Screening And Validation Of Biomarkers For Diagnosis And Prognosis Of Glioblastoma Multiforme

Objectives Glioma is the most common tumor in the central nervous system,and glioblastoma multiforme(GBM,grade IV),being the most maglignant tumor,occurs most commonly.70%-80% of patients with GBM have a survival time of 3-6 months and a 5-year survival rate of less than 10%.Although the integrated treatment mode of microsurgery+local precise radiotherapy+postoperative target gene therapy has greatly improved the prognosis of patients in recent years,these patients have a poorer prognosis influenced by many factors such as high tumor proliferation,aggressive and infiltrative growth,early tolerance of chemotherapy drug “temozolomide”,microvascular hyperplasia and mutation of related pathogenic genes.It is found that the discrepancy of tumor genotypes will have varied effects on the prognosis of patients.In 2016,WHO has applied the molecular pathological features of tumors such as IDH mutation,1p/19 q codeletion and MGMT promoter methylation to the classification of glioma more accurately.Therefore,the study of glioma gene signature provides a new drug target choice for refining its molecular diagnosis and improving the prognosis of patients.Based on this,this study visualizes the current situation and hot spots of GBM biomarkers by bibliometrics;screening the diagnosis and prognosis biomarkers of GBM suitable for both Eastern and Western populations by bioinformatics methods.Finally,the expression of screening biomarkers was verified by molecular biology methods,and the biological mechanism influencing glioma cell survival and its correlation with patient prognosis were explored,so as to provide theoretical basis for clinical diagnosis and treatment.Methods(1)Bibliometric analysis: Key databases were searched to obtain the research literature of GBM biomarkers,which were screened according to the inclusion and exclusion criteria.The analysis software of BICOMS(Bibliographic Item CoOccurrence Mining System)is used to extract and sort out the information includingkeywords,authors,countries and regions,research institutions,publishing time and journals,to generate cooperative network diagram and clustering network diagram for authors,research institutions,countries and regions and keywords,and to sort out HR(Hazard Ratio)and 95%CI(Credible Interval)of differentially expressed genes.(2)Bioinformatic analysis: The phenotype information,gene expression data and survival data of GBM patients were obtained from TCGA(The Cancer Genome Atlas)database,and the phenotype information and gene expression data of normal samples were screened from GTEx(The Genotype-Tissue Expression)project and standardized.Patients information in 693 and 325 datasets is downloaded from CGGA(Chinese glioma genome atlas)database.Differential expression analysis,univariate Cox proportional hazard regression and LASSO(least absolute shrinkage and selection operator)were employed to screen the prognosis-related genes of GBM patients.The prognosis index(PI)were calculated.K-M(Kaplan-Meier)survival curve and ROC(the time-dependent receiver operating characteristic)curve were established to analyze the prediction effect of PI model.Analysis of clinical confounding factors with PI was explored by multivariate Cox proportional hazards regression.The biological functions of prognostic genes were analyzed by GO(Geno ontology)analysis and KEGG(Kyoto Encyclopedia of Genes and Genomes)pathway enrichment analysis.The prognosisrelated genes screened in TCGA were verified in CGGA₃25 and CGGA₆93 datasets by univariate and multivariate Cox proportional hazards regression,K-M curve and ROC curve analysis.