A Circadian Clock Gene Relevant Prognosis Model Of Stomach Adenocarcinoma Using The Cancer Genome Atlas

Background:Gastric cancer is the most common and deadliest malignancy in the world.Although it is too early to talk about a complete cure for gastric cancer,there is an urgent need for new therapies and drugs to control the growth process of the tumour.The Circadian clock gene(CCG)plays an important role in various tumours.Although a number of key CCGs have been shown to be involved in various tumourigenic processes,the molecular function of this genome has not been fully revealed.The aim of our study is to develop a CCG-related prognostic risk model for gastric cancer and to assess the prognosis of gastric cancer patients through risk modelling.Method:The RNA-seq data from The Cancer Genome Atlas(TCGA)of gastric adenocarcinoma(STAD,containing 375 cancer tissue samples and 26 paired normal tissue samples)were selected to screen for differentially expressed genes(DEGs)after data missing value completion and data cleaning.And the list of CCGs was downloaded from Genecards,the differentially expressed CCGs were screened out,and then 17 of the differentially expressed CCGs were screened by the Least Absolute Shrinkage and Selection Operator(LASSO)algorithm to identify the 17 CCGs that were significantly associated with the prognosis of 334 gastric cancer patients with a follow-up time of >30 days,and these 17 CCGs were further included in the multifactorial COX regression analysis.As result,9 CCGs significantly associated with prognosis in gastric cancer were used to construct prognostic risk models,and their regression coefficients in the stepwise multivariate COX regression analysis and their expressions in the matrix file were used to calculate risk scores.The predictive power of the risk model was also assessed by the validation set(GSE84437).In addition,the relationship between CCG-related prognostic risk models and immune infiltration,gene mutations,tumor microenvironment scores,drug sensitivity,and some tumor-related pathways was explored by Spearman correlation analysis and enrichment analysis.Finally,a Protein-Protein Interaction(PPI)network was constructed based on the differential genes between the high-risk and low-risk groups.And by performing Kyoto encyclopedia of genes and genomes(KEGG),Gene Ontology(GO)enrichment analysis on these differential genes,we initially explored some possible changes in cellular biological functions brought about by the CCG prognostic risk model.Result:In this study,256 differentially expressed CCGs were screened,and 9 prognosis-associated CCGs(SERPINE1,GFRA1,SST,GRP,NFKB2,LDHD,CCT6 A,HABP2 and ADH4)were identified by LASSO algorithm and stepwise multivariate COX regression analysis,and the corresponding CCG risk scores(CCGRS)were obtained.9 CCGs was also used to build a prognosis risk model.Kaplan-Meier(KM)analysis by dividing patients into high-risk and low-risk groups revealed that the overall survival of patients in the high-risk group was much lower than that of patients in the low-risk group(p=4.8e-14).The predictive performance of CCGRS for OS was assessed using time-dependent ROC analysis,and the ROC curves showed that the AUC corresponded to 0.68,0.79 and 0.84 at 1,3 and 5 years,respectively.We assessed the predictive power of the model by validation set GSE84437.Both KM analysis and time-dependent ROC analysis showed that the prognosis risk model associated with CCG in dataset GSE84437 still had good predictive power.Our prognosis risk model has good predictive accuracy.In addition,to explore the function of the resulting prognosis model,we used some high-quality survival data such as disease-specific survival(DSS),disease-free interval(DFI)and progression-free interval(PFI)from the TCGA database for GC patients and some clinical data such as clinical tumour stage(TNM stage)and pathological stage for TCGA gastric cancer patients,and conducted a series of analysis combined with the CCGRS and risk groups.The results suggest that the CCG-related prognosis model we developed can be used as an independent risk factor to predict the prognosis of gastric cancer patients.The results of the molecular biological functional study on the risk model showed a significant positive correlation between CCGRS and lipid metabolism GSVA score(r=0.212,p<0.001),iron death GSVA score(r=0.137,p=0.012),cell aging GSVA score(r=0.123,p=0.025)and cellular autophagy GSVA score(r=0.131,p=0.017).CCGRS was significant negative correlatted with tumour mutation burden(TMB)(r=-0.167,p=0.002)and microsatellite instability(MSI)(r=-0.159,p=0.004).What’s more,the most significant mutational mode for the nine gastric cancer prognosisrelated CCG was missense mutation.In addition,the IC50 values of various chemotherapeutic agents such as 5-fluorouracil,belinostat,crizotinib,doxorubicin,linifanib,mitomycin C,sorafenib,tiffinib,tupastin,vincristine,zilburtan,cyclopamine,dasatinib,imatinib,lapatinib and ricitinib were significantly different between high and low risk groups.CCG-related risk models are also involved in multiple ways by modulating immune cell infiltration in patients with gastric cancer.A PPI network was constructed based on 577 differentially expressed genes between high-and low-risk subgroups,and functional enrichment analysis of the differentially expressed genes was performed to further reveal the potential mechanisms of tumorigenesis and development brought by risk models.Conclusion : In this study,a CCG-related prognosis risk model was constructed,and prognosis risk model has promising applications for predicting clinical outcomes and results of GC patients.