Construction of a New Ferroptosis-related Prognosis Model for Survival Prediction in Colorectal Cancer

Aim: This study was designed to develop a ferroptosis-related gene signature for guiding the prognostic prediction in colorectal cancer (CRC) and to explore the potential in the molecular functions of the gene signature. Background: Ferroptosis is mainly characterized by lipid peroxide accumulation on the cell membranes in an iron-dependent manner, resulting in cellular oxidative stress, metabolic disorders, and, ultimately, cell death. This study aimed to develop a prognostic ferroptosis signature in CRC and explore its potential molecular function. Objective: The present work was designed to devise a ferroptosis signature for CRC prognosis and explore its potential molecular function. Methods: Single-cell RNA sequencing data GSE161277 and transcriptome sequencing data GSE17537 and TCGA-CRC from the Gene Expression Omnibus (GEO) and the Cancer Genome Atlas (TCGA) databases were downloaded, respectively. Quality control, dimension reduction, clustering, and clustering of single-cell RNA sequencing (scRNA- seq) data were performed using the Seurat package. A total of 259 ferroptosis-correlated genes from the FerrDb database were acquired. The single sample gene set enrichment analysis (ssGSEA) was performed to calculate the scores of genes related to ferroptosis. ESTIMATE was used to calculate immune infiltration. Independent prognostic factors were determined by performing Weighted Gene Co-Expression Network Analysis (WGCNA), univariate and Cox analyses, and Lasso analyses were used to search for independent prognostic factors. Results: From the scRNA-seq (GSE161277) dataset, 22 cell clusters were initially identified, and according to immune cell markers, only 8 types of cells (Follicular B, central memory T cell, Epithelial, Natural killer T cell, Plasma B, M1 macrophage, Fibroblasts, and Mast cell) were finally determined to be related to CRC prognosis. The results of the scRNA-seq analysis showed that the score of ferroptosis-related genes was higher in tumour tissues and in 8 types of cells in tumour samples. In the TCGA dataset, CRC samples were divided into ferroptosis-related high scores, ferroptosis-related median scores, and ferroptosis-related low scores. Immune cell analysis revealed that ferroptosis- related high scores had the highest abundance of immune cells. An 11-gene signature was developed by WGCNA, univariate Cox, and Lasso Cox regression. The prediction ability of the signature was successfully validated in the GSE17537 dataset. A comprehensive nomogram combining the 11 signature genes and clinical parameters could effectively predict the overall survival of CRC patients. Conclusions: The present molecular signature established based on the 11 ferroptosis-related genes performed well in assessing CRC prognosis. The present discoveries could inspire further research on ferroptosis, providing a new direction for CRC management.