(3)Laboratory detection: q RT-PCR(real-time fluorescence quantitative PCR),Western Blot and immunohistochemistry were used to detect the expression of prognosis-related genes in varying grades of gliomas and normal brain tissues;The FN1 and HSPA5 genes of A172 glioma cell line were silenced.The effects of silencing FN1 and HSPA5 genes on proliferation,apoptosis and migration of A172 glioma cell line were detected by MTT assay,flow cytometry and Transwell cell migration assay.The effects of FN1 and HSPA5 gene silencing on the expression of Bax,Bcl-2,MMP9 and E-cadherin in A172 glioma cell line were detected by q RT-PCR and Western blot.The PFS(Progression-Free Survival)and OS(Overall Survival)of GBM patients were followed up.The influence of gene differential expression on the prognosis of GBM patients was analyzed by univariate and multivariate Cox proportional hazards regression and K-M survival curve,and the correlation between gene differential expression and clinicopathological parameters of patients was analyzed.Results(1)Current research status and hot spots of GBM biomarkers: 1771 articles were finally included in the English database and 123 articles in the Chinese database.The number of Chinese articles is less than 1/10 of that of English articles.116 institutions in 26 provinces and municipalities have participated in Chinese literature research,and 611 institutions in 18 countries have participated in English literature research.Cooperation among provinces,countries and institutions needs to be further strengthened.There are 239 keywords in Chinese literature,which can be classified into 5 categories by cluster analysis.There are 2963 keywords in English literature,which can be classified into 4 categories.The biomarkers based on cells and tissues in Chinese literature include 2-HG,CD133,CXCR4,Vimentin,Nestin,etc.In English literature,there are 271 biomarkers based on tissues research,87 biomarkers based on cells research and 900 biomarkers based on bioinformatics literature.Biomarkers based on cells,tissues and bioinformatics research include VEGF,TGFBI,CD44,COL1A2,COL3A1 and so on.310 micro RNAs have also been mentioned in English literature.Chinese literature focuses on MAPK signal pathway,PI3K-Akt signal pathway,P53 signal pathway and Fas/Fas L signal pathway.English literature focuses on Wnt/?-catenin signal pathway,Akt/ERK signal pathway,Akt/Fox M1 signal pathway,ATM/Chk2/p53 signal pathway,CDK4/6-RB signal pathway and so on.(2)Screening of biomarkers for diagnosis and prognosis of GBM: According to the threshold,581 differentially expressed genes were left.Among them,138 m RNA were highly expressed,443 m RNA were lowly expressed,and 20 m RNAs were finally selected.Sixteen m RNAs(RNF10,MTPN,RTN4,HSPA5,PLD3,GRN,FLII,NDUFB2,DKK3,MAP1LC3 A,SERPINE2,TTYH3,SCG5,FN1,TAGLN2,LY6E)are protooncogenes,with HRs from 1.230 to 2.039.Four m RNAs are antioncogenes(RPS19?EIF3L?EIF4A2?FDPS),HRs ? 0.789⁰.552.The HR of PI was 2.653(95%CI 1.976-4.122)(P<0.05).The results of K-M curve of PI showed that AUC of 3 years and 5 years were 0.824 and 0.820,respectively.Differentially expressed genes involved in 8 biological processes,which are cell adhesion molecule binding,chaperone binding,ubiquitin protein ligase binding,ubiquitin-like protein ligase binding,cadherin binding,translation initiation factor activity,translation factor active-RNA binding and unfolded protein binding.In validation section,we applied above 20 m RNAs from TCGA to test in CGGA.Uni-Cox results showed that 4 genes(MTPN,RTN4,MAP1LC3 A and DKK3)were not significantly correlation with OS,while 7 genes(LY6E,FLII,PLD3,EIF4A2,RNF10,RPS19 and NDUFB2)were statisticallysignificant only in CGGA₃25.The other 9 genes(TTYH3,TAGLN2,FN1,HSPA5,EI3 L,SCG5,SERPINE2,FDPS and GRN)have predictive effects on prognosis in CGGA₆93 and CGGA₃25.SCG5 and SERPINE2 showed opposite effects in CGGA and TCGA,and seven genes(TTYH3,TAGLN2,FN1,HSPA5,EI3 L,FDPS and GRN)showed the same effects(protooncogenes or antioncogenes)between TCGA and CGGA.Multi-Cox regression analysis of 20 m RNAs in CGGA showed that PI remained the key prognostic predictor,taking into account other clinical confounders.