Galactophore Department, The Second Clinical Medical College, Yangtze University, Jingzhou, China.
Drug Dev Res. 2024 Jun;85(4):e22223. doi: 10.1002/ddr.22223.
Breast cancer (BC) is a common cancer for women. This study aims to construct a prognostic risk model of BC and identify prognostic biomarkers through machine learning approaches, and clarify the mechanism by which linalool exerts tumor-suppressive function. Three mRNA microarray/RNA sequencing data sets (GSE25055, GSE103091, and TCGA-BRCA) were obtained from Gene Expression Omnibus database and The Cancer Genome Atlas database, and prognostic genes were obtained by univariate COX analysis. Multiple machine learning methods were used to screen core genes and construct prognostic risk models. The enrichment analysis of crucial genes was analyzed using the DAVID database. UALCAN, human protein atlas, geneMANIA, and LinkedOmics databases were used to analyze gene expression and co-expressed genes. Molecular docking and molecular dynamics simulation was applied to verify the binding affinity between linalool and phosphoglycerate kinase 1 (PGK1). Cell counting kit 8 (CCK-8, Edu, transwell, flow cytometry, and Western blot assay were used to analyze cell activity, apoptosis, cell cycle and protein expression. Eight prognostic genes were obtained by bioinformatics analysis and machine learning, and prognostic risk models were constructed. This model could well predict the prognosis of patients, and the risk score could be used as an independent risk factor for BC. Overall survival (OS) and immune cell infiltration characteristics were distinct between high and low risk groups. PGK1 was highly expressed in BC and the OS of patients with high PGK1 expression was shorter. PGK1 was related to cell cycle and PPAR signaling pathway. Linalool and PGK1 had good binding activity, and linalool could inhibit the viability, proliferation, migration, and invasion of BC cells, promote cell apoptosis, and induce G0/G1 arrest. In addition, linalool can promote PPARγ protein expression and inhibit PGK1 expression. Machine learning and molecular docking were promising for exploration of new drug targets for BC, and linalool exerts tumor-suppressive effects in BC by inhibiting PGK1 expression and activating PPAR signaling pathway.
乳腺癌（BC）是女性常见的癌症。本研究旨在通过机器学习方法构建 BC 的预后风险模型并确定预后生物标志物，并阐明芳樟醇发挥肿瘤抑制功能的机制。从基因表达综合数据库和癌症基因组图谱数据库中获得了三个 mRNA 微阵列/RNA 测序数据集（GSE25055、GSE103091 和 TCGA-BRCA），并通过单变量 COX 分析获得了预后基因。使用多种机器学习方法筛选核心基因并构建预后风险模型。使用 DAVID 数据库分析关键基因的富集分析。使用 UALCAN、人类蛋白质图谱、geneMANIA 和 LinkedOmics 数据库分析基因表达和共表达基因。应用分子对接和分子动力学模拟验证芳樟醇与磷酸甘油酸激酶 1（PGK1）的结合亲和力。细胞计数试剂盒 8（CCK-8、Edu、transwell、流式细胞术和 Western blot 分析用于分析细胞活性、细胞凋亡、细胞周期和蛋白质表达。通过生物信息学分析和机器学习获得了 8 个预后基因，并构建了预后风险模型。该模型可以很好地预测患者的预后，风险评分可以作为 BC 的独立危险因素。高低风险组之间的总生存期（OS）和免疫细胞浸润特征明显不同。PGK1 在 BC 中高表达，PGK1 高表达患者的 OS 更短。PGK1 与细胞周期和 PPAR 信号通路有关。芳樟醇与 PGK1 具有良好的结合活性，芳樟醇可抑制 BC 细胞的活力、增殖、迁移和侵袭，促进细胞凋亡，并诱导 G0/G1 期阻滞。此外，芳樟醇可以促进 PPARγ 蛋白表达并抑制 PGK1 表达。机器学习和分子对接有望成为探索 BC 新药物靶点的方法，芳樟醇通过抑制 PGK1 表达和激活 PPAR 信号通路发挥抑制 BC 的作用。
