3D-Imaging of synapses in neuronal tissues with synchrotron X-ray ptychography

Abstract

AbstractWiring diagrams of neural circuits are of central importance in delineating mechanisms of computation in the brain (Lichtman and Sanes, 2008; Litwin-Kumar and Turaga, 2019). To generate these diagrams, the individual parts of neurons - axons, dendrites and synapses - must be densely identified in 3-dimensional volumes of neuronal tissue. This is typically achieved by electron microscopy (Kornfeld and Denk, 2018), necessitating physical sectioning of the specimen either before or during the image acquisition process using ultrathin sectioning techniques or gallium or gas cluster ion beams (Denk and Horstmann, 2004; Hayworth et al., 2020; Kasthuri et al., 2015; Xu et al., 2017). Here, we demonstrate that X-ray ptychography (Pfeiffer, 2018), a coherent diffractive X-ray imaging technique, can faithfully acquire 3-dimensional images of metal-stained mouse neuronal tissue. Achieving high imaging quality requires minimization of the radiation damage to the sample, which we achieve by imaging at cryogenic temperatures and using specialised tomographic reconstruction algorithms (Odstrcil et al., 2019b). Using a newly identified tri-functional epoxy resin we demonstrate radiation resistance to X-ray doses exceeding 1011Gy. Sub-40 nm resolution makes it possible to densely resolve axon bundles, boutons, dendrites, and synapses without physical sectioning. Moreover, the tissue volumes can subsequently be imaged in 3D using high-resolution focused ion beam scanning electron microscopy (FIB-SEM) (Heymann et al., 2006; Knott et al., 2008) showing intact ultrastructure, suggesting that metal-stained neuronal tissue can be highly radiation-stable. Ongoing improvements in synchrotron, X-ray and detector physics (Yabashi and Tanaka, 2017), as well as further optimization of sample preparation and staining procedures (Hua et al., 2015; Karlupia et al., 2023; Lu et al., 2023; Mikula and Denk, 2015; Pallotto et al., 2015; Song et al., 2022), could lead to substantial improvements in acquisition speed (Du et al., 2021), whilst widening the volumes that can be imaged with X-ray techniques using laminography (Helfen et al., 2005; Helfen et al., 2013; Holler et al., 2020b; Holler et al., 2019) and nano-holotomography (Cloetens et al., 1999; Kuan et al., 2020) could allow for non-destructive X-ray imaging of synapses and neural circuits contained in volumes of increasing size.