Abstract
Oxytocin (OXT), a neuropeptide hormone essential to a wide range of social functions, has drawn increasing attention as a crucial contributor to the neurobiology of autism spectrum disorder (ASD). Central OXT system disruptions have been reported in several genetic mouse models of ASD; however, a detailed and systematic characterization of these phenotypes, and cross-model identification of shared and distinct features, are presently lacking. We integrated whole-brain OXT immunolabeling, SHIELD tissue clearing, light-sheet microscopy, and three-dimensional (3D) machine learning-based cell detection to establish a high-throughput, intact-tissue pipeline and quantified OXT immunopositive (OXT+) neurons across subregions of the paraventricular nucleus of the hypothalamus (PVN) in two genetic mouse models of ASD: Cntnap2 and Fmr1 knockout (KO) mice. We validated this pipeline alongside conventional immunohistochemistry using tissue sections. We show subregion- and sex-specific differences in PVN OXT+ cell counts in the two KO models. Notably, whole-PVN analysis revealed additional subregion- and sex-specific differences that were not evident in section-based quantification. These results identify subregion- and sex-specific differences in PVN OXT+ neuronal distribution as a shared phenotype in two genetic mouse models of ASD. This work highlights the importance of region-specific, high-resolution 3D approaches in intact tissue for quantifying cell populations within anatomically complex brain regions.