基于铁死亡相关基因骨肉瘤预后模型的构建与验证

Construction and validation of osteosarcoma prognostic model based on ferroptosis related gene

目的基于铁死亡相关基因(FRG)构建骨肉瘤(OS)预后模型,探讨FRG在OS中的表达及与患者预后的关系。方法通过生物信息学方法从UCSC Xena数据库中获取88例OS患者的转录组测序数据和其中85例患者的临床资料,与基因型-组织表达(GTEx)数据库中获取的396例正常骨组织样本合并,从FerrDb数据库中获取FRG,从合并后的数据中进行差异分析并提取差异表达的FRG。采用基因本体(GO)和京都基因与基因组百科全书(KEGG)富集分析探索OS中FRG的生物学功能。采用单因素Cox与Lasso回归模型筛选预后相关基因并构建预后预测模型。采用受试者工作特征(ROC)曲线分析预后模型的预测价值。采用单因素及多因素Cox回归模型分析OS患者预后的独立影响因素。采用实时荧光定量聚合酶链反应(qPCR)检测预后相关FRG在人成骨细胞hFOB1.19与OS细胞系U2OS、MG63中的表达情况。结果共获得57个差异表达的FRG。GO与KEGG富集分析发现这些差异基因主要富集在缺氧反应、线粒体外膜、铁离子结合等生物学反应和线粒体自噬、化学致癌-活性氧以及铁死亡等途径。应用单因素Cox及Lasso回归模型共筛选出9个预后相关的FRG来构建预后模型,分别为酰基辅酶A合成酶家族成员2(ACSF2)、芳香烃受体核转运因子样蛋白(ARNTL)、B细胞淋巴瘤/白血病-2相互作用蛋白3(BNIP3)、脂肪酸去饱和酶2(FADS2)、葡萄糖-6-磷酸脱氢酶(G6PD)、磷酸葡萄糖酸脱氢酶(PGD)、细胞因子信号转导抑制因子1(SOCS1)、转化生长因子β受体1(TGFBR1)、血管内皮生长因子A(VEGFA)。ROC曲线和生存分析证实这9个FRG构建的风险模型对OS患者的生存情况具有较好的预测价值。单因素和多因素分析结果显示,转移和风险评分均是OS患者预后的独立影响因素(P﹤0.01)。qPCR结果显示,U2OS和MG63细胞中FADS2 m RNA相对表达量均高于hFOB1.19细胞,差异均有统计学意义(P﹤0.05);MG63细胞中ACSF2 m RNA相对表达量高于hFOB1.19细胞,差异有统计学意义(P﹤0.05)。结论本研究成功构建了基于FRG的OS预后模型,发现ACSF2、ARNTL、G6PD、PGD、FADS2、SOCS1、BNIP3、TGFBR1、VEGFA等9个FRG能够作为OS患者的预后生物标志物,可为OS患者的临床治疗和预后评估提供参考。

Objective To construct the osteosarcoma(OS)prognostic model based on ferroptosis related gene(FRG)and explore the expression of FRG in OS and its relationship with prognosis of patients.Method Transcriptome sequencing data of 88 patients with OS and clinical data of 85 of them were obtained from UCSC Xena database by bioinformatics method,and combined with 396 normal bone tissue samples obtained from Genotype-Tissue Expression(GTEx)database.FRGs were obtained from the FerrDb database,and differential analysis was performed from the combined data to extract differentially expressed FRGs.Gene Ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)enrichment analysis were used to explore the biological function of FRG in OS.Univariate Cox and Lasso regression models were used to screen the prognostic genes and to construct the prognosis predicting model.The receiver operating characteristic(ROC)curve was used to analyze the predictive value of prognostic model.Univariate and multivariate Cox regression models were used to analyze the independent influencing factors for prognosis of OS.The expression of FRG in human osteoblast cell line hFOB1.19 and OS cell lines U2OS and MG63 was detected by real-time quantitative polymerase chain reaction(qPCR).Result A total of 57 differentially expressed FRGs were obtained.GO and KEGG enrichment analysis showed that these differentially expressed genes were mainly enriched in biological reactions of hypoxia response,mitochondrial outer membrane,iron ion binding,and pathways of mitochondrial autophagy,chemical carcinogenic reactive-oxygen and ferroptosis.A total of 9 prognostic related FRGs were screened using univariate Cox and Lasso regression models to construct prognostic models,which were acyl-CoA synthetase family member 2(ACSF2),aryl hydrocarbon receptor nuclear translocator like(ARNTL),B-cell lymphoma/leukemia-2 interacting protein 3(BNIP3),fatty acid desaturase 2(FADS2),glucose-6-phosphate dehydrogenase(G6PD),phosphogluconate dehydrogenase(PGD),suppressor of cytokine signaling 1(SOCS1),transforming growth factor beta receptor 1(TGFBR1),and vascular endothelial growth factor A(VEGFA).The ROC curve and survival analysis showed that the risk models based on these 9 FRGs had good prediction value for survival of OS patients.Univariate and multivariate analysis showed that metastasis and risk score were independent influencing factors for prognosis of OS patients(P<0.01).qPCR results showed that the relative expression of FADS2 mRNA in U2OS and MG63 cells were higher than that in hFOB1.19 cell(P<0.05).The relative expression of ACSF2 mRNA in MG63 cell was higher than that in hFOB1.19 cell(P<0.05).Conclusion The OS prognostic model based on FRG is successfully constructed in this study,which finds that 9 FRGs,including ACSF2,ARNTL,G6PD,PGD,FADS2,SOCS1,BNIP3,TGFBR1 and VEGFA,can be used as prognostic biomarkers for patients with OS.It can provide reference for the clinical treatment and prognosis evaluation of patients with OS.