目的 基于多组学数据与孟德尔随机化（Mendelian randomization，MR）因果推断方法，筛选驱动胃黏膜“炎-癌转化”的特征基因，并探索针对这些基因的潜在中药干预策略。方法 整合转录组数据与大规模遗传数据，通过差异分析、MR分析与蛋白质互作网络构建，筛选与胃癌风险存在因果关联的特征基因。对筛选出的基因进一步进行生存分析、免疫浸润分析及免疫中介机制解析。以核心特征基因为靶点，对中药小分子化合物库进行分子对接虚拟筛选，预测潜在活性成分。结果 在“炎-癌转化”进程中鉴定出30个持续上调的差异基因；MR分析从中确定5个与胃癌风险呈正向因果关联的核心驱动基因（S100A8、CXCR1、S100A9、S100A12、ZBED2）。其中，S100A12的高表达与患者不良预后显著相关，且与中性粒细胞浸润正相关、与B细胞及CD4⁺ T细胞浸润负相关。中介分析提示S100A12可能通过负向调控CD25⁺CD4⁺ T细胞介导促癌效应。全表型组MR分析显示靶向S100A12具有较好的潜在安全性。虚拟筛选获得多个与S100A12蛋白具有高亲和力的中药小分子化合物（C-箭毒碱、酸浆苦味素D等）。结论 S100A12是连接慢性炎症与胃癌发生的关键特征基因，其通过重塑免疫微环境驱动“炎-癌转化”。针对该靶点的中药小分子筛选为干预胃癌前病变提供了先导化合物线索。

Objective To screen characteristic genes driving gastric mucosal “inflammation-cancer transformation” based on multi-omics data and Mendelian randomization (MR) causal inference methods, and to explore potential traditional Chinese medicine (TCM) intervention strategies. Methods Transcriptomic data and large-scale genetic data were integrated. Differential analysis, MR analysis, and protein-protein interaction network construction were performed to screen characteristic genes causally associated with gastric cancer risk. Subsequently, survival analysis, immune infiltration analysis, and immune mediation mechanism analysis were conducted on the identified genes. Finally, molecular docking-based virtual screening was performed on a TCM small-molecule compound library using core characteristic genes as targets to predict potential active components. Results A total of 30 continuously up-regulated differentially expressed genes were identified during the “inflammation-cancer transformation” process. MR analysis identified five core characteristic genes (S100A8, CXCR1, S100A9, S100A12, ZBED2) showing positive causal associations with gastric cancer risk. Among them, high expression of S100A12 was significantly associated with poor prognosis in patients, positively correlated with neutrophil infiltration, and negatively correlated with B-cell and CD4⁺ T-cell infiltration. Mediation analysis revealed that S100A12 may promote carcinogenic effects by negatively regulating CD25⁺CD4⁺ T cells (primarily regulatory T cells). Phenome-wide MR analysis suggested good potential safety for targeting S100A12. Virtual screening identified several TCM small-molecule compounds with high binding affinity to the S100A12 protein (e.g., C-curarine, physalin D). Conclusion S100A12 is a key characteristic gene linking chronic inflammation to gastric cancer development, driving “inflammation-cancer transformation” by reshaping the immune microenvironment. The screening of TCM small molecules targeting this gene provides lead compound candidates for intervening in gastric precancerous lesions. The established research framework integrating multi-omics and reverse drug screening offers a new approach for the precise prevention and drug development of gastric cancer.

Q811.4;R285

国家自然科学基金青年基金项目（82505345）