Identification And Functional Analysis Of DNA Methylation Signatures Derived From Glioma Stem-like Cells For Radiotherapy In Glioblastoma

Background:Glioblastoma(GBM)is the most common and aggressive primary brain cancer with the five-year survival rate of 9%.However,the pathogenesis and underlying molecular mechanisms of most cases are nubilous so far.Surgical excision is the most common measure in GBM clinical treatment.The operation is to take maximum dissection after taking a full account for the functional regions.As one of the typical treatment procedures in GBM,radiotherapy can prolong the survival but fail to prevent the tumor recurrence.The frequent occurrence of radioresistance should be responsible for the high rate of recurrence and mortality.Aberrant DNA methylation can induce many disorders including cancer.A large number of researches indicate that DNA methylation has a tight correlation with the prognosis of GBM.The G-CIMP,Glioma CpG Island Methylation Phenotype,stands for a number of CpG islands hyper-methylated in GBM.Patients with the G-CIMP molecular feature are younger at the time of diagnosis and experience significantly improved outcomes.Temozolomide(TMZ)is one of the most common agents used in chemotherapy for GBM.The therapeutic effect of TMZ depends on its ability to alkylate the DNA molecular and lead to the cell death.Proteins encoded by MGMT can repair lesions on DNA molecules.Hyper-methylation of the promotor region of MGMT has been shown to associated with better prognosis in GBM.In addition,researchers classified the gliomas into low grade gliomas and high grade gliomas clearly based on the methylome profile by unsupervised clustering.The glioma stem-like cells(GSCs)have the ability of tumor initiating cells of infinite proliferation and tumorigenesis.And the GSCs also possess the unabridged genomic characteristics of the GBM genome.The GSCs are proven to play an important role in the resistance to radiation.The GSC hypothesis based on the numerous researches suggests that the small number of GSCs are the tumor initial cells of GBM and play vital roles in the formation,growth and treatment of GBM.Methods:Taking advantage of the robust GSC model,we identify the differentially methylated loci between the radio-sensitive and radio-resistant GSCs:a)get beta values from the raw methylation array data of GSCs,b)calculate the differences between the mean values of radio-sensitive and radio-resistant GSCs,c)generate the cumulative distribution(p values)of the differences.In this study,we set the threshold to 0.001.Probes with p value less than 0.001 are defined as hyper-methylated in radio-sensitive samples while probes with p value greater than 0.999 are defined as hyper-methylated in radio-resistant samples.The TCGA GBM samples are classified by R package ssGSEA according to their enrichment scores(ES)derived from the corresponding methylation profiles and the differentially methylated probes.Survival analysis is conducted by R package survival depending on the clinical information of the classified GBM samples.The gene function enrichment analysis is done using GSEA.The correlations between the methylation levels and the expression abundances are denoted by Pearson correlation coefficients.Results:We reveal that the methylation pattern differs from radio-sensitive GSCs to radio-resistant GSCs,The radio-resistant GSCs have higher methylation levels.The differentially methylated regions contain genes such as ERBB2,WWTR1,HIST1H3G,HMG20B,and so on.These genes are involved in critical biological processes including cell proliferation,cell cycle,DNA repair,chromosome stability,epithelial-mesenchymal transition,and embryonic stem cell self-renewal.The results of survival analysis suggest that the radio-sensitive patients have longer survival time and can benefit from the radiotherapy.Dysregulated genes between radio-sensitive and radio-resistant samples are mainly involved in apoptotic signaling pathway,response to heat,cell cycle and response to oxidative stress.The methylation levels of the 28 common probes differentially methylated in both the GSCs and TCGA datasets have moderate correlations with the expression abundances of certain dysregulated genes.Conclusions:The methylation signatures derived from GSCs can be used to classify GBM samples into radio-sensitive or radio-resistant group and to predict prognosis effectively.The discovery also validates our assumption based on the GSC hypothesis that GBM patients harboring the similar methylation signatures with radio-sensitive GSCs would experience longer survival time.The differentially methylated or dysregulated genes are involved in important biological processes and play crucial roles in the radiation resistance in GBM.