Comprehensive Characterization Of Immunogenic Cell Death In Melanoma

Background:With the in-depth investigation of immunogenic cell death(ICD),increasing evidence has highlighted its critical functions within many tumors.In this study,ICD-related treatment modality for melanoma was developed after investigating the characterization of ICD and its prognostic role in melanoma.Methods:In this research,melanoma samples were acquired from The Cancer Genome Atlas(TCGA)and the Gene Expression Omnibus(GEO)databases.In addition,TCGA database was searched to acquire m RNA expression,copy number variation(CNV),single nucleotide variation(SNV)and methylation data of pan-cancer.34 ICD-related genes were summarized according to a literature.Firstly,the alterations of ICD-related genes of the expression levels,prognostic values,mutation information,and methylation levels were investigated in multiple cancers.Subsequently,the impacts of ICD on the prognosis and immune microenvironment of melanoma were specifically explored.On the basis of the expression of ICD-related genes,TCGA melanoma samples were separated into two subtypes(i.e.C1 and C2)with different prognosis and immune microenvironment utilizing the non-negative matrix factorization(NMF)clustering.Based on the differentially expressed genes(DEGs)between the two subtypes,LASSO-Cox regression analysis was conducted to obtain an ICD-dependent risk signature(ICDRS).To confirm the accuracy of the conclusion,TCGA and GEO samples were separated and train(50% TCGA samples),test1(50% TCGA samples),test2(100% TCGA samples)and test3(100% GEO samples)cohorts were developed.After obtaining the risk score of each sample based on the ICDRS,samples in the four cohorts were labeled with high risk or low risk.Then the following investigations and validations were conducted:(1)principal component analysis(PCA)and t-distributed stochastic neighbor embedding(t-SNE)analysis for the discrimination of high-and low-risk samples;(2)survival analysis among samples with all risk levels;(3)comparison with the instructed signatures for the estimation of the diagnostic value of the ICDRS;(4)collecting the m RNA expression of the ICD-related genes and immune checkpoint genes(ICGs);(5)estimation of the proportion of immune cells by CIBERSORT;(6)assessment of the activities if immune-related pathways by ss GSEA;(7)the Pearson correlation analysis to illustrate the correlation between risk scores and the expression levels of ICD-related genes and ICGs,immune-related pathway activities and the proportion of immune cells.Additionally,the Cancer Immunome Atlas(TCIA)were utilized to predict immunotherapy response in high-and low-risk subgroups.After obtaining the DEGs between the two subgroups,the Connectivity map(CMap)were utilized for drug prediction.Results:The expression levels,SNV,CNV,and methylation of ICD-related genes in tumor tissues statistically differed from normal tissues in pan-cancer.In melanoma,two subtypes with different prognosis and immune microenvironment were identified by the NMF clustering.Subsequently,an ICDRS was constructed to identify the risk score of each sample according to the expression levels of GBP2,THBS4,and APOBEC3 G.Those high-risk samples were combined with poor prognosis,low proportion of M1 macrophage,high proportion of M2 macrophage,decreasing expression of ICGs and ICD-related genes,and low activities of immune-related pathways.In addition,the immunotherapy response prediction suggested that samples in the high-risk subgroup had worse immunotherapy responses.The drug prediction indicated that axitinib,imatinib,sorafenib,sunitinib,and cinobufagin might be helpful for melanoma patients.Conclusions:This study indicated the alterations of the expression levels,SNV,CNV,and methylation of ICD-related genes in pan-cancer.A novel ICDRS was identified successfully and helped to illustrate the influence of ICD on melanoma.The signature could function to determine the prognosis and tumor immune microenvironment and aid in patient classification for individualized treatment in melanoma.