An O-glycosylation-based prognostic signature for biochemical recurrence in prostate cancer identifies GALNTL6 as a potential promoter of malignant phenotypes.

Background: Prostate cancer (PCa) is frequently diagnosed in men, and biochemical recurrence (BCR) remains a persistent problem after clinical treatment. Commonly used indicators do not identify patients who are likely to relapse early. Protein O-linked glycosylation is an important post-translational modification involved in cancer progression, but its prognostic relevance and regulatory mechanisms in PCa remain insufficiently understood. GALNTL6, which encodes polypeptide N-acetylgalactosaminyltransferase-like 6, is putatively involved in mucin-type protein O-linked glycosylation and remains poorly characterized in PCa. To address this issue, we integrated several public transcriptomic cohorts with matched clinical data to develop an O-glycosylation-related prognostic model.

Methods: Different machine-learning approaches were tested during model development. The final score O-glycosylation scores (OGs) was then examined for risk classification, nomogram construction, and prediction of drug sensitivity. GALNTL6 was chosen for experimental validation in DU145 and 22Rv1 cells using both overexpression and knockdown designs, followed by assays for cell proliferation, migration, and invasion. RNA sequencing was also carried out to investigate downstream molecular changes.

Results: The O-glycosylation signature consistently stratified patients into distinct risk groups across multiple independent cohorts (P<0.01), and the derived risk score remained an independent predictor of BCR [hazard ratio (HR) >1.48, P<0.05]. A nomogram integrating the O-glycosylation signature with clinicopathological variables achieved good performance in estimating 1-, 3-, and 5-year BCR probabilities [area under the curve (AUC): 0.747-0.863]. Drug-sensitivity analysis suggested that patients with different OGs scores may show distinct predicted responses to candidate agents, including AZD7762, a checkpoint kinase inhibitor, and tigecycline, a glycylcycline antibiotic with reported antitumor activity. GALNTL6 emerged as the central gene within the signature, and its increased expression was associated with unfavorable prognosis, more advanced disease status, and higher prostate-specific antigen (PSA) levels. Functionally, silencing GALNTL6 reduced PCa cell growth and motility, whereas enforced GALNTL6 expression produced the opposite effects (P<0.05). Transcriptomic analysis further indicated that GALNTL6 overexpression was associated with PI3K-AKT signaling and antioxidant-response genes.

Conclusions: This study establishes an O-glycosylation-related prognostic model for BCR risk stratification in PCa and identifies GALNTL6 as a potential promoter of malignant phenotypes. These findings suggest that GALNTL6 may be associated with AKT pathway activity and oxidative-stress-related transcriptional programs, although further mechanistic validation is required.