ABSTRACT: Late-onset Alzheimer’s disease (LOAD), which accounts for more than 95% of cases, lacks a well-defined gene expression signature, unlike early-onset forms caused by autosomal dominant mutations in PSEN1, PSEN2, and APP. Although early and late onset Alzheimer’s share core neuropathological features, their genetic convergence had not previously been demonstrated. In this study, synchronized skin fibroblasts from autopsy-confirmed LOAD patients revealed 20 significantly dysregulated genes (Adjusted P ≤ 0.05) that clearly distinguished LOAD from non-Alzheimer’s dementias. Among these, NHLH1, URB2, and WASF2 were also dysregulated in blood B lymphocytes, supporting a cross-tissue “modular resilience” framework. An additional 12 dysregulated gene families in skin were validated in blood, resulting in 15 LOAD-associated gene families shared across tissues. Furthermore, five genes—CARNS1, NORAD, SCG2, SNHG14, and WASF2—were differentially expressed in skin fibroblasts and dysregulated in the prefrontal cortex, indicating a multi-layered resilience module connecting peripheral and central tissues. These genes participate in metabolic buffering, genomic stability, secretory organization, epigenetic regulation, and cytoskeletal remodeling. Their coordinated dysfunction suggests a systems-level impairment in stress adaptation relevant to AD biology. Overall, the results reveal previously unrecognized gene-network convergence between familial early-onset and sporadic late-onset Alzheimer’s disease. The robust cross-tissue reproducibility of LOAD dysregulated gene networks highlights the potential of synchronized peripheral cells as accessible sources for disease-specific genomic profiling. This approach may enable the development of personalized biomarkers and targeted therapeutic strategies for Alzheimer’s disease.
