Integrative single-cell and bulk transcriptomics define polyamine-associated cell states in acute myeloid leukemia...

Polyamine metabolism is an important metabolic dependency in cancer, but its cellular organization and clinical relevance in acute myeloid leukemia (AML) remain unclear.

We integrated publicly available single-cell RNA sequencing (scRNA-seq) data from 41 AML samples (198,846 cells) and bulk transcriptomic profiles from 712 AML patients across independent cohorts. Polyamine activity was quantified using a 76-gene signature. Consensus clustering and machine-learning algorithms were used for molecular classification and prognostic modeling. Immune features and gene dependency were further evaluated. CCT6A was further validated by quantitative real-time PCR (qPCR), small interfering RNA (siRNA) silencing, cell viability and apoptosis assays, Western blotting, intracellular polyamine and ODC1 activity measurements, and spermidine rescue.

Polyamine activity was preferentially enriched in AML blasts and associated with increased stemness, proliferation, and DNA damage response, together with impaired differentiation. Two molecular subtypes with distinct metabolic and immune features were identified. The Random Survival Forest model effectively stratified patients into prognostic risk groups, achieving external-validation area under the curve (AUC) values up to 0.774. High-risk patients exhibited distinct immune and stromal microenvironmental features. CCT6A was overexpressed in AML samples, and its silencing suppressed cell viability, promoted apoptosis, and reduced ODC1 activity and intracellular polyamine levels. Importantly, exogenous spermidine substantially restored cell viability following CCT6A knockdown.

Polyamine-active AML blast states have prognostic and immune relevance, while CCT6A is functionally linked to AML cell fitness and polyamine homeostasis.