Table_2_Development of a novel glycolysis-related genes signature for isocitrate dehydrogenase 1-associated glioblastoma multiforme.docx

BackgroundThe significant difference in prognosis between IDH1 wild-type and IDH1 mutant glioblastoma multiforme (GBM) may be attributed to their metabolic discrepancies. Hence, we try to construct a prognostic signature based on glycolysis-related genes (GRGs) for IDH1-associated GBM and further investigate its relationships with immunity.
MethodsDifferentially expressed GRGs between IDH1 wild-type and IDH1 mutant GBM were screened based on the TCGA database and the Molecular Signature Database (MSigDB). Consensus Cluster Plus analysis and KEGG pathway analyses were used to establish a new GRGs set. WGCNA, univariate Cox, and LASSO regression analyses were then performed to construct the prognostic signature. Then, we evaluated association of the prognostic signature with patients’ survival, clinical characteristics, tumor immunogenicity, immune infiltration, and validated one hub gene.
Results956 differentially expressed genes (DEGs) between IDH1 wild-type and mutant GBM were screened out and six key prognostically related GRGs were rigorously selected to construct a prognostic signature. Further evaluation and validation showed that the signature independently predicted GBM patients’ prognosis with moderate accuracy. In addition, the prognostic signature was also significantly correlated with clinical traits (sex and MGMT promoter status), tumor immunogenicity (mRNAsi, EREG-mRNAsi and HRD-TAI), and immune infiltration (stemness index, immune cells infiltration, immune score, and gene mutation). Among six key prognostically related GRGs, CLEC5A was selected and validated to potentially play oncogenic roles in GBM.
ConclusionConstruction of GRGs prognostic signature and identification of close correlation between the signature and immune landscape would suggest its potential applicability in immunotherapy of GBM in the future.

背景：IDH1野生型与IDH1突变型多形性胶质母细胞瘤（glioblastoma multiforme, GBM）的预后存在显著差异，这可能与二者的代谢差异相关。因此，本研究拟基于糖酵解相关基因（glycolysis-related genes, GRGs）构建IDH1相关GBM的预后特征，并进一步探究其与免疫的关联。
方法：本研究基于癌症基因组图谱（The Cancer Genome Atlas, TCGA）数据库与分子特征数据库（Molecular Signature Database, MSigDB），筛选IDH1野生型与突变型GBM之间差异表达的GRGs。通过一致性聚类Plus分析（Consensus Cluster Plus analysis）与京都基因与基因组百科全书（Kyoto Encyclopedia of Genes and Genomes, KEGG）通路分析构建新的GRGs集。随后采用加权基因共表达网络分析（Weighted Gene Co-expression Network Analysis, WGCNA）、单变量Cox回归与最小绝对收缩和选择算子（Least Absolute Shrinkage and Selection Operator, LASSO）回归分析构建预后特征。进一步评估该预后特征与患者生存情况、临床特征、肿瘤免疫原性、免疫浸润的关联，并对1个核心基因进行验证。
结果：本研究共筛选出IDH1野生型与突变型GBM间的956个差异表达基因（differentially expressed genes, DEGs），并严格筛选出6个与预后相关的关键GRGs以构建预后特征。进一步评估与验证结果显示，该特征可独立预测GBM患者的预后，且具有中等准确度。此外，该预后特征还与多项临床特征（性别、O6-甲基鸟嘌呤-DNA甲基转移酶（O6-methylguanine-DNA methyltransferase, MGMT）启动子状态）、肿瘤免疫原性（mRNAsi、EREG-mRNAsi及HRD-TAI）及免疫浸润（干细胞指数、免疫细胞浸润、免疫评分与基因突变）显著相关。在6个预后相关关键GRGs中，CLEC5A被筛选并验证为可能在GBM中发挥致癌作用。
结论：构建GRGs预后特征并明确其与免疫微环境的密切关联，提示该特征未来有望应用于GBM的免疫治疗。