A m6A/m1A/m5C-related eight-gene signature predicts prognosis and correlates with immune microenvironment in pediatric acute myeloid leukemia.
Using public RNA-sequencing data, the study developed and dataset-validated an eight-gene RNA-methylation-related signature associated with survival, immune-cell features, tumor purity, and predicted immunotherapy response in pediatric acute myeloid leukemia.
Open original publication →What the AI sees
Using public RNA-sequencing data, the study developed and dataset-validated an eight-gene RNA-methylation-related signature associated with survival, immune-cell features, tumor purity, and predicted immunotherapy response in pediatric acute myeloid leukemia.
Research significance
The reported evidence supports prognostic associations and correlations with the immune microenvironment; it only raises the untested hypothesis that RNA-methylation genes or the resulting risk score could guide immunotherapy selection or reveal therapeutic targets in pediatric AML.
Source abstract
BACKGROUND: N6-methyladenosine (m6A), 5-methylcytosine (m5C) and N1-methyl-adenosine (m1A) are among predominate forms of RNA methylations implicated in the progression of cancer. The study aims to dissect the associations of m6A/m1A/m5C genes with survival, tumor microenvironment (TME) and predicted immunotherapy response in pediatric patients with acute myeloid leukemia (AML). METHODS: RNA sequencing (RNA-seq) data were downloaded from Xena database. Differentially expressed m6A/m1A/m5C genes were filtered between dead and alive samples. Least absolute shrinkage and selection operator (LASSO) cox regression analysis was used to identify optimal prognostic genes and construct a risk-score model. TME characteristics and immunotherapy response were analyzed across different risk subgroups. RESULTS: Twenty-five differentially expressed m6A/m1A/m5C genes were identified. Eight optimal prognostic genes were selected via LASSO and an 8-gene-based prognostic model was built. It effectively distinguished high- from low-risk patients [survival difference: P value =6.328e-07, area under the curve (AUC) value =0.861 (0.820, 0.778)], with valid verification in a validation dataset [survival difference: P value =2.677e-02, AUC value =0.774 (0.625, 0.833)]. The risk score was an independent prognostic factor and a composite nomogram could more accurately predict survival probability. Low-risk patients had significantly higher levels of monocytes and lower tumor purity compared to high-risk patients. A trend toward differential risk scores was observed between predicted immunotherapy responders and non-responders. ZC3H13, NSUN2, METTL16, TRMT10C and DNMT3B, and TRMT6 possessed stronger correlations with tumor purity. CONCLUSIONS: We propose an 8-gene prognostic signature related to m6A/m5C/m1A and a risk-score model for assessing prognosis in pediatric AML. These genes may serve as promising preliminary prognostic biomarkers and warrant further exploration as potential immunotherapy targets for pediatric AML, pending validation in larger-scale prospective cohorts.