Identification of ferroptosis genes in immune infiltration and prognosis in thyroid papillary carcinoma using network analysis

• Published in: BMC genomics 2021-07, Vol.22 (1), p.1-576, Article 576
• Main Authors: Lin, Ruoting, Fogarty, Conor E, Ma, Bowei, Li, Hejie, Ni, Guoying, Liu, Xiaosong, Yuan, Jianwei, Wang, Tianfang
• Format: Article
• Language: English
• Subjects: Algorithms; Apolipoprotein E; Apoptosis; Cancer; Cancer therapies; Cell death; Cellular control mechanisms; Clusters; Correlation analysis; Datasets; Enrichment; Ferroptosis; Ferroptosis genes; Gene expression; Gene set enrichment analysis; Genes; Genetic aspects; Genomes; Genomics; Immune system; Immune-related genes; Infiltration; Lipids; Medical prognosis; Metastases; Microenvironments; Multivariate analysis; Mutation; Network analysis; Oncology, Experimental; Papillary thyroid cancer; Papillary thyroid carcinoma; Parathyroid; Parathyroid hormone; Patients; Performance evaluation; Peroxides; Principal components analysis; Prognosis; Prognostic model; Regression models; Risk; Survival; Survival analysis; Thyroid; Thyroid cancer; Thyroid papillary carcinoma; Tumors; Tumour microenvironment
• ISSN: 1471-2164; 1471-2164
• DOI: 10.1186/s12864-021-07895-6

Papillary thyroid carcinoma (PTC) is the most common thyroid cancer. While many patients survive, a portion of PTC cases display high aggressiveness and even develop into refractory differentiated thyroid carcinoma. This may be alleviated by developing a novel model to predict the risk of recurrence. Ferroptosis is an iron-dependent form of regulated cell death (RCD) driven by lethal accumulation of lipid peroxides, is regulated by a set of genes and shows a variety of metabolic changes. To elucidate whether ferroptosis occurs in PTC, we analyse the gene expression profiles of the disease and established a new model for the correlation. The thyroid carcinoma (THCA) datasets were downloaded from The Cancer Genome Atlas (TCGA), UCSC Xena and MisgDB, and included 502 tumour samples and 56 normal samples. A total of 60 ferroptosis related genes were summarised from MisgDB database. Gene set enrichment analysis (GSEA) and Gene set variation analysis (GSVA) were used to analyse pathways potentially involving PTC subtypes. Single sample GSEA (ssGSEA) algorithm was used to analyse the proportion of 28 types of immune cells in the tumour immune infiltration microenvironment in THCA and the hclust algorithm was used to conduct immune typing according to the proportion of immune cells. Spearman correlation analysis was performed on the ferroptosis gene expression and the correlation between immune infiltrating cells proportion. We established the WGCNA to identify genes modules that are highly correlated with the microenvironment of immune invasion. DEseq2 algorithm was further used for differential analysis of sequencing data to analyse the functions and pathways potentially involving hub genes. GO and KEGG enrichment analysis was performed using Clusterprofiler to explore the clinical efficacy of hub genes. Univariate Cox analysis was performed for hub genes combined with clinical prognostic data, and the results was included for lasso regression and constructed the risk regression model. ROC curve and survival curve were used for evaluating the model. Univariate Cox analysis and multivariate Cox analysis were performed in combination with the clinical data of THCA and the risk score value, the clinical efficacy of the model was further evaluated. We identify two subtypes in PTC based on the expression of ferroptosis related genes, with the proportion of cluster 1 significantly higher than cluster 2 in ferroptosis signature genes that are positively associated. The