| Objective: Cancer is one of the diseases that threaten human health in the 21 st century.A ccording to statistics,one out of every four people in the world is at risk of developing cancer.The most harmful cancers are breast cancer,colorectal cancer,liver cancer and kidney cancer.The incidence and mortality of cancer in developed countries is significantly higher than that in developing countries.The continuing trend of global cancer morbidity a nd mortality is predicting that cancer will become the number one killer of human beings and poses a huge challenge to public health.Improving the prognosis of cancer patients is a key factor in improving patient survival.Therefore,finding biomarkers related to cancer prognosis,early assessment of the prognosis of cancer patients through biomarkers is very important to improve the prognosis and survival rate of cancer patients.A large number of studies have shown that cancer cells have metabolic pathway changes compared to normal cells: despite the presence of oxygen,cancer cells preferentially met abolize glucose through glycolysis,producing lactic acid as the end product.The abnormally active glycolysis process is the main source of energy supply in cancer cells through which adenosine triphosphate(ATP)is produced.Glycolysis has become an important metabolic feature of cancer,and changes in its metabolic pathways are usually regulated by the expression of specific genes,which may be differentially expressed genes in cancer and adjacent tissues.A gene signature is the combination of a certain number of molecular markers to form a new variable that is used to determine or define certain biological properties.Unlike the molecular pattern of a single marker,molecular tags are not only based on single-gene function,but also focus on the common coordination between genes,describing a specific biological property from the overall and systemic levels.There is growing evidence that molecular tags are becoming a better choice.In recent years,thousands of biomarkers related to survival and prognosis that can be retrieved through database mining have been explored,but most of the markers have not been clinically applied and have been selected and analyzed for a specific feature or a single signaling pathway.The research is very limited,which may be a new entry point for our research.This study intends to construct a gene signature related to glycolysis as a prognosis by mining transcriptome sequencing data of four common cancers(hepatocellular car cinoma,breast invasive carcinoma,renal clear cell carcinoma,colorectal adenocarcinoma).Methods: 1.Download TCGA hepatocellular carcinoma,breast invasive carcinoma,renal clear cell carcinoma,colorectal adenocarcinoma-related m RNA sequencing and chip data.2.GSEA analyzes three glycolysis-related gene sets and screens differentially expressed genes in cancer and adjacent tissues.3.Univariate and multivariate Cox proportional regression models were used to screen for prognostic-related mRNAs,and molecular markers based on mRNAs expression profiles were constructed to predict patient prognosis.4.Kaplan-Meier curve method predicts the predictive power of molecular markers for the prognosis of cancer patients.5.ROC curve method to verify the diagnostic performance of molecular tags.6.Patients were randomly divided into training sets and test sets to verify the effectivene ss of molecular tags in predicting patient prognosis.7.Univariate and multivariate Cox regression analysis of the association of m RNAs with other clinical variables.8.The DAVID online analysis tool performs functional annotation and pathway enrichment analysis of differentially expressed genes in different groups of patients.Results: 1.Different glycolytic-related differentially expressed genes were screened in colorectal adenocarcinoma,hepatocellular carcinoma and breast invasive carcinoma.Compared with adjacent tissues,75 differentially expressed mRNAs were found in colorectal adenocarcinoma tissues(P<0.05).109 differentially expressed mRNAs were found in he patocellular carcinoma tissues(P<0.05),101 differentially expressed m RNAs were found in breast invasive carcinoma tissues(P<0.05),and 0 differentially expressed mRNAs we re found in renal clear cell carcinoma.2.Univariate Cox proportional regression model analysis showed no correlation with prognosis-related m RNA in colorectal adenocarcinoma and breast invasive carcinoma(P>0.05).22 m RNAs were associated with OS in patient s with hepatocellular carcinoma(P<0.001).Therefore,hepatocellular carcinoma was selected for subsequent analysis.3.Further Cox multivariate regression model analysis,obtained molecular markers consisting of 6 m RNAs(DPYSL4,HOMER1,ABCB6,CENPA,CDK1,STMN1).4.Calculate the risk score of each patient based on the regression coefficients of the six genes and the gene expression values of the patient specimens.Risk score =(0.1142 × DP YSL4 expression value)+(0.1982 × HOMER1 expression value)+(0.2647 × ABCB6 ex pression value)+(0.4603 × CENPA expression value)-(0.5359 × CDK1 expression value)+(0.3966 × expression value of STMN1).Based on the median risk score,309 patients with hepatocellular carcinoma were divided into high-risk group(n = 154)and low-risk group(n = 155).5.In the training set,test set and total hepatocellular carcinoma samples,the correlation between different risk groups and patient prognosis was analyzed.The results showed that compared with the low-risk group,the high-risk group and overall survival time were significantly lower(P<0.05).).6.ROC analysis of molecular tag diagnostic efficacy,AUC = 0.765,suggesting that the molecular tag has a good diagnostic value.7.Univariate and multivariate Cox regression analysis evaluated whether the molecular markers of these 7 m RNAs were independent of other clinical variables,and found that in single-factor and multi-factor Cox regression analysis,risk score was still significant with the overall survival time of patients.Correlation suggests that the molecular tag may be an independent prognostic factor for patients with hepatocellular carcinoma.Conclusion: 1.DPYSL4,HOMER1,ABCB6,CENPA,CDK1,STMN1 are highly expre ssed in liver cancer tissues and are associated with poor prognosis of hepatocellular carcinoma.2.The molecular markers consisting of DPYSL4,HOMER1,ABCB6,CENPA,CDK1,and STMN1 are independent risk factors for the prognosis of hepatocellular carcinoma.3.This study newly constructed a molecular tag related to glycolysis as a prognostic risk marker for hepatocellular carcinoma patients,and provided new ideas and molecular targets for the research and individualized treatment of liver cancer. |
PDF Full Text Request