Identification of SLC15A3 as a shared prognostic biomarker for AML and COVID-19 via integrative bioinformatics, machine learning and Mendelian randomization

Abstract

Acute myeloid leukemia (AML) is a lethal hematological malignancy with poor overall prognosis.COVID-19 co-infection further complicates clinical management and worsens survival outcomes of AML patients, highlighting the need to explore the molecular interplay between the two diseases. In this study, we initially screened 109 common differentially expressed genes shared by AML and COVID-19 from public transcriptomic datasets, which were closely involved in tumor progression and viral pathogenesis. We further characterized their biological functions, constructed protein interaction and TF-miRNA regulatory networks, and established an AML prognostic model to evaluate the tumor immune microenvironment. Combining survival analysis, clinical correlation, immune infiltration assessment and m 6 A modification analysis, SLC15A3 was preliminarily screened as a candidate prognostic gene for AML. Subsequent weighted gene co-expression network analysis (WGCNA), machine learning (ML) and Mendelian randomization (MR) analyses further identified SLC15A3 as a promising prognostic biomarker for AML. Finally, RT-qPCR and Western blotting assays verified its aberrant expression in AML cells. Consistent with bioinformatic prediction, SLC15A3 serves as a shared candidate biomarker for AML and COVID-19,although its mechanistic role linking AML and SARS-CoV-2-related inflammation requires further functional validation.