乳腺癌脑转移相关基因的生物信息学分析

Bioinformatic Analyses of Key Genes in Breast Cancer Brain Metastases

目的:通过生物信息学挖掘乳腺癌脑转移中关键基因,为乳腺癌脑转移的预防和治疗提供理论依据。方法:从GEO数据库下载乳腺癌脑转移相关数据集(GSE52604、GSE26338、GSE43837、GSE100534)基因表达谱。GEO2R分析乳腺癌原发肿瘤与脑转移瘤差异表达基因,韦恩图筛选4个数据集中共表达差异基因(DEGs),UALCAN和乳腺癌基因表达Miner分析差异基因与肿瘤分期、肿瘤病理分级的相关性,Kaplan-Meier与cBioPortal分析差异基因对预后的预测价值,TIMER分析差异基因与肿瘤免疫细胞浸润的相关性。结果:血小板源性生长因子受体α(PDGFRA)、皮连蛋白(DPT)和软骨间层蛋白(CILP)为共表达下调差异基因,PDGFRA和CILP与肿瘤分期显著相关(均P<0.05),DPT和CILP与肿瘤病理分级明显相关(均P<0.001)。预后分析中,DPT和CILP下调与乳腺癌脑转移的总生存率(OS)显著相关(均P<0.05),PDGFRA、DPT和CILP共突变与乳腺癌脑转移的OS显著相关(均P<0.05),PDGFRA、DPT和CILP与肿瘤中CD4+T细胞、CD8+T细胞、巨噬细胞、中性粒细胞和树突状细胞的浸润明显相关(均P<0.01)。结论:PDGFRA、DPT和CILP差异基因与免疫细胞浸润相关可能是乳腺癌脑转移的分子机制,可以作为乳腺癌脑转移的预测和预后指标。

Objective: To explore the key genes in breast cancer brain metastases by bioinformatic analyses, providing theoretical foundation for prevention and potentially gene therapy in breast cancer brain metastases.Methods: Gene expression profiles of breast cancer brain metastasis datasets (GSE52604, GSE26338, GSE43837, GSE100534) were downloaded from the GEO database.GEO2R was used to analyze differentially expressed genes in primary tumors and brain metastases of breast cancer.Four datasets were screened by venn diagram to identify coexpressed differentially expressed genes (DEGs).UALCAN and breast cancer gene expression Miner were used to analyze correlation of DEGs and tumor grading, pathological grading.Kaplan-Meier and cBioPortal were used to analyze prognostic value of DEGs.The correlation between DEGs and cancer immune cell infiltration was investigated via TIMER.Results: Platelet-derived growth factor receptor alpha (PDGFRA), dermatopontin (DPT) and cartilage intermediate layer protein (CILP) were down-regulated in breast cancer brain metastases.PDGFRA and CILP were significantly associated with tumor grade (All P <0.05).DPT and CILP were significantly associated with tumor pathological grade (All P < 0.001 ).DPT and CILP down-regulation was associated with overall survival (OS) of breast cancer brain metastasis in prognosis analysis (All P <0.05).PDGFRA, DPT, and CILP co-mutation were significantly associated with OS in breast cancer brain metastases (All P <0.05).PDGFRA, DPT and CILP expression were positively correlated with infiltrating levels of CD4 +T and CD8 +T cells, macrophages, neutrophils, and dendritic cells (DCs)(All P <0.01).Conclusion: Differentially expressed PDGFRA, DPT and CILP may be related to immune cell infiltration, which would be the potential molecular mechanism of breast cancer brain metastasis, and can be used as a predictor and prognosis indicator for brain metastasis of breast cancer.