Integrating transcriptomics and machine learning to predict ferroptosis-related genes and analyzing the role of GOT1 in gastric cancer progression

Abstract

Background: Gastric cancer (GC) is one of the leading causes of cancer-related mortality worldwide. Despite advances in treatment, the prognosis for advanced GC remains poor, highlighting the need for new therapeutic targets. Ferroptosis, a form of programmed cell death characterized by iron-dependent lipid peroxidation, has emerged as a potential pathway for cancer therapy. This study aims to identify ferroptosis-related genes in GC using transcriptomics and machine learning, and to validate the role of one key gene, Glutamic-Oxaloacetic Transaminase 1 (GOT1), in GC progression.

Methods: The Sangerbox platform was employed to analyze differentially expressed genes between GC tissues and adjacent non-cancerous tissues in the GSE184336 dataset. Weighted Gene Co-expression Network Analysis (WGCNA) was used to identify co-expressed gene modules from the GSE184336 dataset and their associations with phenotypes. Support Vector Machine (SVM) and Random Forest (RF) regression algorithms were applied to screen for core target genes. Quantitative Real-Time PCR (qRT-PCR) was used to analyze mRNA expression, while western blotting and immunohistochemistry (IHC) assays were used to determine protein expression. Cell proliferation was assessed using a 5-Ethynyl-2'-deoxyuridine assay, and cell migration was analyzed using a Transwell assay. Flow cytometry was used to quantify cell death. Fluorometric and colorimetric assays were performed to analyze Reactive Oxygen Species (ROS) and Fe^{2 +} levels, respectively. A xenograft mouse model was used to evaluate the impact of GOT1 overexpression on tumor formation in vivo. Haematoxylin and eosin (HE) staining was used to analyze the pathological conditions of tumors.

Results: The analysis of the GSE184336 dataset identified 2297 dysregulated genes in GC tissues compared to adjacent non-cancerous tissues. WGCNA revealed strong correlations between gene significance and module membership in the "green," "ivory," and "lightsteelblue1" modules, encompassing 350 genes. Subsequent analysis identified 14 intersection genes among the 2297 dysregulated genes in GC tissues, the 350 genes from WGCNA, and the 1467 genes related to ferroptosis. Machine learning algorithms and protein-protein interaction analysis identified IDH2, BGN, IGFBP7, and GOT1 as key genes. GOT1 expression was downregulated in GC tissues with the lowest error value. Overexpression of GOT1 inhibited GC cell proliferation, migration, and induced ferroptosis, whereas these effects were reversed by treatment with Fer-1. Further, GOT1 overexpression suppressed the malignant phenotype of GC cells in vivo.

Conclusion: This study identified several ferroptosis-related genes in GC, with GOT1 being a critical regulator. Overexpression of GOT1 significantly inhibited GC cell proliferation and migration, and induced ferroptosis. These findings suggest that GOT1 could serve as a potential therapeutic target for GC treatment.

Keywords: Ferroptosis; Gastric cancer; Glutamic-oxaloacetic transaminase 1; Machine learning; Transcriptomics.