Integrated analysis of programmed cell death-related genes identifies CORO1A as an apoptosis-associated gene in acute myeloid leukemia

Background Programmed cell death (PCD) governs tissue homeostasis and shapes tumor–immune interactions. In acute myeloid leukemia (AML), apoptosis evasion drives resistance and poor outcomes, yet the PCD landscape and the role of Coronin 1A (CORO1A) remain insufficiently defined. Methods We integrated AML bone marrow single-cell RNA-seq (GSE154109) with bulk RNA-seq from TCGA (training) and GSE71014 (testing), and GTEx normal marrow. After batch correction, we profiled 3,524 PCD- related genes. Single-cell analyses (Seurat, Harmony) resolved hematopoietic and malignant populations; pathway activity was quantified by ssGSEA/GSVA. Survival modeling used a multi-algorithm framework (10 learners, 117 combinations). Immune features were inferred via IOBR; immunotherapy response via TIDE; drug sensitivity via OncoPredict. We analyzed DepMap AML cell lines ( n = 29), correlating CORO1A CRISPR dependency (CERES) with CORO1A expression (log 2 TPM) by Pearson’s correlation in R. CORO1A function was tested by shRNA knockdown in HL-60 and THP-1 cells with CCK-8, Annexin V, and Western blot. Results Malignant cells and monocytes/macrophages showed higher PCD enrichment than lymphoid clusters. Malignant cells upregulated MYC/E2F targets, DNA repair, fatty acid metabolism, with attenuated interferon signaling. From 3,524 genes, we derived a five-gene prognostic signature—CORO1A and PECAM1 (risk), and CLEC11A, ITGA4, AGTPBP1 (protective)- via optimal LASSO plus stepwise Cox (mean C-index 0.72; TCGA 0.74; GSE71014 0.70). Discrimination was strong (AUCs: training 0.816/0.803/0.887 at 1/3/5 years; testing 0.711/0.765/0.749) and independent of age and sex. Risk score—associated transcripts were enriched for apoptosis, FoxO signaling, senescence, proteoglycans in cancer, and HTLV-1 pathways; GSEA indicated hematopoietic stemness and NRF2-upregulated programs. High-risk tumors had higher stromal/immune scores, lower purity, increased monocytes/macrophages/neutrophils, and reduced B/T/NK infiltration; the risk score correlated with cytokines (CXCR3, IL10, PF4, PPBP). TIDE predicted greater immunotherapy benefit in high-risk patients. OncoPredict suggested higher sensitivity in the high-risk group to 5-fluorouracil, PI3K-AKT-mTOR inhibitors (afuresertib, pictilisib, taselisib, dactolisib), and the MET inhibitor savolitinib. CORO1A was highly expressed in AML cell lines; knockdown reduced proliferation, decreased Bcl-2/Bax ratio by 1.5-fold, and increased apoptosis. Conclusions Multi-omic integration of PCD-related genes delineates PCD-driven heterogeneity in AML and yields a robust five-gene prognostic model with therapeutic implications. CORO1A emerges as a potential apoptosis-associated oncogene that promoting AML cell survival.