基于生物信息学筛选肝细胞癌血管侵袭特征基因和预后标志物

Screening characteristic genes and prognostic biomarkers of vascular invasion in hepatocellular carcinoma based on bioinformatics

目的:挖掘肝细胞癌(HCC)血管侵袭相关的特征基因并筛选其预后标志物。方法:使用R软件limma包筛选出癌症基因组图谱数据库(TCGA)、基因表达数据库系列(GSE19977和GSE20017)中HCC血管侵袭相关差异表达基因,对其进行GO功能富集和KEGG信号通路分析。采用最小绝对收缩选择算子(LASSO)和支持向量机递归特征消除(SVM-RFE)筛选重叠的HCC血管侵袭特征基因。采用单因素Cox回归分析筛选出的特征基因与患者预后的关系。结果:3个数据库中上下调表达一致的差异表达基因有517个。GO富集结果显示3个数据库交集的差异基因主要富集于线粒体基质、核糖体,血红素结合、氧化还原酶活性等。KEGG分析结果显示交集的差异表达基因主要富集在新陈代谢、过氧化物酶体增殖物激活受体(PPAR)、补体和凝血级联等信号通路。两种算法共筛选出10个重叠的特征基因:爱帕琳肽受体(APLNR)、残疾基因同源物1(DAB1)、分泌磷蛋白2(SPP2)、甲状腺激素应答蛋白(THRSP)、溶质载体家族22成员7(SLC22A7)、溶质载体家族16成员2(SLC16A2)、内皮细胞特异性分子1(ESM1)、外泌体成分8(EXOSC8)、同源盒基因D10(HOXD10)和TB(POZ)结构域6蛋白(KBTBD6)。APLNR、SPP2、SLC22A7低表达[HR(95%CI)分别为0.85(0.71~1.02),0.87(0.80~0.95),0.89(0.82~0.97)]及HOXD10高表达[HR(95%CI)为3.26(1.84-5.77)]是HCC患者预后的危险因素。结论:APLNR、DAB1、SPP2、THRSP、SLC22A7、SLC16A2、ESM1、EXOSC8、HOXD10和KBTBD6可作为HCC血管侵袭特征基因,其中APLNR、SPP2、SLC22A7及HOXD10与HCC患者预后有关。

Aim:To explore the characteristic genes related to vascular invasion in hepatocellular carcinoma(HCC)and screen the prognostic biomarkers of HCC.Methods:The R software limma package was utilized to identify the differentially expressed genes related to HCC vascular invasion in the Cancer Genome Atlas(TCGA)and gene expression database series(GSE19977 and GSE20017)datasets,and GO function enrichment and KEGG signaling pathway analysis were performed.Two algorithms,least absolute shrinkage and selector operation(LASSO)and support vector machine-recursive feature elimination(SVM-RFE),were used to select overlapping genes.Univariate Cox regression analyses were performed to obtain the prognostic value of overlapping genes.Results:There were 517 differential genes with consistent up-and down-regulation in the three datasets.The GO enrichment results showed that the overlapping differential genes were enriched in the mitochondrial matrix and ribosomes,heme binding and oxidoreductase activity.KEGG signaling pathway analysis showed that the differentially expressed genes mainly enriched in signaling pathways such as metabolism,peroxidase proliferatoractivated receptor(PPAR),complement and coagulation cascade.A total of 10 overlapping characteristic genes were screened by the two algorithms:apelin receptor(APLNR),disabled homolog 1(DAB1),secreted phosphoprotein 2(SPP2),thyroid hormone responsiveprotein(THRSP),solute carrier family 22 member 7(SLC22A7),solute carrier family 16 member 2(SLC16A2),endothelial cell specific molecule 1(ESM1),exosome component 8(EXOSC8),homeobox D10(HOXD10),kelch repeat and BTB domain-containing protein 6(KBTBD6).Low expressions of APLNR,SPP2,SLC22A7[HR(95%CI)0.85(0.71-1.02),0.87(0.80-0.95),0.89(0.82-0.97)]and high expression of HOXD10[HR(95%CI)3.26(1.84-5.77)]were risk factors for the prognosis of HCC patients.Conclusion:APLNR,DAB1,SPP2,THRSP,SLC22A7,SLC16A2,ESM1,EXOSC8,HOXD10 and KBTBD6 may be the characteristic genes related to HCC vascular invasion,among which APLNR,SPP2,SLC22A7 and HOXD10 are associated with the prognosis of HCC patients.