
We identified a novel start-codon mutation in TREM2 (TREM2c·3G>A, M1I) that causes complete loss of TREM2 expression and is associated with early-onset Alzheimer’s disease (AD) in humans. This mutation triggers profound blood-brain barrier (BBB) impairment in both humans and mice, positioning BBB impairment as an early trigger rather than a downstream consequence of neurodegeneration in AD. Mechanistically, TREM2 deficiency causes NF-κB-driven upregulation of microglial CCL2, which activates CCR2 on endothelial cells and induces tight junction internalization. The resulting BBB compromise permits the leakage of peripheral pathological proteins, including β2-microglobulin, accelerating amyloid plaque formation and premature Tau pathology. Pharmacological inhibition of CCR2 and endothelial cell-specific Ccr2 knockdown restored BBB function and reduced amyloid plaque burden, highlighting this pathway as a tractable therapeutic target. Our findings establish a direct mechanistic link between microglial dysfunction, BBB breakdown, and pathogenic protein accumulation, thereby redefining the vascular role of TREM2 in neurodegenerative disease.
Alzheimer’s disease (AD) is a multifactorial disorder in which cerebrovascular abnormalities frequently co-occur with β-amyloid (Aβ) and tau pathology. White matter hyperintensities (WMH), a common marker of cerebral small vessel disease, are associated with cognitive impairment and are highly prevalent across the AD continuum. However, the extent to which WMH burden is associated with spatial patterns of tau pathology and its longitudinal accumulation remains unclear. We examined 1,455 participants with Aβ PET, tau PET, T2-FLAIR MRI, and cognitive assessments from four independent cohorts: ADNI (n = 710), A4/LEARN (n = 440), GHABS (n = 249), and PREVENT-AD (n = 56). Resting-state functional MRI data were available in subsets of ADNI and A4, and CSF SomaScan proteomic data were available in ADNI and PREVENT-AD. Associations of WMH burden with longitudinal tau accumulation and network-based tau propagation were examined. A data-driven approach was used to identify distinct tau subtypes. WMH burden, tau accumulation, and clinical and biological characteristics were evaluated across subtypes. Greater total WMH burden was associated with faster tau accumulation (ADNI, A4) or higher subsequent tau burden (GHABS) in the temporal meta-ROI among Aβ-positive (A+) individuals. These associations were attenuated after accounting for Aβ burden. However, greater WMH burden, particularly in the occipital lobe, remained significantly associated with faster tau accumulation and network-based tau propagation in occipital regions after adjustment for Aβ burden. An occipital-dominant tau subtype was identified, characterized by greater occipital than parietal tau burden. This subtype showed more widespread associations between WMH and tau accumulation, an older estimated age of Aβ positivity onset, and a higher prevalence of hypertension, together with slower global tau accumulation and cognitive decline than the parietal-dominant subtype. CSF proteomic analyses further identified vascular and extracellular matrix pathways related to tenascin-C (TNC) in the occipital-dominant subtype. Greater WMH burden was associated with faster tau accumulation and network-based tau propagation, with the most consistent associations involving occipital regions. These findings suggest that the association between WMH burden and occipital tau progression may reflect vascular contributions to regional tau vulnerability in the context of AD pathology.
Gene signatures of Alzheimer's disease (AD) brains reflect the output of a complex interplay of genetic, epigenetic, epi-transcriptomic, and post-transcriptional regulations. To nominate candidate factors modulating these signatures, we developed a machine learning model to integrate cellular and molecular features explaining differential gene expression in AD. Among the features tested, YTHDF proteins, the canonical readers of N6-methyladenosine (m6A) RNA modification, are among the most influential predictors of AD gene signatures. Protein modules containing YTHDFs were downregulated in human AD brains, and knockdown or pharmacological inhibition of YTHDFs in iPSC-derived 2D and 3D neuronal models recapitulated key AD-associated gene signatures. Furthermore, eCLIP-seq revealed altered YTHDF binding to transcripts in AD brains, at both m6A-dependent and m6A-independent sites. Together, these results support an important role for YTHDF proteins in modulating AD-associated gene signatures in the human brain.
The APOE4 allele is the predominant genetic risk factor for late-onset Alzheimer’s disease (AD), and its pathological influence is increasingly understood through the lens of lipid biology. APOE functions as the primary lipoprotein in the central nervous system, coordinating the distribution of cholesterol, phospholipids, and fatty acids among neurons and glial cells to sustain membrane integrity, synaptic function, and myelin maintenance. The E4 isoform carries structural alterations that compromise its lipid carrying capacity and receptor interactions, with consequences that extend across multiple brain cell types. In this review, we synthesize current evidence on how APOE4 disrupts lipid homeostasis in a cell type specific manner, covering astrocytic lipid buffering and oxidative stress responses, microglial phagocytic and lysosomal dysfunction, oligodendrocyte cholesterol handling and myelin biosynthesis, and neuronal lipid clearance. We also address the relationship between APOE isoforms and the hallmark pathologies of AD, including amyloid beta accumulation and tau propagation. Insights from protective natural variants such as APOE2, Christchurch, and Jacksonville are discussed in the context of therapeutic target identification. Evidence drawn from human postmortem studies, neuroimaging, single-cell transcriptomics, and iPSC derived brain cell models converges on a model suggests that disrupted glial lipid metabolism may contribute early in disease, although whether it initiates or amplifies pathology remains unresolved. Targeting APOE-mediated lipid pathways therefore offers a promising and mechanistically grounded avenue for disease modifying intervention in AD.