Integrative analysis of machine learning and single-cell sequencing identifies IGHG1 as a therapeutic target for epimedium in diabetes-gastric cancer comorbidity.

BACKGROUND: The clinical comorbidity of diabetes mellitus (DM) and gastric cancer (GC) presents a significant healthcare challenge, as these two conditions often synergistically promote disease progression. Although Epimedium is known for its anti-tumor and metabolic regulatory properties, the precise molecular mechanism by which it intervenes in the DM-GC comorbidity remains poorly understood. METHODS: We integrated transcriptomic datasets of DM and GC to identify common differentially expressed genes (DEGs). Machine learning algorithms, enhanced by SHapley Additive exPlanations (SHAP) analysis, were utilized to isolate core genes within both disease contexts. These candidates were then intersected with the predicted targets of Epimedium. To refine the targets, single-cell RNA sequencing (scRNA-seq) analysis was performed to evaluate cell-specific expression patterns. Additionally, CCK-8, Western blot, and qPCR assays were conducted in human HGC-27 cells to validate the therapeutic efficacy of Epimedium in vitro. RESULTS: Eight common DEGs were identified, leading to the selection of GABRA1 and IGHG1 as primary hub genes. scRNA-seq analysis revealed that IGHG1 was significantly overexpressed in GC tissues compared to normal tissues, particularly within the B-cell population, whereas GABRA1 showed insufficient expression levels for further analysis. In cellular verification, HGC-27 cell viability was significantly inhibited by Epimedium in a dose-dependent manner, with a half-maximal inhibitory concentration (IC50) of 632.3 μg/ml. Furthermore, the expression of IGHG1 was markedly downregulated at both the transcriptional and translational levels following Epimedium treatment. CONCLUSIONS: Our findings identify IGHG1 as a critical molecular bridge in the DM-GC comorbidity. Epimedium exerts potent anti-tumor effects by targeting and downregulating IGHG1 expression. This study provides a novel therapeutic strategy and a robust theoretical basis for the treatment of patients suffering from concurrent diabetes and gastric cancer.