A non-destructive workflow for simultaneous reconstruction of nuclear and dendritic architecture in intact brain...

Neurons display elaborate dendritic and axonal structures which serve as input and output sites for the transmission of information across the nervous system. Changes to the structure and complexity of both are commonly reported in genetic and acquired brain diseases with neuropathology studies also reporting changes to nuclear architecture. However, a methodological limitation has prevented direct study of this relationship: resolving nuclear geometry conventionally requires physical sectioning that severs dendrites, whereas tissue-clearing approaches that preserve dendrites introduce anisotropic dimensional distortion. No existing method simultaneously quantifies nuclear geometry and dendritic morphology within the same intact neuron at sub-micrometre resolution. We exploit a property of Golgi-Cox impregnation that has historically been treated as a staining artefact. Namely, the consistent exclusion of silver-chromate precipitate from the nucleus, which appears as a geometrically precise, low-attenuation void within the high-contrast soma in X-ray microscopy (XRM). Using this void as a label-free nuclear imaging channel, we performed compartment-level morphometry on mouse neurons across two independent XRM scans at sub-micrometre voxel resolution (0.619 and 0.726µm isotropic), and full whole-neuron reconstructions including nucleus, soma, and dendritic arbour. Across 26 CA1 neurons, nuclear morphometrics were consistent: median nuclear sphericity 0.893, median soma-to-nucleus centroid offset 1.35µm, median perinuclear cytoplasmic thickness 3.96µm, and median nucleus-to-soma volume ratio 4.2%. Nuclear metrics did not differ significantly between scans after correction for multiple comparisons. This feasibility study establishes Golgi-XRM for simultaneous, non-destructive reconstruction of nuclear geometry and dendritic architecture as a tool for investigating nuclear positioning as a disease-sensitive variable.