Prognostic prediction and immune infiltration analysis based on ferroptosis and EMT state in hepatocellular carcinoma

Abstract

BackgroundFerroptosis is one of the main mechanisms of sorafenib against hepatocellular carcinoma (HCC). Epithelial-mesenchymal transition (EMT) plays an important role in the heterogeneity, tumor metastasis, immunosuppressive microenvironment, and drug resistance of HCC. However, there are few studies looking into the relationship between ferroptosis and EMT and how they may affect the prognosis of HCC collectively.MethodsWe downloaded gene expression and clinical data of HCC patients from the Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC) databases for prognostic model construction and validation respectively. The Least absolute shrinkage and selection operator (LASSO) Cox regression was used for model construction. The predictive ability of the model was assessed by Kaplan–Meier survival analysis and receiver operating characteristic (ROC) curve. We performed the expression profiles analysis to evaluate the ferroptosis and EMT state. CIBERSORT and single-sample Gene Set Enrichment Analysis (ssGSEA) methods were used for immune infiltration analysis.ResultsA total of thirteen crucial genes were identified for ferroptosis-related and EMT-related prognostic model (FEPM) stratifying patients into two risk groups. The high-FEPM group had shorter overall survivals than the low-FEPM group (p<0.0001 in the TCGA cohort and p<0.05 in the ICGC cohort). The FEPM score was proved to be an independent prognostic risk factor (HR>1, p<0.01). Furthermore, the expression profiles analysis suggested that the high-FEPM group appeared to have a more suppressive ferroptosis status and a more active EMT status than the low- FEPM group. Immune infiltration analysis showed that the myeloid-derived suppressor cells (MDSCs), and regulatory T cells (Tregs) were highly enriched in the high-FEPM group. Finally, a nomogram enrolling FEPM score and TNM stage was constructed showing outstanding predictive capacity for the prognosis of patients in the two cohorts.ConclusionIn conclusion, we developed a ferroptosis-related and EMT-related prognostic model, which could help predict overall survival for HCC patients. It might provide a new idea for predicting the response to targeted therapies and immunotherapies in HCC patients.

Bidirectional crosstalk between the epithelial-mesenchymal transition and immunotherapy: A bibliometric study

“…Establishing gene signatures with EMT based on The Cancer Genome Atlas (TCGA) or the Gene Expression Omnibus (GEO) database can provide a basis for the evaluation of prognosis and the development of novel targets for immunotherapy. 25 , 72 Overall, all the topics identified as hotspots in recent years were related to the mechanism of EMT in immunotherapy and the development of therapeutic strategies for improving immunotherapeutic approaches for cancer treatment.…”

Identification of biomarkers co-associated with M1 macrophages, ferroptosis and cuproptosis in alcoholic hepatitis by bioinformatics and experimental verification

“…In previous studies, a prognostic model of hepatocellular carcinoma was established using WGCNA analysis of macrophage-related genes ( 41 ). The prognostic model based on ferroptosis and epithelial-mesenchymal transition state helps predict overall survival of hepatocellular carcinoma ( 42 ). Cuproptosis-related subtypes predict tumor microenvironments and drug candidates in hepatocellular carcinoma ( 43 ).…”

Sensitivity Analysis for Survival Prognostic Prediction with Gene Selection: A Copula Method for Dependent Censoring

“…Usually, a multi-gene predictor is more accurate than a single-gene predictor as the prognostic ability of a single gene is limited. A predictor optimally constructed from the selected genes becomes a useful biomarker for predicting survival in breast cancer [ 10 , 11 , 12 , 13 , 14 , 15 ], lung cancer [ 6 , 7 , 8 , 9 , 16 , 17 ], gastric cancer [ 18 , 19 ], ovarian cancer [ 20 , 21 , 22 , 23 , 24 ], skin cancer [ 25 ], liver cancer [ 26 , 27 ], bladder cancer [ 28 ], head and neck cancer [ 29 , 30 ], glioma [ 31 ], myeloproliferative neoplasms [ 32 ], kidney cancer [ 33 ], and cancers of mixed types [ 34 , 35 ]. These analyses were performed mostly based on univariate Cox regression with the significance scaling of p -values, such as 0.05, 0.01, and 0.001, followed by cross-validation and/or external validation.…”