Bioinformatics Analysis Of Prognostic Genes Related To Tumor Microenvironment In Pancreatic Cancer

Objective:Pancreatic adenocarcinoma(PAAD)is known as the “king of cancer”and has a very poor prognosis.A special pathological feature of PAAD is a denser fibroproliferative response in the tumor stroma,forming a unique tumor microenvironment(TME).Traditional treatments for pancreatic cancer,including surgery and chemotherapy,have not seen significant results in improving patient survival.Targeted therapy and immunotherapy targeting TME has become a major hotspot in the current treatment of pancreatic cancer.The purpose of our study is to find TME-related prognostic genes in pancreatic cancer using bioinformatics analysis,and is expected to provide new biomarkers for targeted therapy of TME.Methods:The data of PAAD patients in TCGA database were collected and divided into high/low immune score and high/low stromal score group by ESTIMATE algorithm.The correlation between immune/stromal score and PAAD prognosis was analyzed by R software.Differentially expressed genes(DEGs)in each group were screened.Weighted gene co-expression network analysis(WGCNA)method was used to extract the module genes most correlated with immune/stromal score,respectively.The intersection was taken to obtain co-expressed differential genes.GO(Gene Ontology),KEGG(Kyoto Encyclopedia of Genes and Genomes)enrichment analysis and protein-protein interaction network(PPI)were used for further functional validation.Lasso-Cox regression analysis was used to determine the prognostic genes associated with overall survival,and then a risk score model was constructed to evaluate the application effect of the risk score model by KaplanMeier curves and receiver operating characteristics(ROC)curves.Multivariate analysis was performed to screen independent prognostic factors for PAAD using the Cox hazard proportional regression model.GEPIA(Gene Expression Profiling Interactive Analysis),HPA(Human Protein Atlas),TIMER(Tumor IMmune Estimation Resource),KM(KaplanMeier Plotter)and other databases were used to validate the prognostic genes in terms of gene transcription level,protein expression level,immune infiltration analysis,and prognostic analysis,and finally clinical experimental validation was performed by immunohistochemistry.Results:1.A total of 178 patients with PAAD were included in this study.High immune score and high stromal score were both associated with poor prognosis in PAAD(p=0.012,0.025).1248 DEGs were selected for the immnue group and 1525 for the stromal group.Combined with WGCNA analysis 250 co-expressed differential genes were finally obtained.2.3 TME-related genes KYNU(Kynureninase),ZBED2(Zinc Finger BED-Type Containing2),CXCL10(C-X-C Motif Chemokine Ligand 10)were associated with poor prognosis in PAAD,risk score model: Risk Score(RS)=0.207739268683587*KYNU+0.0484033837407738*ZBED2+0.0317084731747505*CXCL10.There were significant prognostic differences between the high and low RS groups(p=6.6e-5).ROC analysis showed that the area under the curve(AUC)at 365,1095 and 1825 time points was 0.71,0.79 and 0.87,respectively.Further multivariate regression analysis showed that risk score and previous history of chronic pancreatitis were two independent risk factors for PAAD(p < 0.001,0.001).3.KYNU,ZBED2 and CXCL10 expression were negatively correlated with overall survival(OS)of PAAD(p=0.0029,0.001,0.0022);ZBED2 and CXCL10 were also negatively correlated with recurrence-free survival(RFS)of PAAD(p=1.7e-05,0.00082).The expression levels of KYNU,ZBED2 and CXCL10 were negatively correlated with tumor purity but positively correlated with immune cell infiltration in TME.4.KYNU,ZBED2 and CXCL10 were highly expressed in pancreatic cancer and lowly expressed in normal tissues,and the differences were statistically significant(p <0.05).Immunohistochemistry further confirmed that KYNU was highly expressed in PAAD and lowly expressed in adjacent tissues(p=1.7e-5).5.KYNU,ZBED2 and CXCL10 are differentially expressed not only in pancreatic cancer,but also in other cancers,and are associated with the prognosis of numerous cancers,suggesting a potential role of the above genes in tumor development.Conclusion:TME is directly related to PAAD prognosis,and the more immune/stromal components in TME,the worse the PAAD prognosis.KYNU,ZBED2 and CXCL10 are PAAD tumor microenvironment independent prognostic genes that are differentially expressed in PAAD.The high expression of KYNU,ZBED2 and CXCL10 indicates the poor prognosis of pancreatic cancer.The risk scores of KYNU,ZBED2 and CXCL10 can accurately evaluate the prognosis of patients with pancreatic cancer.In summary,KYNU,ZBED2 and CXCL10 are independent prognostic genes associated with the tumor microenvironment in pancreatic adenocarcinoma,are differentially expressed in pancreatic adenocarcinoma,and may be potential biomarkers for pancreatic adenocarcinoma.