基于内质网应激应答的胶质瘤风险模型构建

Establishment of endoplasmic reticulum stress response⁃based grading prognostic signature for the malignant phenotype of glioma

目的探索内质网应激应答(ERS)相关调控基因表达对胶质瘤(不包括胶质瘤母细胞瘤)病理分级、预后和恶性进展表型的预测价值。方法排除重复和无生存期的样本后,分别选取美国癌症基因组图谱和中国脑胶质瘤基因组图谱中胶质瘤临床样本测序数据作为训练集和验证集。通过差异基因富集和蛋白互作分析,构建与预后相关的ERS风险回归模型。ROC曲线分析、实时荧光定量PCR和免疫组化验证该模型对胶质瘤预后及恶性进展表型的预测价值。结果研究结果显示,7个ERS相关风险因子表达随胶质瘤级别上升而上调,并精准预测患者不良预后(在1、3和5年预测准确度均高于0.7),差异均有统计学意义。进一步验证显示,ERS风险基因与胶质瘤恶性表型标志物CD44呈正相关,与临床有利预后标志物GPR158呈负相关(P<0.05)。最后,临床样本的基因表达和免疫组化分析证实,ERS相关风险因子在较高级别胶质瘤中高表达,与CD44表达量呈正相关(P<0.05)。结论基于ERS应答的风险回归模型具备病理分级和预后的能力,并与胶质瘤的恶性进展表型呈正相关,可为胶质瘤精准靶向诊疗提供参考。

Objective This study aimed to explore the predictive value of endoplasmic reticulum stress response(ERS)-related regulatory gene expression for the pathological grading,prognosis,and malignant progres⁃sion phenotype of gliomas(excluding glioblastomas).Methods After excluding duplicate and non-survival samples,clinical sequencing data of glioma samples from the American Cancer Genome Atlas and the Chinese Brain Glioma Genome Atlas were selected as the training and validation sets.A prognosis-related ERS risk regression model was constructed through differential gene enrichment and protein interaction analysis.The predictive value of ERS Cox model for glioma prognosis and malignant progression phenotype was validated using ROC curve analysis,real-time fluorescence quantitative PCR,and immunohistochemistry.Results The study findings reveal that the expression of 7 ERS-related risk factors increases with the rise in glioma grade and accurately predicts unfavorable patient prognosis(with accuracies higher than 0.7 for predictions at 1,3,and 5 years),all of which are statistically significant.Further validation demonstrates a positive correlation between ERS risk genes and the glioma malignant phenotype marker CD44,as well as a negative correlation with the clinically favorable prognosis marker GPR158,both of which have statistical differences(P<0.05).Finally,gene expression and immunohistochemistry analysis of clinical samples confirm that ERS-related risk factors are highly expressed in higher-grade gliomas,positively correlated with CD44 expression,with statistical significance(P<0.05).Conclusion The preliminary results of the study suggest that the risk regression model based on ERS response exhibits the capacity to predict pathological grading and prognosis.Moreover,it demonstrates a positive correlation with the malignant progression phenotype of gliomas.These findings offer insights for precise and targeted diagnosis and treatment of gliomas.