Fucosyltransferase 4 as a candidate diagnostic biomarker and potential therapeutic target in osteosarcoma: a multi-dataset transcriptomic and structural bioinformatics analysis

Abstract Background Osteosarcoma, the most common primary malignant bone tumor in children and young adults, has a poor prognosis in metastatic or recurrent disease. Fucosyltransferase 4 ( FUT4 ) mediates fucosylation and has been implicated in several cancers, but its role in osteosarcoma remains unclear. Methods Using two public Gene Expression Omnibus microarray cohorts (GSE36001, GSE16088), we assessed FUT4 expression and diagnostic performance by limma-based differential analysis and ROC curves with AUC estimation, and profiled function through GO, KEGG, GSEA and GSVA. We further constructed protein-protein interaction and transcription factor/microRNA/drug regulatory networks, characterized immune-cell signatures by ssGSEA, and prioritized candidate binding compounds by DrugCLIP virtual screening followed by AutoDock Vina docking. Results FUT4 transcripts were significantly down-regulated in osteosarcoma in both cohorts (GSE36001, p < 0.05; GSE16088, p < 0.01), with moderate-to-high discriminatory capacity, although the high AUC in GSE16088 may be confounded by non-osseous control tissues. FUT4 -associated genes were enriched in ossification, osteoblast differentiation and extracellular matrix-related terms. GSEA linked high FUT4 to epithelial-mesenchymal transition and inflammatory-response signatures, and exploratory GSVA suggested differences in cytoskeletal, apoptotic and Wnt/beta-catenin signalling between FUT4 -low and FUT4 -high subgroups, which also showed distinct immune-cell correlation patterns. Structure-based screening prioritized ST50926609 as the lead candidate FUT4 -binding compound (Vina score = − 7.35 kcal/mol). Conclusions Collectively, these computational findings support FUT4 as a candidate diagnostic biomarker and potential therapeutic target in osteosarcoma, but they remain hypothesis-generating and require validation in clinical cohorts, cell and animal models.