Identification and immunological features of anoikis-related clusters in Alzheimer’s disease

Objective: Anoikis, a form of apoptosis triggered by cell detachment from the extracellular matrix, has been implicated in neurodegenerative disorders, yet its role in Alzheimer’s disease (AD) remains unclear. This study aimed to identify anoikis-related genes (ARGs) in AD, characterize their immunological associations, and establish a diagnostic model for clinical prediction. Methods: Gene expression data from GEO datasets were analyzed to identify differentially expressed ARGs (DEARGs). Chromosomal distribution, correlation, and immune infiltration analyses were performed. Clustering was applied to stratify AD patients into subtypes. Gene set variation analysis (GSVA) and weighted gene co-expression network analysis (WGCNA) were conducted to explore functional pathways and hub genes. Machine learning algorithms, including random forest (RF), support vector machine, eXtreme gradient boosting, and generalized linear model, were compared to construct an optimal diagnostic model, which was validated using an external dataset. Results: Ten DEARGs were identified, of which nine were upregulated and one was downregulated in AD. DEARGs correlated with immune alterations, including increased resting natural killer (NK) cells and neutrophils, alongside reduced CD4 naïve T cells, CD4 memory activated T cells, M1 macrophages, and activated dendritic cells. Clustering divided AD patients into two subtypes with distinct immune infiltration patterns; cluster 2 was enriched with neutrophils and associated with heightened risk. Functional enrichment highlighted pathways involving cell cycle regulation, immune activation, and metabolic processes. The RF model demonstrated superior diagnostic performance, and a five-gene signature (choline ethanolamine phospho-transferase (CEPT1), coilin (COIL), ADRM1 26S proteasome ubiquitin receptor (ADRM1), ADP ribosylation factor 3 (ARF3) and mitochondrial calcium uptake 1 (CBARA1)) was incorporated into a nomogram with good predictive efficiency, validated by external data (area under the curve (AUC)=0.639). Conclusions: This study reveals that DEARGs and immune heterogeneity play pivotal roles in AD progression. The diagnostic model provides a promising tool for AD risk prediction.