FASEB J. 2026 Sep 15;40(17):e72240. doi: 10.1096/fj.202504766R.
ABSTRACT
The core pathological features of Alzheimer's disease (AD) include Aβ plaques and neurofibrillary tangles, which collectively drive the neurodegenerative process. Meanwhile, the phagocytic function of microglia plays a dual role in AD: it attempts to clear pathological proteins such as Aβ, but its chronic activation may also exacerbate neuroinflammation and synaptic damage. This study integrated microarray data from AD cohorts in the Gene Expression Omnibus (GEO) database. Through differential expression analysis, protein-protein interaction network construction, and three machine learning algorithms (Least Absolute Shrinkage and Selection Operator, Support Vector Machine-Recursive Feature Elimination, and Extreme Gradient Boosting), key microglial phagocytosis-related signature genes were identified. Diagnostic models were subsequently developed and validated, with further investigation of immune infiltration patterns, regulatory networks, and molecular subtypes. Based on comprehensive analysis, this study identified five core signature genes (HLA-DPA1, IL4R, ITGAM, SPP1, TNFRSF1B) that form a highly accurate diagnostic model validated in independent cohorts. Immune infiltration analysis revealed significant increases in neutrophils and M2 macrophages in AD brains, with these genes showing strong correlations with immune cell abundance. The study further identified two molecular subtypes with distinct immune features, constructed regulatory networks, and predicted potential therapeutics including Tamibarotene. By integrating transcriptomics and machine learning, this study identifies key molecular features of microglial phagocytosis in AD, providing a novel diagnostic framework and insights into the immune mechanisms of the disease.
PMID:42663999 | DOI:10.1096/fj.202504766R