Matrix stiffness-related transcriptomic signatures reveal astrocyte-centered neuroimmune alterations in Alzheimer’s disease

Abstract Background Extracellular matrix remodeling and altered tissue mechanics are increasingly implicated in Alzheimer’s disease (AD), but links between matrix stiffness and AD pathology remain unclear. Objective To identify matrix stiffness-related transcriptomic signatures in AD and characterize their diagnostic potential, functions, cell-type specificity, and intercellular communication patterns. Methods Bulk transcriptomic datasets from the Gene Expression Omnibus were integrated with literature-derived matrix stiffness-related genes. Differential expression analysis, three machine-learning algorithms, and cross-dataset validation were used to identify candidate genes and construct a multigene classification model. Functional enrichment and immune-cell deconvolution analyses were performed. Single-cell RNA sequencing data were analyzed to assess cell-type-specific expression, inferred astrocyte trajectories, and predicted cell–cell communication. Results Eighty-one differentially expressed matrix stiffness-related genes were identified in AD brain tissue. FREM3, DUSP4, and FAM107A were selected and showed concordant differential expression across discovery and validation datasets. The three-gene model achieved an area under the receiver operating characteristic curve of 0.963 in the discovery cohort. Gene set enrichment analysis linked all three genes to oxidative phosphorylation, although FAM107A showed an enrichment direction opposite to DUSP4 and FREM3. Immune deconvolution suggested associations with immune-cell signatures. Single-cell analysis identified astrocytes as a major population expressing the three-gene signature. AD astrocytes showed altered pseudotime distributions and stronger predicted communication with immune cells, particularly T cells and macrophages. Conclusions FREM3, DUSP4, and FAM107A are candidate matrix stiffness-related transcriptomic signatures in AD and may connect extracellular matrix remodeling with mitochondrial metabolism, immune dysregulation, and astrocyte-associated changes. Independent clinical validation and experimental confirmation are required.