Integration of transcriptomic data reveals lipid metabolic heterogeneity and identifies GSTO1 as a therapeutic target in acute myeloid leukemia.
Acute myeloid leukemia (AML) is an aggressive hematologic malignancy with poor prognosis and significant heterogeneity. Lipid metabolic reprogramming is a key hallmark of cancer, yet its systemic characterization and clinical relevance in AML remain largely unexplored.
Multi-omics data were integrated, including one single-cell RNA-seq dataset and bulk transcriptomes from nine AML cohorts. Lipid metabolism activity was assessed using GSVA. Consensus clustering based on lipid metabolism pathways identified molecular subtypes. A lipid metabolism-related prognostic signature (LMRS) was constructed via machine learning algorithms and validated across nine independent cohorts. Functional validation was performed in AML cell lines using GSTO1 inhibition.
Single-cell analysis revealed significant upregulation of lipid metabolism pathways in AML malignant cells, particularly in progenitor-like subpopulations. Three lipid metabolism-based subtypes (C1-C3) were identified, with the C3 subtype exhibiting the highest metabolic activity, an immunosuppressive microenvironment, and the worst prognosis. A robust nine-gene LMRS model was developed, which effectively stratified patients into high- and low-risk groups with distinct survival outcomes. LMRS demonstrated superior predictive accuracy over existing models, was independently prognostic, and correlated with chemotherapy and immunotherapy resistance. Inhibition of GSTO1 significantly induced apoptosis and ROS production in AML cells.
This study comprehensively defines lipid metabolic heterogeneity in AML, establishes a clinically applicable prognostic signature, and underscores lipid metabolism as a key driver of AML progression and immunosuppression. Targeting lipid metabolism, particularly through GSTO1 inhibition, represents a promising therapeutic strategy.
Multi-omics data were integrated, including one single-cell RNA-seq dataset and bulk transcriptomes from nine AML cohorts. Lipid metabolism activity was assessed using GSVA. Consensus clustering based on lipid metabolism pathways identified molecular subtypes. A lipid metabolism-related prognostic signature (LMRS) was constructed via machine learning algorithms and validated across nine independent cohorts. Functional validation was performed in AML cell lines using GSTO1 inhibition.
Single-cell analysis revealed significant upregulation of lipid metabolism pathways in AML malignant cells, particularly in progenitor-like subpopulations. Three lipid metabolism-based subtypes (C1-C3) were identified, with the C3 subtype exhibiting the highest metabolic activity, an immunosuppressive microenvironment, and the worst prognosis. A robust nine-gene LMRS model was developed, which effectively stratified patients into high- and low-risk groups with distinct survival outcomes. LMRS demonstrated superior predictive accuracy over existing models, was independently prognostic, and correlated with chemotherapy and immunotherapy resistance. Inhibition of GSTO1 significantly induced apoptosis and ROS production in AML cells.
This study comprehensively defines lipid metabolic heterogeneity in AML, establishes a clinically applicable prognostic signature, and underscores lipid metabolism as a key driver of AML progression and immunosuppression. Targeting lipid metabolism, particularly through GSTO1 inhibition, represents a promising therapeutic strategy.