Construction Of CircRNA-miRNA-mRNA Regulatory Network And Prognostic Risk Assessment Model For Colorectal Cancer

Objectives: In order to deeply investigate the expression distribution and functional analysis of circular RNAs（circ RNAs）in colorectal cancer（CRC）,this article takes differentially expressed circ RNAs（DEcirc RNAs）in tissues as the entry point,and explores the occurrence and development process of colorectal cancer based on the mechanism of ce RNAs containing micro RNA response elements（MREs）,and constructs a Cox risk regression clinical prediction model,Provide new ideas for future research and clinical treatment of CRC.Methods: First,download the circ RNA sequencing data and circ RNA chip data related to colorectal cancer from the gene expression comprehensive database（GEO）（all original expression data have been normalized and log2 converted）,and then apply the GEO2 R online tool and the online website Biojupices Circ RNA differentially expressed between colorectal cancer and normal tissues was screened from two data sets.Through the circ RNA database CRCD and the network tool circ Interactome,MRes（potential binding sites of mi RNA）of the six different circ RNAs screened were predicted,and mi RNA-m RNA was further predicted.The differentially expressed genes of CRC were collected in the TGCA database and intersected with the predicted mi RNA-targeted m RNA.After that,the genes obtained from the intersection were analyzed to study the relevant mechanisms involved in colorectal cancer.The m RNA input STRING software was used to predict their interactions and get gene-gene interaction network.The network map was scored by MCODE,which was used by Cytoscape software.The genes with high scores were identified and listed as hub-genes,then the circ RNA-mi RNA-m RNA critical regulatory network was mapped by Cyto Scape software.Then,LASSO COX regression analysis was used to construct risk prognostic models,Receiver operating characteristic（ROC）curves,survival analyses,nomograms to predict model efficacy and the ability to distinguish high-risk from low-risk populations.In addition,we performed enrichment analysis of function and pathway of differential genes between high-and low-risk groups,and further analyzed immune checkpoint expression in high-and low-risk groups based on the results of enrichment analysis.Finally,we assessed differences in Drug sensitivity between high-and low-risk groups according to the Half Maximal Inhibitory Concentration（IC50）available in the Genomics of Drug sensitivity in Cancer（GDSC）database,to obtain potential therapeutic drugs.Results: 1.The circRNA sequencing dataset GSE205241 and the circRNA microarray dataset GSE197991 screened out the differentially expressed circ RNAs（3up-regulated circ RNAs and 3 down-regulated circ RNAs）between colorectal cancer and normal tissues.By predicting the circ RNA-mi RNA of the six selected circ RNAs,six mi RNAs interacting with the differentially expressed circ RNAs in colorectal cancer and normal tissues were screened according to the binding sites and the mi RNA action pathway,including: hsa-mi R-3185、 hsa-mi R-635、hsa-mi R-499a-3p、has-mi R-920、 hsa-mi R-211、hsa-mi R-765.2.The differential genes were found to be involved in cell cycle-related biological processes through signal pathway analysis.Furthermore,nine hub genes（Node Score cutoff & GT;0.2 k-code & GT;2）were further screened from the differential m RNA network map,including CD4、 MAD2L1、 CXCL12、 CXCL13、CCL8、TOP2A、CCR8、IL17A and CCL20.The resulting circ RNA-mi RNA-m RNA regulatory network consists of three circrnas,three mirnas,and nine hub genes.3.Twelve survival-related mrnas were obtained by Univariate Cox analysis,and the optimized model was constructed by LASSO-COX analysis.Seven mrnas（BTN1A1、 SLCO1A2、 ZIC2、 AGAP3、 PLD5、LRAT、CD1B）were used as risk prognostic models.ROC curves（AUC1year = 0.751,AUC3 year = 0.740）,survival analysis,independent prognostic analysis,nomograms,and correction curves all demonstrated the predictive power of the risk prognostic model.4.Model-based GSEA analysis of high and low risk gene set enrichment pathway,it was found that the high-risk genes were enriched in lipid metabolism,nucleosome assembly protein,DNA packaging complex,nucleosome and chromatin structure-component-related pathways Low-risk gene sets are enriched for cell adhesion molecule,chemokine signaling pathways,hematopoietic cell lineages,intestinal immune networks for IGA production,antigen receptor mediation,immune response regulation,cell surface receptor signaling pathways,immunoglobulin production,and immunoglobulin complex-related pathways.In addition,the expression of TNFRSF25 in high-risk group was significantly higher than that in low-risk group,indicating that anti-TNFRSF25 immunotherapy may be effective in high-risk patients.5.By predicting the drug sensitivity of high-and low-risk groups,it was found that there were four chemotherapeutic drugs（KU-55933、AZD6482、Navitoclax、ERK₂440）with different sensitivity to CRC patients.The CRC patients in high-risk group were more sensitive to KU-55933、AZD6482,patients in the low-risk group were more sensitive to Navitoclax、ERK₂440.Conclusions: In this study,preliminary bioinformatic analysis suggested that HSA,HSA and HSA may influence the development of colorectal cancer through regulating transcription.The risk prognostic model composed of ABTN1A1,SLCO1A2,ZIC2,AGAP3,PLD5,LRAT and CD1 B was screened,which can provide some references for the prognosis and clinical treatment of patients.