Genome-wide methylation analysis of cerebrospinal fluid circulating tumor DNA: a new biomarker for recurrent glioblastoma [mRNA-Seq]

Background: Glioblastoma (GBM) is a common malignancy of the central nervous system (CNS) that is prone to recurrent and has a short survival period. Clinical biomarkers for the diagnosis and prognosis of recurrent GBM are lacking.Methods: We collected 4 cerebrospinal fluid (CSF) samples and 1 normal CSF sample from recurrent GBM, as well as paired tissue samples. Genome-wide methylation profiles of CSF circulating tumor DNA (ctDNA) and tissue mRNA transcription profiles were analyzed, and genes were screened by univariate Cox analysis, Lasso regression analysis, and multivariate Cox analysis by the Chinese Glioma Genome Atlas (CGGA) database. Data analysis and visualization were performed using R software, SPSS software, and Graphpad Prism.Results: By taking the intersection of differentially methylated regions (DMRs) and differentially expressed genes (DEGs), 892 genes were selected for Lasso regression analysis and multivariate Cox analysis, and 12 hub genes were finally screened to construct diagnostic and prognostic models. The diagnostic (AUC=0.982) and prognostic (5-years AUC=0.931) models based on the 12 hub genes had high accuracy.Conclusions: This study reveals that 12 hub genes in CSF ctDNA can diagnose and prognosticate recurrent GBM and provide new biomarkers for clinical research into the mechanisms of GBM recurrent. Overall design: In this study, we collected cerebrospinal fluid samples as well as matched tumor tissue samples from four patients with glioblastoma, in addition to one normal cerebrospinal fluid sample and one normal brain tissue sample as controls. We sequenced the methylation of four glioblastoma cerebrospinal fluid samples and one normal cerebrospinal fluid sample, and the other four glioblastoma tissue samples and one normal brain tissue sample were sequenced for gene expression profiling.

背景：胶质母细胞瘤（Glioblastoma, GBM）是中枢神经系统（Central Nervous System, CNS）常见恶性肿瘤，易复发且生存期较短，目前临床尚无针对复发型胶质母细胞瘤的诊断及预后相关生物标志物。
方法：本研究收集了复发型胶质母细胞瘤患者的4份脑脊液（Cerebrospinal Fluid, CSF）样本、1份正常脑脊液样本，以及配对的组织样本。对脑脊液循环肿瘤DNA（circulating tumor DNA, ctDNA）开展全基因组甲基化谱分析，对组织样本开展mRNA转录谱分析，并依托中国胶质瘤基因组图谱（Chinese Glioma Genome Atlas, CGGA）数据库，通过单变量Cox分析、Lasso回归分析及多变量Cox分析筛选候选基因。数据分析与可视化工作采用R软件、SPSS软件及GraphPad Prism完成。
结果：本研究将差异甲基化区域（Differentially Methylated Regions, DMRs）与差异表达基因（Differentially Expressed Genes, DEGs）取交集，筛选出892个候选基因用于Lasso回归分析及多变量Cox分析，最终得到12个核心基因以构建诊断与预后模型。基于该12个核心基因构建的诊断模型（曲线下面积（Area Under Curve, AUC）=0.982）与预后模型（5年随访AUC=0.931）均具有较高预测准确度。
结论：本研究表明，脑脊液ctDNA中的12个核心基因可用于复发型胶质母细胞瘤的诊断及预后评估，为胶质母细胞瘤复发机制的临床研究提供了新型生物标志物。
整体设计：本研究收集了4例胶质母细胞瘤患者的脑脊液样本及配对肿瘤组织样本，同时纳入1份正常脑脊液样本与1份正常脑组织样本作为对照。对4例胶质母细胞瘤患者的脑脊液样本及1份正常脑脊液样本进行甲基化测序，对另外4例胶质母细胞瘤组织样本及1份正常脑组织样本开展基因表达谱测序。