Salmon, Christopher K., Syed, Tabish A., Kacerovsky, J. Benjamin, Alivodej, Nensi, Schober, Alexandra L., Sloan, Tyler F.W., Pratte, Michael T., Rosen, Michael P., Green, Miranda, Chirgwin-Dasgupta, Adario, Mehta, Shaurya, Jilani, Affan, Wang, Yanan, Vali, Hojatollah, Mandato, Craig A., Siddiqi, Kaleem, and Murai, Keith K.
Astrocytes are increasingly understood to be important regulators of central nervous system (CNS) function in health and disease; yet, we have little quantitative understanding of their complex architecture. While broad categories of astrocytic structures are known, the discrete building blocks that compose them, along with their geometry and organizing principles, are poorly understood. Quantitative investigation of astrocytic complexity is impeded by the absence of high-resolution datasets and robust computational approaches to analyze these intricate cells. To address this, we produced four ultra-high-resolution datasets of mouse cerebral cortex using serial electron microscopy and developed astrocyte-tailored computer vision methods for accurate structural analysis. We unearthed specific anatomical building blocks, structural motifs, connectivity hubs, and hierarchical organizations of astrocytes. Furthermore, we found that astrocytes interact with discrete clusters of synapses and that astrocytic mitochondria are distributed to lie closer to larger clusters of synapses. Our findings provide a geometrically principled, quantitative understanding of astrocytic nanoarchitecture and point to an unexpected level of complexity in how astrocytes interact with CNS microanatomy. [Display omitted] • High-resolution serial electron microscopy resolves astrocytic structural features • Computer vision methods are used to quantitatively understand nanoarchitecture • Astrocytic structure is hierarchical with hubs and specific organelle arrangements • Astrocytes organize brain connectivity through astrocyte-defined synaptic clusters Astrocytes are master coordinators of brain function; yet, a quantitative investigation of their intricate nanoarchitecture is missing. Salmon et al. use astrocyte-tailored computer vision methods to challenge the present understanding of an astrocyte's distinct architecture, unique interactions with synapses and vasculature, and functional organization. [ABSTRACT FROM AUTHOR]