Construction And Verification Of The Prognostic Model Based On Autophagy-Related Long Non-coding RNA In Colorectal Cancer

Objective:Autophagy plays a key role in maintaining the stability of internal environment.Current studies have confirmed that autophagy is significantly related to the formation and development of malignant tumors.This study further explored the potential prognostic value of autophagy-related long non-coding RNA(ATRlncRNA)in colorectal cancer and related molecular regulation,providing a new basis for colorectal cancer prognosis monitoring,individualized treatment and tumor progression mechanisms.Method:Pearson correlation analysis were performed for autophagy genes(ATGs)and lncRNA expression matrix of TCGA colorectal cancer to obtain ATRlncRNA expression matrix,and then competitive endogenous RNA(ceRNA)network mediated by ATRlncRNA was constructed.The clinical samples obtained from TCGA were randomly divided into training group and test group at a ratio of 1:1,and the data set obtained from GEO was used as the validation group.Univariate Cox regression analysis,LASSO regression analysis,and partial likelihood deviance were performed on the training group to obtain prognostic-related ATRlncRNA,which were then conducted with multivariate Cox regression analysis to obtain the prognostic model based on the ATRlncRNA.And according to the median value of risk score of the prognostic model,the samples were classified into the low-risk group and the high-risk group.Kaplan-Meier survival analysis,independent prognosis analysis and multi-factor receiver operating characteristic(ROC)curves were conducted on the prognostic model to explore its prognostic value,clinical applicability and accuracy,whoseVresults were verified by test group and validation group.Gene Set Enrichment Analysis(GSEA)was performed on the gene expression matrix of the high-and low-risk group to explore the existing gene function positioning in colorectal cancer.The immune infiltration analysis containing 22 immunocytes and tumor immune check points were performed on CRC samples based on the prognostic signature.Moreover,the expression of six ATRlncRNAs of the prognostic model was conducted with expression difference validation with ln CAR database and Spearman correlation analysis for cancer stem cell(CSC)index and the tumor microenvironment(TME)to verify the expression differences of the six ATRlncRNAs of the prognostic model and to explore the relationship between ATRlncRNA and the progression of colorectal cancer at the cellular level.Result:(1)210 autophagy-related genes in CRC were obtained.After Pearson correlation analysis for ATG expression matrix and lncRNA expression matrix,3145 ATRlncRNA were obtained.Meantime,we constructed a ceRNA network mediated by ATRlncRNA,which contained 30 ATRlncRNA,18 miRNA and 29 mRNA.(2)The 496 CRC samples obtained from the TCGA database were divided into training group(n=248)and test group(n=248)at the ratio 1:1.294 CRC samples obtained from GEO were used as the validation group.Univariate Cox regression analysis,LASSO regression analysis,partial likelihood deviance as well as multivariate Cox regression analysis were performed on the training group to obtain six-ATRlncRNA(ALMS1-IT1,FGD5-AS1,FLG-AS1,MIR210 HG,MIR31HG,PINK1-AS)prognostic model.Kaplan-Meier survival analysis was performed on samples of training group,test group and validation group.It was observed that the survival rate of samples in the high-risk group was lower than that in the low-risk group(P<0.05).In the independent prognosis analysis of theVItraining group,test group and validation group,two independent prognostic factors were obtained: age and the risk score of the prognostic model,which proved that the prognostic model could be directly used as an independent factor to predict the prognosis of patients.The AUCs of multi-factor ROC curves including the risk score of the prognostic model was respectively0.711,0.776 and 0.693 in the training group,test group and validation group,showing the good accuracy of the model.(3)About the results of GSEA for the high-and low-risk group,the gene set of high-risk group was mainly active in the hallmark gene set "Hypoxia",but not enriched in KEGG pathways.The gene set of low-risk group was enriched in KEGG pathways,including "Peroxisome","The citrate cycle(TCA cycle)" and "Other glycan degradation".(4)In the immune infiltration analysis of TCGA CRC high-risk group and low-risk group,the contents of plasma cells and memory resting CD4+T cells in the low-risk group were higher than those in the high-risk group,and Spearman correlation analysis for the the relative contents of 22 kinds of immune cells and risk score showed that the relative contents of plasma cells and memory resting CD4+ T cells were negatively correlated with the risk score.The expression levels of PDCD1(PD-1),CTLA4(CTLA-4)and HAVCR2(TIM-3)in the high-risk group were higher than those in the lowrisk group.The expression levels of CTLA4 and HAVCR2 were positively correlated with the risk score of the prognostic model.(5)The expression quantities of ALMS1-IT1 and MIR31 HG were higher in tumor tissues than that in normal tissues,and the expression quantities of FGD5-AS1,FLG-AS1,MIR210 HG and PINK1-AS were higher in normal tissues compared with that in tumor tissues.Moreover,there were a positive correlation between CSC index and expression quantities of ALMS1-IT1,and a negative correlation between CSC index and expression quantities of FGD5-AS1,FLG-AS1 and MIR31 HG.The stromal score was positively correlated with the expression of FGD5-AS1,FLG-AS1 and MIR31 HG.The immune score and estimate score were positively correlated with the expression of FLG-AS1 and MIR31 HG,and negatively correlated with the expression of ALMS1-IT1.Conclusion:The prognostic model constructed based on six ATRlncRNAs reflected auspicious clinical applicability and prognostic accuracy,and might become an indicator for independently predicting the prognosis of patients.The ceRNA network mediated by ATRlncRNA revealed the potential interaction between ATRlncRNA and other molecules.The hierarchical GSEA analysis of the high-risk group and the low-risk group further explored the biological functions related to the progression of colorectal cancer and provided new perspectives for individualized treatment.The correlation analysis for the prognostic model and tumor immune infiltration validated the prognostic value of immunocytes.In addition,the correlation analysis between the expression of the six ATRlncRNAs of the prognostic model and the CSC index and TME score provided new clues for the study of colorectal cancer progression.