(3)Experimental verification of biomarker for diagnosis and prognosis of GBM: the results of q RT-PCR and Western blot showed that the m RNA and protein expression of FN1,GRN,HSPA5,TTYH3 and TAGLN2 in glioma tissues were significantly higher than those in normal brain tissues(P<0.05).With the increase of WHO classification,the m RNA and protein expression of FN1,GRN,HSPA5,TTYH3 and TAGLN2 in glioma increased gradually.Immunohistochemical results also showed that the positive expression rate of FN1,GRN,HSPA5,TTYH3 and TAGLN2 in glioma tissues were significantly higher than those in normal brain tissues(P<0.05).With the increase of WHO classification,the positive expression rate of FN1,GRN,HSPA5,TTYH3 and TAGLN2 protein in glioma increased gradually.Compared with A172-sh FN1-NC cells and A172-sh HSPA5-NC cells,the proliferation rates of A172-sh FN1 cells and A172-sh HSPA5 cells in 24 h,48h and 72 h decreased significantly(P<0.05),the apoptosis rates in 48 h increased significantly(P<0.05),the numbers of migrating cells in 48 h decreased significantly(P<0.05),the m RNA and protein expression of Bax and E-cadherin increased significantly(P<0.05),while the m RNA and protein expressions of Bcl-2 and MMP9 decreased significantly(P<0.05).Compared with the A172 cells,there were no significant changes in all cell indexes in A172-sh FN1-NC cells and A172-sh HSPA5-NC cells.Univariate Cox analysis showed that the median survival time and average survival time of PFS in GBM patients with FN1,HSPA5,TAGLN2 and GRN positive expression were significantly decreased(P<0.05).The results of multivariate Cox analysis showed that the expression of FN1,HSPA5 and TAGLN2 had independent influence on PFS of GBM patients.Univariate Cox analysis showed that the median survival time and average survival time of OS in patients with FN1,HSPA5,TAGLN2 and GRN positive expression were significantly decreased(P<0.05).The results of multivariate Cox analysis showed that the expressions of FN1,HSPA5,GRN and TAGLN2 had independent influence on OS of GBM patients(P<0.05).Conclusions(1)The Chinese documents available on the research of GBMbiomarker is small in quantity while the English literature is relatively large.However,there is still much room for cooperation between authors,research institutions and countries.Chinese research topics are relatively scattered,while English research topics are relatively concentrated;but there is room for both of them to extend their research.We should pay more attention to the influence of biomarker on the prognosis of GBM patients while conducting tissue and cell research of biomarker.(2)Through bioinformatics analysis,20 m RNA are related to the prognosis of patients,namely RNF10,MTPN,RTN4,HSPA5,PLD3,GRN,FLII,NDUFB2,DKK3,MAP1LC3 A,SERPINE2,TTYH3.They were mainly involved in cell adhesion molecule binding,chaperone binding,ubiquitin-like protein ligase binding and other biological functions,involving tumor proliferation,invasion,migration and other mechanisms.The PI prediction model can effectively predict the incidence risk of GBM in TCGA,as well as in Chinese patients with GBM.(3)There are 7 genes which have the same expression trend in TCGA and CGGA databases and have the same effect on GBM,namely TTYH3,TALLN2,FN1,HSPA5,EI3 L,FDPS and GRN,among which TTYH3,TALLN2,FN1,HSPA5 and GRN are protooncogenes,while EI3 L and FDPS are antioncogenes.These seven genes may be the suitable biomarker for diagnosis and prognosis of GBM for both eastern and western populations.(4)The high expression of FN1,GRN,HSPA5,TTYH3 and TAGLN2 in glioma tissue may promote the occurrence and development of glioma.The high expression of FN1 and HSPA5 may affect the biological behavior of tumor cells and promote the occurrence and development of glioma by enhancing their proliferation,inhibiting apoptosis and promoting metastasis.(5)The high expression of FN1,HSPA5,TAGLN2,GRN affects the progression-free survival time and overall survival time of GBM patients,and is negatively correlated with the prognosis of GBM patients.FN1,HSPA5,TAGLN2 and GRN can be used as double biomarkers for diagnosis and prognosis of GBM,providing reference for clinical diagnosis and treatment to explore new therapeutic targets.