Background: Emerging evidence highlights the critical involvement of noncoding RNAs (ncRNAs) in the pathologic process of Alzheimer disease (AD). However, the precise functional contributions and regulatory landscapes of long noncoding RNAs (lncRNAs) in AD remain unclear. Objective: This study aims to delineate the regulatory roles of functional lncRNAs in AD by systematically analyzing lncRNA-mRNA interactions within an AD-associated network, leveraging the competing endogenous RNA (ceRNA) framework. Methods: We constructed an AD-specific lncRNA-mRNA network by integrating a probe reannotation pipeline with experimentally validated microRNA (miRNA)-lncRNA/mRNA interactions. Key lncRNAs were identified through topological analysis, while bidirectional hierarchical clustering was applied to define functional modules. Pearson correlation coefficients were computed to assess the association patterns of lncRNA-mRNA pairs. Results: Using a structured analytical approach, we identified 31 differentially expressed lncRNAs and 1045 mRNAs within the AD network. Topological analysis revealed SNHG12, MIR17HG, and GAS5 as central regulatory nodes, indicating their potential roles in network stability and transcriptional regulation. A functionally coherent module of lncRNA-mRNA interactions was identified through clustering, with enrichment in multiple AD-relevant signaling pathways, including neuroinflammation and synaptic dysfunction, suggesting a mechanistic role for lncRNAs in disease-associated processes. Furthermore, leveraging the ceRNA model, we mapped dysregulated ceRNA interactions, uncovering significant alterations in ceRNA crosstalk between AD and non-AD conditions, which may reflect broader disruptions in posttranscriptional gene regulation. Conclusions: Our findings provide key insights into the functional architecture of lncRNA regulatory networks in AD, offering a refined perspective on their contributions to disease pathology and highlighting potential biomarkers and therapeutic targets.
Su et al. (Mon,) studied this question.