| Objective: The Cancer Genome Atlas（TCGA）was used to explore the genomic,transcriptomic,and proteomic profiles of chemokine family members that changed in pancreatic adenocarcinoma（PDAC）.We also defined the chemokine expression profiles that affect the overall survival（OS）of PDAC patients by unsupervised Hierarchical Clustering（HC）.In turn,a chemokine-related gene prognostic model was established.The local immune profile of PDAC patients was also investigated in order to further elucidate the association between chemokine family expression profile and local immune profile in PDAC patients.Methods:Part Ⅰ: Chemokine expression profiles affecting overall patient survival in the TCGA pancreatic cancer data setGEPIA2 was applied to analyze the differences in RNA expression of 44 members of the chemokine family in pancreatic cancer and healthy controls.The impact of full chemokine family members in PAAD on survival was analyzed by Surv Express,while their genomic alterations were extracted using tools from c Bio Portal.Pearson correlation coefficients（r-values）and p values corresponding to chemokine gene combinations were computed by parsing PAAD transcriptional data and clinical data in RStudio.Both principal component analysis（PCA）and unsupervised hierarchical clustering（HC）were carried out with Facto Miner R-package upon the PAAD cohort.Survival analysis of the patient cohort was accomplished by utilizing the software packages Survival and Surv Miner.Finally,to investigate the protein level expression of different chemokines in human pancreatic cancer tissue samples,we retrieved immunohistochemical data of chemokine genes from the Human Protein Atlas（HPA）on tissue arrays of PDAC samples.Part Ⅱ: Prognostic modeling and immune characterization of chemokine-related genesAnalysis of Differential Expressed Genes（DEGs）was undertaken using the DESeq2 package to characterize the separate clusters of TCGA pancreatic cancer chemokines.The DEGs were also subjected to gene ontology（GO）and genomic enrichment analysis（GSEA）to determine the tumor features of each patient cluster.Genes physically and functionally associated with chemokines were depicted by systematic co-expression analysis of STRING and Msig DB literature mining.Feature genes were screened from chemokines and their related genes by LASSO regression.The univariate and multivariate Cox regression analyses of the signature genes were also performed in the TCGA-PAAD cohort to establish a prognostic risk model.Immune-related features and tumor purity were assessed in PDAC patients by the estimate package.The TIMER2.0database was introduced to analyze the infiltration of immune cells and stromal cells in the tumor microenvironment to examine their potential association with chemokine family expression.Finally,we verified the expression of chemokine-related genes in pancreatic cancer cell lines by quantitative real-time fluorescence PCR（qRT-PCR）.Results:Part Ⅰ: Chemokine expression profiles affecting overall patient survival in the TCGA pancreatic cancer data set1.In the expression analysis of each chemokine family member in PDAC,we found that CCL2,CCL3,CCL4,CCL5,CCL11,CCL13,CCL15,CCL17,CCL18,CCL19,CCL20,CCL21,CCL22,CCL24,CCL26,CCL28,CXCL1,CXCL3,CXCL4,CXCL5,CXCL6,CXCL8,CXCL9,CXCL10,CXCL13,CXCL14,CXCL16,CXCL17,and CX3CL1 were upregulated in pancreatic cancer,and other chemokine family genes did not show significant expression differences.2.In the analysis of the effect of full chemokine family members on survival in PDAC patients,patients with significantly high expression of CCL5,CCL18,CCL20,CCL28,CXCL3,CXCL5,CXCL8,CXCL9,CXCL10,CXCL11,CXCL17,CCL21,CCL22,CXCL1,CXCL6,CXCL16,CX3CL1 showed shorter overall survival,and conversely,patients with high expression of CCL3 and CCL14 showed a better prognosis.3.In the genomic alteration analysis of chemokines in the TCGA-PAAD cohort,there was no significant difference in overall survival and disease-free survival（DFS）between the group with and without genomic alteration of chemokines.However,18.12%of PDAC samples（27/149）had at least one chemokine genomic alteration event,with CCL1,CCL11 and CCL27 being the most frequently altered genes in PDAC patients（all 8%）.4.In the Pearson correlation coefficient analysis of chemokine combinations,we identified 635 chemokine gene combinations with positive correlation（0.15 |
PDF Full Text Request