Whole-Brain Single-Neuron Atlas Reveals Microglial Security Hole Accelerating Neuronal Vulnerability

Neurodegenerative diseases are characterized by progressive neuronal loss, yet subtle degeneration remains difficult to detect, hindering advancements in early diagnosis and intervention strategies. Here, we present a 3D, whole-brain, single-neuron atlas that aligns and compares neurons at single-cell resolution, surpassing 2D spatial limitations to precisely detect heterogeneous neurodegenerative lesions. Using the AppNL-G-F model of Alzheimer’s disease, we found that neuronal loss begins concurrently with amyloid-β deposition, indicating a more acute timeline than proposed by the amyloid hypothesis. We also identified age-dependent microglial redistribution from gray-matter–rich to white-matter–rich regions in wild-type mice, accelerated in AppNL-G-F, potentially reducing neuroprotective function. Spatial single-cell-resolution risk analysis revealed microglial depletion, rather than proliferation, as an early risk indicator for neuronal loss, and microgliosis often coexists with microglial loss, suggesting heightened microglial vulnerability. Integration with spatial transcriptomics showed neurons near microglia with reduced homeostatic gene expression are especially vulnerable. Together, these findings demonstrated that amyloid pathology skews microglial distribution and promotes their exhaustion, leading to microglial spatial disorganization and neuronal vulnerability, resulting in “microglial security holes”. Our study underscores the transformative potential of 3D analysis in neurodegenerative research, offering a versatile platform for organ-level spatial multi-omics integration to advance detection and therapeutic strategies.