The Investigation Of Mechanism And Application Value Of MiRNA In Brain Tumors By Network Analysis

Background: Intracranial tumor known as intracranial neoplasm or brain tumor,occurs in the central nervous system.There are two main types of brain tumors: malignant intracranial tumors and benign intracranial tumors.According to the overview of WHO central nervous system classification in 2016,the CNS tumor include: glioma;ependymal tumor;choroid plexus tumours;neuronal and mixed neuronal-glial tumours;pineal region tumor;Embryonal tumour;medulloblastomas;cranial and paraspinal nerve tumor;meningioma;mesenchymal,non-meningothelial tumor;melanocytic tumor;lymphoma;Histiocytic tumours;germ cell tumour;sellar region tumor etc.According to the source of tissue cells,glioma can be divided to diffuse astrocytes,oligodendroglioma,other astrocytic tumors,ependymal tumors,other gliomas,Neuronal and mixed neuronal-glial tumours and other different types.The latest 2016 central nervous system tumor classification established on the new molecular concept of tumor diagnosis.Combined with genetic characteristics,adjust the diagnosis of partial glioma and medulloblastoma.Mutation of IDH,IDH-wildtype,H3 K27M-mutation,IDH-mutation and 1p19q-codeleted were added to the glioma grade criteria.In the traditional histological classification of medulloblastoma include typical medulloblastoma,multifocal/nodular proliferative medulloblastoma,desmoplastic/nodular medulloblastoma,large cell/anaplastic medulloblastoma.The WNT-activated,SHH-activated and TP53 mutant,SHH-activated and TP53 wild-type,non-WNT / non-SHH activators were used to study the molecular characteristics of the medulloblastoma classification and diagnosis.Grade III glioma as a central nervous system malignant brain tumors,mainly divided into anaplastic astrocytoma,IDH-mutant,anaplastic oligodendroglioma,IDH-mutation and 1p19q-codeleted type,pleomorphic xanthoastrocytoma and anaplastic ganglion glioma.MiRNA is a small non-coding RNA molecule(containing about 22 nucleotides),can be found in plants,animals and some viruses.MiRNAs may regulate the gene expression by silencing RNA and influence post-transcriptional gene expression.While the majority of miRNAs are located in the cell,some mi RNAs,usually known as circulating miRNAs or extracellular miRNAs,have also been found in extracellular environment,including various biological fluids,such as cerebrospinal fluid(CSF).Many researches have revealed that abnormal expressed mi RNA in tumor tissues and biological fluids.Besides,some studies concluded that mi RNAs can be biomarkers for brain tumors.Objective:(1)To better understand the tumorigenesis,we integrated the RNAseq and clinical data to build a co-expression scale-free network of gliomas by Weight gene co-expression network analysis(WGCNA)and identified specific gene modules and valuable hub genes associated with gliomagenesis for grade III gliomas.(2)Combined the clinical data of those patients,identify specific genes signature intensively related with clinical prognosis for grade III gliomas patient.(3)By mi RNA data array analysis,figured out different expression microRNAs in the medulloblastoma and clarified the networks and hub genes of tumorigenesis with Ingenuity Pathway Analysis(IPA).(4)Based on the bioinformation analysis,defined differentially expressed set of microRNAs in cerebrospinal fluid(CSF),which can differentiate CNS tumor,such as meningiomas,low grade gliomas,glioblastomas,metastases,medulloblastomas and lymphomas.In addition,validated those biomarkers by fluorescent quantitative PCR and investigated the application value of mi RNA in brain tumor clinical preoperative diagnosis.Methods:(1)We used the TCGA-Assembler download level-3 RNASeqV2 gene expression data,miRNA-seq data of samples and clinical information of those patients.We deleted expressed data close to zero,and selected round numbers of all array.Compared the normal group with grade III gliomas,we used the “DESeq” package in R software(3.3.0)to identify the differently expressed genes(DEG)and mi RNAs with foldChange>2.0,and adjusted P-value<0.05.The WGCNA was employed to find the co-expression modules and visualized it with Cytoscape software.The function of different module genes were annotated by gene ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG).(2)In order to identify prognostic mRNA and miRNA signature,combining the clinical data of those patients in TCGA,we established the life curves of those samples with those hub genes or some different expressed genes.(3)The data was downloaded from NCBI GEO database(GSE42657).We downloaded the data Series Matrix File and performed a log2 transformation.The data includes 7 control samples(consist of 3 cerebellum and 4 frontal lobe control)and 9 medulloblastoma samples.All samples data was normalized by using limma package in R software.The differentially expressed genes(DEGs)were obtained using Linear models and empirical bayes methods for assessing differential expression in microarray experiments in Limma package.Finally,we write the heatmap of different group to visualize the different expression genes and clustered the corresponding group with the different expressed mi RNAs tumor tissue by R package of gplot.Then we uploaded the different expression microRNAs to the IPA(Ingenuity Pathway Analysis)software,using the microRNA Target Filter to find the targeting information.Finally,we use the core analysis(Rapid assessment of the signaling and metabolic pathways,molecular networks,and biological process that are most significantly perturbed in the dataset of interest)to analyze those microRNAs associated network functions.