Abstract

Increasing evidence suggests that disulfidptosis plays a crucial role in tumorigenesis and progression. This study aimed to identify biomarkers closely associated with disulfidptosis in anaplastic thyroid carcinoma (ATC). Utilizing ATC-related datasets (GSE65144, GSE9115, GSE27155, and GSE53072) in conjunction with disulfide bond-related genes (DRGs) identified in the literature, differentially expressed genes (DEGs) were screened from the GSE65144 and GSE9115 datasets. A total of 113 common DEGs were identified through cross-sectional analysis. Weighted gene co-expression network analysis (WGCNA) was employed to screen genes related to disulfidptosis and ATC, and five biomarkers—ATP1B3, TFF3, LGALS1, ADAM12, and COL1A2—were identified using machine learning algorithms. A nomogram model constructed based on these markers demonstrated high accuracy. In vitro validation revealed that ATP1B3 knockdown significantly inhibited tumor growth, indicating its potential anti-ATC activity. Furthermore, laser confocal microscopy, flow cytometry, and other experimental methods suggested a correlation between ATP1B3 and disulfidptosis. These findings highlight ATP1B3, TFF3, LGALS1, ADAM12, and COL1A2 as potential disulfidptosis-related biomarkers in ATC. This study provides a theoretical foundation for understanding the role of disulfidptosis in ATC pathogenesis and suggests that ATP1B3 may serve as a promising therapeutic target.