Abstract
INTRODUCTION: Gastric cancer (GC) is a highly heterogeneous malignancy with poor prognosis, underscoring the urgent need for reliable biomarkers to guide precise stratification and therapy. Transfer RNA-derived small RNAs (tsRNAs) have emerged as potential key regulators in cancer, yet their systematic role in defining GC subtypes remains unexplored. METHODS: We profiled tsRNA expression in GC using transcriptomic data from TCGA and GEO databases. Unsupervised consensus clustering identified tsRNA-based subtypes. A prognostic model was constructed using machine learning algorithms and validated across multiple cohorts. The functional role of a key tsRNA, tsRNA-Asp-3-0024, was investigated through Pandora-seq, qRT-PCR, and in vitro and organoid-based assays. RESULTS: Three distinct tsRNA-mediated subtypes (Stromal_H, Stromal_L, Stromal_M) were identified, exhibiting significant differences in stromal activity, tumor microenvironment, and clinical outcomes. The Stromal_H subtype demonstrated the poorest prognosis, characterized by an immunosuppressive microenvironment and dysregulated DNA repair pathways. A random survival forest (RSF)-based prognostic signature (GCtsRNAscore) effectively stratified patients into high- and low-risk groups, with high-risk patients showing increased sensitivity to targeted therapies (axitinib, bexarotene, dasatinib) and low-risk patients benefiting more from immunotherapy. Furthermore, tsRNA-Asp-3-0024 was significantly upregulated in GC tissues and cell lines, where it promoted proliferation and inhibited apoptosis. DISCUSSION: Our study establishes tsRNAs as powerful biomarkers for molecular subtyping and prognostic prediction in GC. The tsRNA-defined subtypes and GCtsRNAscore model provide a novel framework for personalized treatment strategies. The functional characterization of tsRNA-Asp-3-0024 highlights its potential as both a therapeutic target and a prognostic indicator, paving the way for tsRNA-based precision medicine in GC.