miRNA-Based Signature to Predict the Development of Alzheimer's Disease

Background: Patients with mild cognitive impairment (MCI) suffer from a high risk of developing Alzheimer's disease (AD). Cumulative evidence has demonstrated that the development of AD is a complex process that could be modulated by miRNAs. Here, we aimed to identify miRNAs involved in the pathway, and interrogate their ability to predict prognosis in patients with MCI. Methods: We obtained the miRNA-seq profiles and the clinical characteristics of patients with MCI from the Gene Expression Omnibus (GEO). Cox regression analysis was used to construct a risk level model. The receiver operating characteristic (ROC) curve was used to assess the performance of the model for predicting prognosis. Combined with clinical characteristics, factors associated with prognosis were identified and a predictive prognosis nomogram was developed and validated. Through Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis, we evaluated molecular signatures for the candidate miRNAs. Results: Our analysis identified 120 DEmiRNAs. The Cox regression analysis showed that two miRNAs could serve as risk factors for disease development. A risk level model was constructed. Age, apoe4, and risk level were associated with the prognosis. We developed a nomogram to predict disease progression. The calibration curve and concordance index (C-index) demonstrated the reliability of the nomogram. Functional enrichment analysis showed that these miRNAs were involved in regulating both cGMP-PKG and Sphingolipid signaling pathways. Conclusion: We have identified miRNAs associated with the development of MCI. These miRNAs could be used for early diagnosis and surveillance in patients with MCI, enabling prediction of the development of AD.