. Single‐cell analysis was conducted to assess the pharmacological targets of butein in CRC, with validation performed using the external dataset GSE40967. Results A total of 232 target genes for butein were identified. Functional enrichment analysis revealed significant enrichment of signaling pathways, including mitogen‐activated protein kinase, JAK‐STAT and NF‐ κ B, among these genes. Differential analysis, in conjunction with WGCNA, yielded 520 disease‐related genes. Subsequently, a disease‐drug‐gene network consisting of 727 targets was established, and a subnetwork containing 56 crucial genes was extracted. Important pathways such as the FoxO signaling pathway exhibited significant enrichment within these key genes. Machine learning applied to the 56 important genes led to the identification of a potential biomarker, UBE2C. Receiver operating characteristic analysis demonstrated the excellent clinical predictive utility of UBE2C. Single‐cell analysis suggested that butein’s therapeutic effects might be linked to its influence on epithelial and T cells, with UBE2C expression associated with these cell types. Validation using the external dataset GSE40967 further confirmed the exceptional clinical predictive capability of UBE2C. Conclusion This study combines network pharmacology with single‐cell analysis to unravel the mechanisms underlying butein’s effects in CRC. Notably, UBE2C emerged as a promising biomarker with superior clinical efficacy. These research findings contribute significantly to our understanding of specific molecular mechanisms, potentially shaping future clinical practices.