Sun Ruohan, Pan Yujun, Mu Long, Ma Yaguang, Shen Hong, Long Yu
Department of Neurology, the First Affiliated Hospital of Harbin Medical University, Harbin, China.
Department of Neurosurgery, the First Affiliated Hospital of Harbin Medical University, Harbin, China.
Front Genet. 2021 Oct 18;12:768930. doi: 10.3389/fgene.2021.768930. eCollection 2021.
Glioblastoma multiforme (GBM) is the most widely occurring brain malignancy. It is modulated by a variety of genes, and patients with GBM have a low survival ratio and an unsatisfactory treatment effect. The irregular regulation of RNA binding proteins (RBPs) is implicated in several malignant neoplasms and reported to exhibit an association with the occurrence and development of carcinoma. Thus, it is necessary to build a stable, multi-RBPs signature-originated model for GBM prognosis and treatment response prediction. Differentially expressed RBPs (DERBPs) were screened out based on the RBPs data of GBM and normal brain tissues from The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression Program (GTEx) datasets. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses on DERBPs were performed, followed by an analysis of the Protein-Protein Interaction network. Survival analysis of the DERBPs was conducted by univariate and multivariate Cox regression. Then, a risk score model was created on the basis of the gene signatures in various survival-associated RBPs, and its prognostic and predictive values were evaluated through Kaplan-Meier analysis and log-rank test. A nomogram on the basis of the hub RBPs signature was applied to estimate GBM patients' survival rates. Moreover, western blot was for the detection of the proteins. BICC1, GNL3L, and KHDRBS2 were considered as prognosis-associated hub RBPs and then were applied in the construction of a prognostic model. Poor survival results appeared in GBM patients with a high-risk score. The area under the time-dependent ROC curve of the prognostic model was 0.723 in TCGA and 0.707 in Chinese Glioma Genome Atlas (CGGA) cohorts, indicating a good prognostic model. What was more, the survival duration of the high-risk group receiving radiotherapy or temozolomide chemotherapy was shorter than that of the low-risk group. The nomogram showed a great discriminating capacity for GBM, and western blot experiments demonstrated that the proteins of these 3 RBPs had different expressions in GBM cells. The identified 3 hub RBPs-derived risk score is effective in the prediction of GBM prognosis and treatment response, and benefits to the treatment of GBM patients.
多形性胶质母细胞瘤（GBM）是最常见的脑恶性肿瘤。它受多种基因调控，GBM患者的生存率较低且治疗效果不理想。RNA结合蛋白（RBPs）的异常调控与多种恶性肿瘤有关，据报道与癌症的发生和发展有关。因此，有必要建立一个稳定的、基于多个RBPs特征的模型来预测GBM的预后和治疗反应。基于来自癌症基因组图谱（TCGA）和基因型-组织表达计划（GTEx）数据集的GBM和正常脑组织的RBPs数据，筛选出差异表达的RBPs（DERBPs）。对DERBPs进行基因本体论和京都基因与基因组百科全书分析，随后分析蛋白质-蛋白质相互作用网络。通过单变量和多变量Cox回归对DERBPs进行生存分析。然后，基于各种与生存相关的RBPs中的基因特征创建风险评分模型，并通过Kaplan-Meier分析和对数秩检验评估其预后和预测价值。应用基于核心RBPs特征的列线图来估计GBM患者的生存率。此外，通过蛋白质印迹法检测蛋白质。BICC1、GNL3L和KHDRBS2被认为是与预后相关的核心RBPs，随后被应用于构建预后模型。高风险评分的GBM患者生存结果较差。在TCGA队列中，预后模型的时间依赖性ROC曲线下面积为0.723，在中国胶质瘤基因组图谱（CGGA）队列中为0.707，表明这是一个良好的预后模型。此外，接受放疗或替莫唑胺化疗的高风险组的生存时间短于低风险组。列线图对GBM具有很强的鉴别能力，蛋白质印迹实验表明这3种RBPs的蛋白质在GBM细胞中具有不同的表达。所确定的由3个核心RBPs得出的风险评分在预测GBM预后和治疗反应方面是有效的，有助于GBM患者的治疗。
