An Immune Gene-Related Five-lncRNA Signature for to Predict Glioma Prognosis.
An Immune Gene-Related Five-lncRNA Signature for to Predict Glioma Prognosis.
复制标题
用于预测神经胶质瘤预后的免疫基因相关 5-lncRNA 特征
DOI：
10.3389/fgene.2020.612037
复制
发表时间：
2020
影响因子：
3.7
通讯作者：
Zhao S
中科院分区：
文献类型：
--
作者：
Wang X;Gao M;Ye J;Jiang Q;Yang Q;Zhang C;Wang S;Zhang J;Wang L;Wu J;Zhan H;Hou X;Han D;Zhao S
背景肿瘤免疫微环境与胶质瘤的恶性进展和治疗耐药性密切相关。长链非编码RNA（lncRNA）在这一过程中起着调节作用。我们研究了神经胶质瘤微环境中的病理机制和与lncRNAs相关的潜在免疫治疗抗性。方法下载神经胶质瘤患者的数据集，并对其进行系统聚类分析。接下来，我们分析了胶质瘤的免疫微环境、相关基因表达和患者生存率。分析共表达的lncRNA以产生lncRNA和免疫相关基因的模型。我们使用生存和考克斯回归分析模型。采用单变量、多变量、受试者工作特征（ROC）和主成分分析（PCA）等方法对模型的准确性进行验证。最后，GSEA被用来评估哪些功能和途径与差异基因相关。结果正常脑组织免疫功能处于低-中水平，胶质瘤免疫功能明显分为低-高三组。间质、免疫和估计分数随着免疫沿着增加，而肿瘤纯度降低。此外，人白细胞抗原（HLA）、程序性细胞死亡-1（PDL 1）、T细胞免疫球蛋白和粘蛋白结构域3（TIM-3）、B7-H3和细胞毒性T淋巴细胞相关抗原-4（CTLA 4）表达伴随免疫状态增加，并且患者预后恶化。筛选5种免疫基因相关lncRNA（AP001007.1、LBX-AS 1、MIR 155 HG、MAPT-AS 1和LINC 00515）构建风险模型。我们发现风险评分与患者预后和临床特征相关，并与PDL 1，TIM-3和B7-H3表达呈正相关。这些lncRNA可能通过补体-细胞因子受体相互作用、补体和凝血级联反应调节肿瘤免疫微环境，并可能促进高免疫组中的CD 8 + T细胞、调节性T细胞、M1巨噬细胞和浸润性中性粒细胞的活性。在体外，通过q-PCR和免疫组化（IHC）验证免疫相关lncRNA的异常表达以及危险评分与免疫相关指标（PDL 1、CTLA 4、CD 3、CD 8、iNOS）的关系。结论首次构建了免疫基因相关lncRNA风险模型。风险评分可能成为肿瘤免疫亚型的新生物标志物，为胶质瘤免疫治疗提供分子靶点。
Background The tumor immune microenvironment is closely related to the malignant progression and treatment resistance of glioma. Long non-coding RNA (lncRNA) plays a regulatory role in this process. We investigated the pathological mechanisms within the glioma microenvironment and potential immunotherapy resistance related to lncRNAs. Method We downloaded datasets derived from glioma patients and analyzed them by hierarchical clustering. Next, we analyzed the immune microenvironment of glioma, related gene expression, and patient survival. Coexpressed lncRNAs were analyzed to generate a model of lncRNAs and immune-related genes. We analyzed the model using survival and Cox regression. Then, univariate, multivariate, receiver operating characteristic (ROC), and principle component analysis (PCA) methods were used to verify the accuracy of the model. Finally, GSEA was used to evaluate which functions and pathways were associated with the differential genes. Results Normal brain tissue maintains a low-medium immune state, and gliomas are clearly divided into three groups (low to high immunity). The stromal, immune, and estimate scores increased along with immunity, while tumor purity decreased. Further, human leukocyte antigen (HLA), programmed cell death-1 (PDL1), T cell immunoglobulin and mucin domain 3 (TIM-3), B7-H3, and cytotoxic T lymphocyte-associated antigen-4 (CTLA4) expression increases concomitantly with immune state, and the patient prognosis worsens. Five immune gene-related lncRNAs (AP001007.1, LBX-AS1, MIR155HG, MAPT-AS1, and LINC00515) were screened to construct risk models. We found that risk scores are related to patient prognosis and clinical characteristics, and are positively correlated with PDL1, TIM-3, and B7-H3 expression. These lncRNAs may regulate the tumor immune microenvironment through cytokine–cytokine receptor interactions, complement, and coagulation cascades, and may promote CD8 + T cell, regulatory T cell, M1 macrophage, and infiltrating neutrophils activity in the high-immunity group. In vitro, the abnormal expression of immune-related lncRNAs and the relationship between risk scores and immune-related indicators (PDL1, CTLA4, CD3, CD8, iNOS) were verified by q-PCR and immunohistochemistry (IHC). Conclusion For the first time, we constructed immune gene-related lncRNA risk models. The risk score may be a new biomarker for tumor immune subtypes and provide molecular targets for glioma immunotherapy.