Department of Orthopedics, Renmin Hospital of Wuhan University, Wuchang, Wuhan, China.
Department of Urology, The First Affiliated Hospital of Hainan Medical University, Haikou, Hainan, China.
J Cell Physiol. 2021 Jan;236(1):294-308. doi: 10.1002/jcp.29842. Epub 2020 Jun 8.
Neuroblastoma (NBL) exists in a complex tumor-immune microenvironment. Immune cell infiltration and tumor-immune molecules play a critical role in tumor development and significantly impact the prognosis of patients. However, the molecular characteristics describing the NBL-immune interaction and their prognostic potential have yet to be investigated systematically. We first employed multiple machine learning algorithms, such as Gene Sets Enrichment Analysis and cell type identification by estimating relative subsets of RNA transcripts, to identify immunophenotypes and immunological characteristics in NBL patient data from public databases and then investigated the prognostic potential and regulatory networks of identified immune-related genes involved in the NBL-immune interaction. The immunity signature combining nine immunity genes was confirmed as more effective for individual risk stratification and survival outcome prediction in NBL patients than common clinical characteristics (area under the curve [AUC] = 0.819, C-index = 0.718, p < .001). A mechanistic exploration revealed the regulatory network of molecules involved in the NBL-immune interaction. These immune molecules were also discovered to possess a significant correlation with plasma cell infiltration, MYCN status, and the level of chemokines and macrophage-related molecules (p < .001). A nomogram was constructed based on the immune signature and clinical characteristics, which showed high potential for prognosis prediction (AUC = 0.856, C-index = 0.755, p < .001). We systematically elucidated the complex regulatory mechanisms and characteristics of the molecules involved in the NBL-immune interaction and their prognostic potential, which may have important implications for further understanding the molecular mechanism of the NBL-immune interaction and identifying high-risk NBL patients to guide clinical treatment.
神经母细胞瘤（NBL）存在于复杂的肿瘤免疫微环境中。免疫细胞浸润和肿瘤免疫分子在肿瘤的发展中起着关键作用，并显著影响患者的预后。然而，描述 NBL-免疫相互作用的分子特征及其预后潜力尚未得到系统研究。我们首先使用多种机器学习算法，如基因集富集分析和通过估计 RNA 转录本的相对子集来识别细胞类型，来识别公共数据库中 NBL 患者数据中的免疫表型和免疫学特征，然后研究了参与 NBL-免疫相互作用的鉴定免疫相关基因的预后潜力和调控网络。结合九个免疫基因的免疫特征被证实比常见的临床特征更有效地用于个体风险分层和 NBL 患者的生存结果预测（曲线下面积 [AUC] = 0.819，C 指数 = 0.718，p < .001）。机制探索揭示了参与 NBL-免疫相互作用的分子的调控网络。这些免疫分子也被发现与浆细胞浸润、MYCN 状态以及趋化因子和巨噬细胞相关分子的水平具有显著相关性（p < .001）。基于免疫特征和临床特征构建了列线图，显示出很高的预后预测潜力（AUC = 0.856，C 指数 = 0.755，p < .001）。我们系统地阐明了参与 NBL-免疫相互作用的分子的复杂调控机制和特征及其预后潜力，这可能对进一步理解 NBL-免疫相互作用的分子机制和识别高危 NBL 患者以指导临床治疗具有重要意义。
