Construction Of Prognostic Prediction Model For Lung Adenocarcinoma Based On Epithelial-mesenchymal Transition Related Genes

Objective: The foremost cause of mortality from malignant tumors globally is lung cancer,with lung adenocarcinoma(LUAD)being the most prevalent non-small cell cancer.The prognosis of lung adenocarcinoma patients is dismal,yet it is a remarkably diverse and aggressive malignancy.The identification of successful prognostic biomarkers could be advantageous in the clinical diagnosis and treatment of lung adenocarcinoma.The prognosis of lung adenocarcinoma is closely linked to the abnormal activation of epithelialmesenchymal transition(EMT),which is a major factor in its occurrence,development and metastasis.This study seeks to utilize LUAD’s transcriptome sequencing data from The Cancer Genome Atlas(TCGA)to construct a prognostic model based on EMT-related genes,thereby furnishing a foundation for forecasting the survival outcome of lung adenocarcinoma patients.Methods: 1.Construction of prognostic risk model:Gen Clip3 was employed to screen EMT-related genes that were differentially expressed,while the common genes were obtained from TCGA-LUAD and GSE10072 datasets.Cox regression analysis of univariate was conducted to identify prognostic genes associated with EMT,and the prognostic risk model was constructed using LASSO and multivariate Cox regression.The model’s reliability was tested by selecting GSE50081,GSE68465,and GSE72094 datasets from Gene Expression Omnibus(GEO)as external validation,with lung adenocarcinoma patients divided into high and low risk groups based on the median risk score.The K-M survival curves were plotted to compare the prognostic difference.The ROC curve and risk curve were employed to designate the predictive efficiency in TCGA-LUAD and validation cohorts.Independent prognostic analysis was conducted to ascertain if the model could be employed as a distinct prognostic factor for those suffering from lung adenocarcinoma.By combining the risk score and clinical information of TCGA-LUAD,a nomogram was created to quantitatively forecast the survival probability of individual patients.Analysis of GSEA、drug sensitivity and correlation between immune systems were applied to the TCGA-LUAD high and low risk groups.The GSCALite tool was used to describe the related characteristics of the 9 genes in the model in lung adenocarcinoma.Analysis of mRNA and protein levels expression of TIMP1 genes in the model was conducted using the GEPIA2,UALCAN and HPA databases in both the lung adenocarcinoma tissues and the normal tissues.2.RT-PCR: Analysis of TIMP1 mRNA expression in 20 pairs of lung adenocarcinoma and paracancerous tissues was conducted using RT-PCR.3.Statistical analysis: The Graph Pad Prism software was used for analysis,and the statistical method was paired t test.P<0.05 was considered statistically significant.Results: 1.Prognostic model: In this study,an EMT-related prognostic risk model composed of 9 EMT-related genes was constructed,and the risk score was calculated by the formula: Risk Score=0.273×LGR4-0.136×MRC1+ 0.289×TIMP1+0.197×GPI+0.084×ID1+ 0.164×CAV1+ 0.217×CDKN3+ 0.147×EFNB2-0.086×NDNF.The median risk score was used to divide the patients into two groups: those with high and low risk.In GSE50081,GSE68465,GSE72094 datasets and TCGA-LUAD,the K-M curve,ROC curve,and risk curve were all drawn.The area under curve of ROC is greater than 0.6,indicating that the prediction accuracy of this model was satisfactory,and the K-M curve indicated that median survival time with high-risk scores was lower than that of those with low-risk scores in each group,P < 0.05.Independent prognostic analysis showed that the risk score HR > 1,indicating that the prognostic model is a risk factor for the prognosis of lung adenocarcinoma.Calibration curve and decision curve show that the nomogram has good predictive performance.2.Correlation analysis of high and low risk group: GSEA enrichment analysis found that cell cycle,extracellular matrix-receptor interaction and cancer pathway were active in the high risk group.The immune function of MHC class I and Parainflammation were more active in the high risk group.The results of drug sensitivity analysis showed that Gefitinib,Crizotinib,Trametinib,Erlotinib,Epirubicin and Cisplatin were more sensitive in the high-risk group.3.Characteristic analysis of model genes in lung adenocarcinoma: Compared with normal tissues,TIMP1,LGR4,CDKN3 and GPI genes were highly expressed in lung adenocarcinoma tissues.The results of single nucleotide variation analysis showed that The mutation frequency of gene LGR4 was as high as 31%.The copy number variation analysis of the model genes showed that the proportion of heterozygous genes was significantly higher than that of homozygous genes.The methylation levels of CAV1,EFNB2 and ID1 genes were higher in tumor tissues.4.TIMP1 gene expression analysis: The analysis results showed that TIMP1 gene and TIMP1 protein were significantly overexpressed in lung adenocarcinoma compared with normal lung tissue.RT-PCR was used to verify the results of tumor tissues and paracancerous tissues samples from 20 patients with lung adenocarcinoma.TIMP1 gene expression was significantly increased in lung adenocarcinoma tissues(P < 0.05).Conclusion: The prognostic model of EMT-related genes was constructed based on TCGA-LUAD and GSE10072 dataset,with a precise prediction accuracy of the model.RT-PCR results showed that it has a significant difference in the expression of TIMP1 gene between lung adenocarcinoma tissues and paracancerous tissues,and TIMP1 was highly expressed in tumor tissues.