Identification Of A CD4~+T Cell-based Prognostic Gene Signature Associated With Microenvironment Of Pancreatic Cancer

Background and aim:Increasing evidences have shown that the study of immune cells in the microenvironment of pancreatic cancer is of great significance in clinical treatment.However,reliable prognostic characteristics based on immune cell infiltration have not been established.This research aims to develop a genetic signature based on the status of immune cells for the risk stratification of pancreatic cancer.Methods:“x Cell”R packages wasappliedto quantize immune status.Immune cell status was estimated with microarray gene expression data for a discovery cohort from the TCGA database.The Cox regression model and the LASSO were applied to obtain hub genes and to develop an immune-based gene signature.In addition,the finding was validated in one independent cohort.Results:Five types of immune cells weredemonstrated to be closely related to the prognosis of pancreatic cancer patients(CD4~+memory T-cells,Th2 cells,CD4~+naive T-cells,CD4~+TCM,CD8~+T-cells).The LASSO model identified nine genes(SGSM2,MAK,RBM5,SH2B1,CEP55,ACTL6A,SLC25A44,HMGB2 and CD2AP)to construct a gene signature for risk stratification.Moreover,as a result of multivariate analysis,immune-based genetic markers were identified as independent prognostic factors(p=3.37e-14).Similar results were validated in patient data from the First Hospital of China Medical University(n=42)(p=0.005).Conclusions:We have established and verified a promising tool for risk stratification of pancreatic cancer based on immune cell characteristics of patients.Further,the genes in the signature might provide targets for immune treatment of pancreatic cancer.