(4)Cerebrospinal fluid(CSF)microRNA array data was downloaded from GEO database,and normalized by using limma package in R Language.we write the heatmap of different group to visualize the different expression genes and clustered the corresponding group with the different expressed miRNAs in CSF.Through cluster analysis,we found miRNA molecules that could correctly cluster samples.Then,collected the cerebrospinal fluid samples from patients who admitted in neurosurgery clinical center of Lanzhou University Hospital.The specific miRNAs was defined and the specific primers were designed.Finally,verified the reliability of the markers by fluorescence quantitative PCR.Results:(1)In grade III glioma,a total of 2036 differently expressed mRNAs and 50 miRNAs were confirmed by “DESeq” package in R.The 2036 genes are clustered into five modules by weighted gene co-expression networks analyses(WGCNA).In addition,we build the co-expression network of different expressed genes by the weight gene co-expression network analysis and visualized it with Cytoscape software.As is showed in the network,BUB1 B,KIFC1,TOP2 A,BUB1,SLC12A5,ESCO2,ESPL1,EPR1,KIF15,CASC5,SGOL1,NUSAP1,CCNB2,NUF2,TTK,KIF2 C are on the central of network.Moreover,we found that the network included two center,the downregulated genes and upregulated genes constitute the regulatory network respectively.BUB1 B,KIFC1,TOP2 A,BUB1,ESPL1,EPR1 are the center of overexpressed genes network;SLC12A5,VSNL1,SULT4A1,TMEM130,SNAP25 are on the central of lower expressed genes network.However,when we merged the different regulated mRNAseq and miRNAseq data to build the co-expression network,SLC12A5,MAL2,VSNL1,A2BP1,EPB49,SULT4A1,TMEM130,ADAM11,SNAP25,C1orf115,DNM1,SYT1 are on the central of network and mir-128,mir-129 are involved in.We can hypothesize that the genes in the central of network may be the hub genes of high grade LGG pathological process.(2)Merged the clinical data and gene expressed data,we find that KIF4 A,NCAPG,SGOL1,KLK7,SULT4A1 and TSHR intensively related with clinical prognosis.Mir-10 b,mir-27 a,mir-329-1 and mir-138-2 intensively related with survival time of glioma patients.(3)Using mi RNA microarrays data from the GEO database,we identify the 48 significantly differential expression miRNAs in medulloblastoma compared with the total normal group.The core analysis showed the molecular network interactions and signaling pathways associated with 28 differential expression microRNAs of medulloblastoma and their predicted molecular targets were rebuild using IPA.The network of ‘Organismal Injury and Abnormalities,Reproductive System Disease,Cancer' with IPA score of 41,Focus 17 miRNA Molecules,and ‘Cancer,Organismal Injury and Abnormalities,Cell Death and Survival' with IPA score of 23,Focus 11 mi RNA Molecules.The most impacted biological processes and diseases regulated by the analyzed miRNAs included: Organismal Injury and Abnormalities,Reproductive System Disease,Cancer.The molecular network maps in the networks 1 showed 3 main components were found to be at the central hub of the most significant network with a score of 41,which were TP53,AGO2 and ERK1/2.TP53,SIRT1 and YBX1 located in the central of networks 2 with a core analysis score of 23.(4)We think that current miRNAs cannot afford to distinguish the various brain tumor so far.Data mining of GSE62381 showed that the mir-205,-664 and-148 also can be treated as biomarkers to distinguish the metastasis from other intracranial tumor.After analysis the glioma and glioblastoma group with normal control.We found that miR-21,miR-16,miR-125 b,miR-223,and miR-142-3p were upregulated in glioblastoma.MiR-451,miR-142-3p,miR-25,miR-15 a,miR-16 and miR-144 high upregulated in glioma.Beside,we infer that mir-21,miR-16 were upregulated in many malignant brain tumors and mir-125 b may be the specific mi RNA biomarker of glioblastoma or poor prognostic glioma when compared with the low grade glioma.MiR-711 and-886-3p were down-regulated in primary brain Lymphoma respect to the normal and other cancer groups.Furthermore,mir-886-3p was downregulated in the Lymphoma group comparing with glioblastoma and medulloblastoma.Conclusions:(1)By the system biostatistics analysis,we may provide a better understand way to find the mechanisms of high grade gliomas tumorigenesis.Our result may predict that,two aspects involved in the glioma deterioration process,the downregulated genes and up regulated genes both play a vital character.(2)Clinical prognosis studies should be multidimensional,the prognosis of malignant tumors is not only related to the pathological glioma grade,but the specific molecular markers,since these molecular markers may play an important role in the development of tumors.(3)Mi RNAs may play an important role in the development and progression of medulloblastoma by acting on TP53,AGO2,ERK1 / 2,SIRT1 and YBX1.(4)In summary,the results of our analysis indicated that mi RNAs have great potential as noninvasive biomarkers for CNS cancers detection.However,present microRNA variety cannot afford to make an accurate diagnosis.The miRNAs assays seem to be more sensitive in the diagnosis of lymphoma and metastatic brain cancer than of glioma.By all means,further validation based on a larger sample of patients is still required,and with the development of the next generation sequence technology,the new mi RNA will be appeared to promote the accuracy of diagnosis in CNS disease.