基于肿瘤组织与血清miRNA表达数据的肝癌诊断模型构建

Construction of diagnosis signature of hepatocellular carcinoma based on tissue and circulating miRNA expression data

目的探究肝癌患者组织及血清中共同差异表达的miRNA,构建肝癌诊断模型。方法从美国国立生物技术信息中心数据库(Gene Expression Omnibus)下载芯片数据,从癌症基因组图谱(The Cancer Genome Atlas,TCGA)下载肝癌及正常样本的miRNA表达数据和mRNA表达数据,利用R软件进行差异分析;LASSO-logistic回归分析构建肝癌miRNA诊断模型;ROC曲线评价模型的预测效果;最后在40例肝癌患者及40例健康者血清中通过实时定量荧光PCR检测关键差异miRNA的表达。结果共筛选出36个同时在肝癌患者组织及肿瘤血清中均差异表达的miRNA,包括31个上调miRNA和5个下调miRNA。GO分析显示差异miRNA主要富集在胞质、胞核区域,参与包括转录因子活性调节、γ-干扰素等多种信号通路的转导。进一步经LASSO-logistic回归构建含有8个miRNA(hsa-miR-139-3p、 hsa-miR-10b-5p、 hsa-miR-500a-3p、 hsa-miR-224-5p、 hsa-miR-532-5p、hsa-miR-103a-3p、hsa-let-7c-3p、 hsa-miR-30c-1-3p)的诊断模型,ROC曲线显示该诊断模型在训练集、测试集以及GSE112264及GSE113486数据集中曲线下面积分别为0.992、0.997、0.953、0.916。收集临床样本检测得到肝癌患者组血清中关键差异miRNA的表达显著高于对照组(P <0.05)。结论由8个同时在肝癌患者血清及肿瘤组织中均差异表达的miRNA构成的诊断模型在鉴定肝癌与正常样本中具有良好的区分效果,对于肝癌患者的早期诊断提供借鉴意义。

Objective Detecting the differentially expressed miRNAs in the tissues and serum of patients afflicted with liver cancer and construction of a liver cancer diagnostic model. Methods The miRNA profiles of serum samples with hepatocellular carcinoma(HCC) GSE112264 and GSE113486 were downloaded from the National Center for Biotechnology Information Database(Gene Expression Omnibus). The miRNA expression data and m RNA expression profiles of tissues with liver cancer and normal samples were extracted from the Cancer Genome Atlas(TCGA). The differentially expressed miRNA and mRNA were analyzed using the R software. miRNA diagnostic signature was constructed using a LASSO-logistic regression analysis and was evaluated by the area under the curve(AUC) of receiver operating characteristic curve(ROC). Finally, the expression of key differential miRNA from the serum of 40 liver cancer patients and 40 healthy samples by RT-qPCR. Results A total of 36 miRNAs that were differentially expressed in hepatocellular carcinoma tissues as well as serum were screened, including 31 up-regulated miRNAs and 5 down-regulated miRNAs. GO analysis showed that the differential miRNAs were mainly located in the cytoplasm and nucleus, and enriched in the transduction of multiple signaling pathways including the IFN-gamma pathway. The differential miRNA was analyzed by using the LASSO-logistic regression to establish of a diagnostic model containing 8 differential miRNAs. The ROC curve displayed that the area under the curve(AUC) of the diagnostic model in the training set, test set, GSE112264 cohort, and GSE113486 cohort were 0.992, 0.997, 0.953, 0.916, respectively. The relative expression level of the key differential miRNA in the serum of patients with liver cancer cohort was significantly higher than that in the control group(P<0.05). Conclusion The diagnostic model composed of 8 miRNAs that were differentially expressed in the serum and tumor tissues of HCC patients had a good distinguishing effect in the identification of HCC and normal samples, providing reference significance for the early diagnosis of HCC patients.