| Objective: Gastric cancer(GC)is now considered one of the most highly aggressive malignant tumors worldwide.Most patients with GC are the advanced stage at the time of diagnosis and even lost the opportunity for surgical resection.Due to recurrence and metastasis,patients with advanced GC have a lower five-year overall survival rate.Therefore,early diagnosis and prognosis evaluation of GC is particularly important.With the rapid development of genomics,microarray and RNA sequencing technologies are widely used.In the research and treatment of malignant tumors,the above two technical methods also have important research significance.The emergence of bioinformatics has raised the awareness and understanding of malignant tumors to a new level.Tumor information data mining has enhanced and deepened the deep understandings of the molecular mechanisms of cancer etiology.Selecting differentially expressed genes(DEGs)based on integrated bioinformatics analyses has been used in recent studies to explore potential biomarkers in GC with microarray and RNA sequencing data.However,the results obtained by the study may be inaccurate due to defects in analytical methods and insufficient clinical sample size.Therefore,it is particularly important to look for multi-level and multi-layered factors as molecular biomarkers for diagnosis,prediction,prognosis and treatment of GC.However,cancer is a complex disease with multiple molecules involved.A single gene or molecule can have certain limitations even as a potential prognostic marker.In recent years,researchers have continued to understand and explore cancer etiology using database sequencing data from the Gene Expression Omnibus(GEO)database or The Cancer Genome Atlas(TCGA),combined with effective bioinformatics analysis.By various methods,we have discovered more and more combinations of genes,which were considered as gene signatures.In these models,there are several or even dozens of genes that can be used as diagnostic,prognostic and therapeutic targets for cancer patients.Therefore,the aim of this study was to find robust DEGs and effective gene signature with prognostic value for GC.Methods: In the first part of the study,the GC gene expression profiles of eight GEOdata sets were first downloaded,including GSE19826,GSE33335,GSE63089,GSE27342,GSE56807,GSE54129,GSE26942,and GSE79973.The limma method was used to screen DEGs in each data set,and then the Robust Rank Aggregation(RRA)algorithm was then employed to integrate all the datasets above and to select final robust DEGs.In order to describe and understand the biological functions and the signaling pathways involved of these DEGs,we finally used bioinformatics analysis to annotate and performed enrichment analysis,such as Gene Ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)pathway enrichment analysis.In the second part of the study,we used comprehensive bioinformatics analysis such as protein network data mining and Fisher's exact test to screen hub genes.Next,we validated the differential expression of above hub genes using the GC expression profile in the TCGA database.At the same time,based on the stepwise multivariate COX proportional hazard regression analysis,a seven-gene signature with prognostic value was established.The area(AUC)under the receiver operating characteristic curve(ROC)was used to predict the 5-year patient survival rate.The Kaplan-Meier(K-M)curves were used to compare survival outcomes between two different groups.In the end,this gene signature was validated by an GEO external data set.In the analysis of the third part,based on the results of the above analysis,we selected the PLAU gene for biological functions researches.First,we analyzed the relationship between PLAU gene expression and clinicopathological data in patients with GC.Next,in the MGC-803 and SGC-7901 cell lines,CCK-8 assay was used to detect the cell proliferation in the knockdown of PLAU(si-PLAU)group and the corresponding negative control(NC)group.Transwell cell migration and invasion assays were used to detect cell migration and invasion between si-PLAU and NC groups,respectively.Results: In the first part of the analysis,we first screened and downloaded a total of 8GEO GC microarray data sets.After the limma method,the corresponding DEGs were obtained in their respective GEO data sets.Using the RRA algorithm,we obtained 346 robust DEGs,including 140 up-regulated genes and 206 down-regulated genes.Based on the GO and KEGG enrichment analysis,the results showed that these up-regulated DEGs were significantly correlated with biological processes such as cell adhesion,cytoskeletal activity and binding.Meanwhile,these up-regulated genes were significantly enriched inECM-receptor interaction,Focal adhesion,PI3K-Akt signaling pathway,Cell adhesion molecules,p53 signaling pathway,Tight junction and other biological pathways.The down-regulated DEGs were significantly involved in biological functions such as energy metabolism and binding.We found that down-regulated genes were significantly enriched in Metabolic pathways,Chemical carcinogenesis,Retinol metabolism,Gastric acid secretion,Glycolysis and other substance energy metabolic pathways.In the second part of the study,based on the 346 robust DEGs above,we used the HIPPIE protein interaction network database and Fisher's exact test algorithm to screen out 11 hub genes.They are THBS1,SPARC,COL1A1,COL4A1,PLAU,COL1A2,MMP1,FBN1,ATP4 A,COL2A1,and MYOC.Pathway enrichment analysis showed that the above genes were significantly enriched in ECM receptor interactions,Focal adhesions,PI3K-Akt signaling pathways,and Cancer proteoglycans,which are closely related to the development of cancer.Next,we used a stepwise multivariate COX proportional hazard regression analysis,the expression of the above 11 hub genes and patient survival information,we screened a total of 7 genes(FBN1,MMP1,PLAU,SPARC,COL1A2,COL2A1 and ATP4A),and a 7 gene signature with prognostic value in GC was constructed.We found that the model can predict a 5-year survival rate of0.816 for patients.Next,the K-M curve showed that overall survival was significantly worse in the high-risk group compared with the low-risk group(log-rank test p-value <0.001).The above results indicate that this gene signature has a good prognostic value.Finally,we validated the prognostic value of this model using the independent dataset GSE62254 from the GEO database.In the third part of the experiment,first,using the expression profile and clinicopathological data of GC patients in the TCGA database,we found that the expression level of PLAU gene was related to the age and pT classification of patients,and GC patient survival outcomes.Next,we used CCK-8 in SGC-7901 and MGC-803 cell lines to find that there was no significant difference in cell proliferation ability between si-PLAU group and NC group.Using the Transwell cell migration assay,the si-PLAU group significantly inhibited the migration of cells compared with the NC group.The cell invasion assay also revealed that the si-PLAU group significantly inhibited the invasion ability of cells compared with the NC group.Conclusion: In this study,a total of 346 significant DEGs were identified using the RRA algorithm,including 140 significantly up-regulated DEGs and 206 significantly down-regulated DEGs.Based on a series of bioinformatics analysis,it was found that up-regulated genes were significantly involved in biological processes such as cell adhesion and cytoskeletal activity,while down-regulated genes were significantly involved in various substance metabolism and glycolysis pathways.Based on the above analysis and screening,a 7 prognostic gene signature was established,which can effectively evaluate the prognosis of GC patients.The expression level of PLAU gene was correlated with the age and pT classification of patients,and there was a significant correlation between the expression of PLAU gene and patient survival.Knockdown of the PLAU gene inhibited migration and invasion of GC cells,but had no significant effect on cell proliferation. |
PDF Full Text Request