DNMT1 Mediated CAHM Repression Promotes Glioma Invasion via SPAK/JNK Pathway.

Gliomas are the most common and fatal brain tumors worldwide. Abnormal DNA promoter methylation is an important mechanism for gene loss of tumor suppressors. A long non-coding RNA colorectal adenocarcinoma hypermethylated (CAHM) has been reported to be nearly deleted in glioblastomas (GBMs). Nevertheless, the roles of CAHM in gliomas remain unknown up to now. In the present study, 969 glioma samples downloaded from the CGGA and Gravendeel databases were included. We found that CAHM expression was correlated with glioma grades, molecular subtype, IDH mutation status, and 1q/19p codel status. In glioma cells, CAHM is hypermethylated by DNA methyltransferase1 (DNMT1) and the loss of CAHM expression could be reversed by 5-Aza-2'-deoxycytidine (5-Aza), a specific inhibitor of DNA methyltransferases. Besides, the expression of CAHM was negatively associated with overall survival in both primary and recurrent gliomas. Moreover, the result of Gene Ontology (GO) analysis suggested that CAHM participated in negatively regulating cell development, nervous system development, neurogenesis, and integrin-mediated signaling pathway. Overexpression of CAHM inhibited glioma cell proliferation, clone formation, and invasion. Further exploring results showed that CAHM overexpression suppressed glioma migration and invasion through SPAK/MAPK pathway. Collectively, this study disclosed that CAHM might be a suppressor in gliomas.

Identification of a mesenchymal-related signature associated with clinical prognosis in glioma.

Malignant glioma with a mesenchymal (MES) signature is characterized by shorter survival time due to aggressive dissemination and resistance to chemoradiotherapy. Here, this study used the TCGA database as the training set and the CGGA database as the testing set. Consensus clustering was performed on the two data sets, and it was found that two groups had distinguished prognostic and molecular features. Cox analysis and Lasso regression analysis were used to construct MES signature-based risk score model of glioma. Our results show that MES signature-based risk score model can be used to assess the prognosis of glioma. Three methods (ROC curve analyses, univariate Cox regression analysis, multivariate Cox regression analysis) were used to investigate the prognostic role of texture parameters. The result showed that the MES-related gene signature was proved to be an independent prognostic factor for glioma. Furthermore, functional analysis of the gene related to the risk signature showed that the genes sets were closely related to the malignant process of tumors. Finally, FCGR2A and EHD2 were selected for functional verification. Silencing these two genes inhibited the proliferation, migration and invasion of gliomas and reduced the expression of mesenchymal marker genes. Collectively, MES-related risk signature seems to provide a novel target for predicting the prognosis and treatment of glioma.