Zhang PeiPei, Ma KeXin, Ke XiaoFei, Liu Liu, Li Ying, Liu YaJuan, Wang YouJun
Department of Pediatrics, The Fifth Affiliated Hospital of Zhengzhou University, Zhengzhou University, Zhengzhou, China.
Front Genet. 2021 Oct 20;12:685646. doi: 10.3389/fgene.2021.685646. eCollection 2021.
Neuroblastoma (NBL) originating from the sympathetic nervous system is the most prevalent solid tumor in infancy. Although there is sufficient variability in prognosis among different age pyramids, age-related gene expression profiles and biomarkers remain poorly explored. The present study aimed to construct a signature based on differentially expressed genes (DEGs) between two age groups in NBL. Univariate Cox regression, multivariate Cox regression, and LASSO analyses were used to identify the optimal prognostic factors. The prediction ability of the model was assessed using the receiver operating characteristic (ROC) curve and C-index. Functional enrichment analysis was performed using the Kyoto Encyclopedia of Genes and Genomes and gene ontology databases. A total of 1,160 DEGs were identified between the two groups, and 204 DEGs impacted the survival of NBL. Functional enrichment analysis revealed that the DEGs were involved in retinol metabolism, cholesterol metabolism, and glycolysis/gluconeogenesis pathways. Five RNAs, namely F8A3, PDF, ANKRD24, FAXDC2, and TMEM160 were recruited into the signature. They were correlated with COG risk classification, INSS stage, and histology. MYCN amplification was linked to FAXDC2, TMEM160, PDF, and F8A3. The expression levels of ANKRD24, PDF, and TMEM160 were lower in the hyperdiploid groups. Only FAXDC2 levels were different in the different MKI grades. The ROC curve showed that the five-RNA-based signatures effectively predicted the OS of NBL (3-years AUC = 0.791, 5-years AUC = 0.816) in the TARGET cohort. The predictive capability was also validated by the GSE49711 cohort (3-years AUC = 0.851, 5-years AUC = 0.848). The C-index in the TARGET and GSE49711 cohorts was 0.749 and 0.809, respectively. The potential mechanisms of the five RNAs were also explored gene set enrichment analysis, and candidate drugs targeting the five genes, including dabrafenib, vemurafenib, and bafetinib, were screened. In conclusion, we constructed a five-RNA-based signature to predict the survival of NBL and screened candidate agents against NBL.
神经母细胞瘤（NBL）起源于交感神经系统，是婴儿期最常见的实体瘤。尽管不同年龄组的预后存在足够的差异，但与年龄相关的基因表达谱和生物标志物仍未得到充分探索。本研究旨在基于NBL两个年龄组之间的差异表达基因（DEG）构建一个特征。使用单变量Cox回归、多变量Cox回归和LASSO分析来确定最佳预后因素。使用受试者工作特征（ROC）曲线和C指数评估模型的预测能力。使用京都基因与基因组百科全书和基因本体数据库进行功能富集分析。两组之间共鉴定出1160个DEG，其中204个DEG影响NBL的生存。功能富集分析表明，这些DEG参与视黄醇代谢、胆固醇代谢和糖酵解/糖异生途径。五个RNA，即F8A3、PDF、ANKRD24、FAXDC2和TMEM160被纳入该特征。它们与COG风险分类、INSS分期和组织学相关。MYCN扩增与FAXDC2、TMEM160、PDF和F8A3相关。在超二倍体组中，ANKRD24、PDF和TMEM160的表达水平较低。在不同的MKI分级中，只有FAXDC2水平不同。ROC曲线显示，基于五个RNA的特征有效地预测了TARGET队列中NBL的总生存期（3年AUC = 0.791，5年AUC = 0.816）。GSE49711队列也验证了该预测能力（3年AUC = 0.851，5年AUC = 0.848）。TARGET和GSE49711队列中的C指数分别为0.749和0.809。还通过基因集富集分析探索了这五个RNA的潜在机制，并筛选了靶向这五个基因的候选药物，包括达拉非尼、维莫非尼和巴非替尼。总之，我们构建了一个基于五个RNA的特征来预测NBL的生存，并筛选了抗NBL的候选药物。
