The therapeutic and biomarker significance of ferroptosis in chronic myeloid leukemia

Fangmin Zhong,Xueru Zhang,Zihao Wang,Xiaolin Li,Bo Huang,Guangyao Kong,Xiaozhong Wang

DOI: https://doi.org/10.3389/fimmu.2024.1402669

2024-07-04

Abstract:Background: The relationship between ferroptosis and the progression and treatment of hematological tumors has been extensively studied, although its precise association with chronic myeloid leukemia (CML) remains uncertain. Methods: Multi-transcriptome sequencing data were utilized to analyze the ferroptosis level of CML samples and its correlation with the tumor microenvironment, disease progression, and treatment response. Machine learning algorithms were employed to identify diagnostic ferroptosis-related genes (FRGs). The consensus clustering algorithm was applied to identify ferroptosis-related molecular subtypes. Clinical samples were collected for sequencing to validate the results obtained from bioinformatics analysis. Cell experiments were conducted to investigate the therapeutic efficacy of induced ferroptosis in drug-resistant CML. Results: Ferroptosis scores were significantly lower in samples from patients with CML compared to normal samples, and these scores further decreased with disease progression and non-response to treatment. Most FRGs were downregulated in CML samples. A high ferroptosis score was also associated with greater immunosuppression and increased activity of metabolic pathways. Through support vector machine recursive feature elimination (SVM-RFE), least absolute shrinkage selection operator (LASSO), and random forest (RF) algorithms, we identified five FRGs (ACSL6, SLC11A2, HMOX1, SLC38A1, AKR1C3) that have high diagnostic value. The clinical diagnostic value of these five FRGs and their effectiveness in differentiating CML from other hematological malignancies were validated using additional validation cohorts and our real-world cohort. There are significant differences in immune landscape, chemosensitivity, and immunotherapy responsiveness between the two ferroptosis-related molecular subtypes. By conducting cellular experiments, we confirmed that CML-resistant cells are more sensitive to induction of ferroptosis and can enhance the sensitivity of imatinib treatment. Conclusion: Our study unveils the molecular signature of ferroptosis in samples from patients with CML. FRG identified by a variety of machine learning algorithms has reliable clinical diagnostic value. Furthermore, the characterization of different ferroptosis-related molecular subtypes provides valuable insights into individual patient characteristics and can guide clinical treatment strategies. Targeting and inducing ferroptosis holds great promise as a therapeutic approach for drug-resistant CML.