Combination of Metabolomics and Bioinformatics to Reveal the Mechanism of Luteolin in the Treatment of Cervical Cancer

ABSTRACT The incidence of cervical cancer is high among women globally. The potential therapeutic efficacy of luteolin in the treatment of cervical cancer has been identified. Therefore, we aim to elucidate the mechanism of action of luteolin in the treatment of cervical cancer through a comprehensive approach that integrates metabolomics with bioinformatics. The first step involved the identification of differential metabolites through UHPLC‐Q‐Orbitrap‐MS, which were then utilized for enrichment analysis of metabolic pathways and to determine the targets associated with these differential metabolites. Subsequently, the differential analysis and WGCNA were employed to identify DEGs and functional module genes respectively. The common targets were obtained by intersecting the results from the aforementioned three analyses, followed by conducting GO and KEGG pathway enrichment analysis on these targets. Subsequently, PPI networks were constructed using these common targets, and key targets were identified utilizing the MCC, EPC, Degree, Closeness Centrality, Betweenness Centrality, and Bottleneck algorithms in the CytoHubba plug‐in. The subsequent steps involved the analysis of key genes for constructing a nomogram, conducting a ROC curve, examining content expression and survival analysis, and ultimately employing molecular docking to investigate the interaction between luteolin and crucial targets. The metabolomics analysis revealed the identification of a total of 45 distinct metabolites in this study, primarily associated with amino acid and nucleotide metabolism. The intersection of 773 differential metabolite targets, 3493 cervical cancer differential genes, and 3245 WGCNA‐associated module genes yielded a set of 32 target genes associated with luteolin therapy for cervical cancer. The GO and KEGG pathway enrichment analysis also revealed that these targets were primarily associated with amino acid metabolism and nucleotide metabolism. The CytoHubba plug‐in was utilized to identify three key genes (DMNT1, EZH2, and GMPS) through the application of multiple algorithms. Additionally, the datasets GSE63514, GSE67522, and GEPIA2 revealed a significant upregulation of all three genes in tumor tissue. ROC analysis demonstrated the good predictive ability of these three hub genes. Finally, the molecular docking results demonstrated the high binding affinity of luteolin towards DMNT1, EZH2, and GMPS. In conclusion, this study has unveiled the potential of luteolin in modulating amino acid and nucleotide metabolism for the treatment of cervical cancer, thereby providing a theoretical foundation for further investigation into the intricate association between luteolin and cervical cancer.