Integrative analysis for identification of key miRNA-mRNA regulatory axes in esophageal cancer and preliminary validation of the regulatory role of miR-15b-5p/ BTG2 therein

Background Esophageal cancer (ESCA), a leading cause of cancer-related mortality, lacks reliable biomarkers for early detection and prognosis. Dysregulated microRNAs (miRNAs) have emerged as pivotal regulators of tumor progression, yet their context-specific roles and interactions with target genes in ESCA remain underexplored. Methods Multi-omics data from The Cancer Genome Atlas-esophageal cancer (TCGA-ESCA) and Gene Expression Omnibus (GEO) datasets were integrated to identify differentially expressed miRNAs and mRNAs. A miRNA-mRNA regulatory network was constructed using FunRich and validated through functional assays, including dual-luciferase reporter, quantitative reverse transcription polymerase chain reaction (qRT-PCR) and in vitro proliferation/migration/invasion experiments. Prognostic signatures were developed using Cox regression, least absolute shrinkage and selection operator (LASSO)-Cox and nomogram analysis. Results We identified 1,131 differentially expressed mRNAs and 69 miRNAs in ESCA. The miR-15b-5p/ BTG2 axis emerged as a central regulatory hub. miR-15b-5p was significantly upregulated in ESCA tissues and showed an inverse correlation with B-cell translocation gene 2 ( BTG2 ) expression. Survival analyses established both molecules as independent prognostic factors. Mechanistically, miR-15b-5p directly targeted BTG2 3′UTR, suppressing its expression. Functional studies demonstrated that miR-15b-5p overexpression promoted proliferation, migration and invasion in ESCA cells, whereas BTG2 restoration reversed these effects. A prognostic nomogram integrating miR-15b-5p, BTG2 and clinical parameters demonstrated robust predictive accuracy (C-index: 0.78). Conclusions The miR-15b-5p/ BTG2 axis represents a novel regulatory mechanism in ESCA progression with significant potential as both a prognostic biomarker and therapeutic target.