PGM2 在脑胶质瘤中的表达及临床意义

目的 探讨磷酸葡糖变位酶 2（PGM2）在脑胶质瘤中的表达及临床意义。方法 提取中国胶质瘤基因组图谱计划（CGGA）数据库和癌症基因组图谱计划（TCGA）数据库中脑胶质瘤患者的临床数据和 mRNA 测序数据，将测序数据与临床数据相匹配后提取PGM2基因的表达情况，并剔除临床数据缺失的病例。根据受试者工作特征（receiver operating characteristic，ROC）曲线获得的临界值（Cut-off值）将病例分为PGM2高表达组和PGM2低表达组。采用 ROC 曲线下面积（AUC）和 Kaplan-Meier 生存曲线评价PGM2表达水平对胶质瘤患者总生存率的预测价值。采用单因素分析影响PGM2表达水平的胶质瘤临床病理因素。采用单因素、多因素 Cox 回归分析脑胶质瘤相关临床病理因素和PGM2表达与胶质瘤患者预后的关系。通过筛选共表达基因，进行基因本体（GO）富集分析和 KEGG 通路分析，探讨PGM2参与生物学过程以及调控的相关信号通路。采用 CIBERSORT 算法分析胶质瘤PGM2表达与免疫细胞浸润的关系。结果 CGGA-325 数据库剔除缺失值后余 273 例，TCGA 数据库剔除缺失值后余 560 例。PGM2表达量预测胶质瘤患者总生存率的 ROC 曲线分析显示，其在 CGGA-325 数据库中的AUC为 0.705（95%CI：0.640～0.769），Cut-off 值为 10.1；在 TCGA 数据库中的AUC为 0.739（95%CI：0.699～0.778），Cut-off 值为8.9。根据 Cut-off 值将病例分为PGM2高表达组和 PGM2 低表达组。Kaplan-Meier 生存分析显示，PGM2高表达组患者的总生存率较PGM2低表达组患者下降，差异有统计学意义（P＜ 0.001）。CGGA-325 和TCGA 数据库中，PGM2低表达与高表达组的胶质瘤级别、IDH分型以及 1p/19q 缺失情况比较，差异均有统计学意义（均P＜ 0.05）。多因素 Cox 回归分析显示，CGGA-325 数据库与 TCGA 数据库中胶质瘤共同独立预后因素是胶质瘤级别与1P/19q共缺失，PGM2表达在CGGA-325数据库中显示为独立预后因素（均P＜ 0.05）。GO 富集分析和 KEGG 通路分析显示，PGM2表达与内质网蛋白转运、Toll 样受体信号通路以及 I-κB 激酶 /NF-κB 信号通路相关。CIBERSORT 分析显示，PGM2低表达组与高表达组的初始 CD4+ T细胞、活化的记忆 CD4+ T 细胞、调节性 T 细胞、单核细胞、巨噬细胞（M0、M1、M2）、中性粒细胞等的免疫浸润程度比较，差异均有统计学意义（均P＜ 0.05）。结论 PGM2的高表达与胶质瘤较差的预后相关，可能作为判断胶质瘤患者预后的指标，未来有望成为胶质瘤潜在的治疗靶点。

Expression and clinical significance of PGM2 in glioma

Objective To explore the expression and clinical significance of phosphoglucomutase 2 （PGM2） in glioma. Methods Clinical data and mRNA sequencing data of glioma patients were collected from the Chinese Glioma Genome Atlas （CGGA） database and the Cancer Genome Atlas （TCGA） database. After matching the sequencing data with clinical data， the gene expression of PGM2 was extracted， and cases with missing clinical data were excluded. According to the cut-off value obtained from the receiver operating characteristic （ROC） curve， the cases were divided into PGM2 high expression group and PGM2 low expression group. The predictive value of PGM2 expression on the overall survival rate of glioma patients was evaluated using area under the ROC curve （AUC） and Kaplan Meier survival curve. Univariate analysis was used to explore the clinical and pathological factors affecting the expression level of PGM2 in gliomas. Univariate and multivariate Cox regression were used to analyze the relationship between prognosis in glioma patients and clinical and pathological factors related to glioma， as well as PGM2 expression. Gene Ontology （GO） enrichment analysis and KEGG pathway analysis were performed by screening co-expressed genes to explore the involvement of PGM2 in biological processes and related signaling pathways regulation. CIBERSORT algorithm was used to analyze the relationship between PGM2 expression in gliomas and immune cell infiltration. Results A total of 273 cases were left after the deletion of missing values from the CGGA-325 and 560 cases were left after the deletion of missing values from the TCGA. ROC curve analysis of PGM2 expression predicting the overall survival rate of glioma patients showed that the AUC in the CGGA-325 database was 0.705 ［95%CI（0.640， 0.769）］， with a cut-off value of 10.1， and the AUC in the TCGA database was 0.739 ［95%CI （0.699， 0.778）］， with a cut-off value of 8.9. According to the cut off value， the cases were divided into PGM2 high expression group and PGM2 low expression group. Kaplan Meier survival analysis showed that the overall survival rate of patients in PGM2 high expression group decreased compared to those in PGM2 low expression group， and the difference was statistically significant （P ＜ 0.001）. In the CGGA-325 and TCGA databases， there were statistically significant differences in glioma grade， IDH type， and 1p/19q deletion between PGM2low expression group and PGM2 high expression group （both P ＜ 0.05）. Multivariate Cox regression analysis showed that the common independent prognostic factors in the CGGA-325 and TCGA databases were glioma grade and co-deletion of 1P/19q， and the expression of PGM2 was an independent prognostic factor in the CGGA-325 database， and the difference was statistically significant （all P ＜ 0.05）. GO enrichment analysis and KEGG pathway analysis showed that PGM2 expression was associated with endoplasmic reticulum protein transportation， Toll-like receptor signaling pathway， and I-κB kinase/NF-κB signal pathway. CIBERSORT analysis showed that the initial CD4+ T cells， activated memory CD4+ T cells， regulatory T cells， monocytes， macrophages （M0， M1， M2）， and neutrophils in PGM2 low expression group and PGM2 high expression group were compared in terms of immune infiltration degree， and the differences were statistically significant （all P ＜ 0.05）. Conclusions High expression of PGM2 is associated with a poorer prognosis in glioma and may serve as a prognostic indicator for glioma patients and is expected to become a potential therapeutic target for glioma in the future.