State of the art of volume imaging and CLEM of organoids

Abstract text (incl. figure legends and references)

Organoids recapitulate tissue architectures and represent an opportunity to investigate healthy or diseased tissues. Understanding organoid complexity requires the integration of diverse imaging modalities covering different spatial and temporal resolutions. Fluorescence light microscopy allows studying the cellular sociology in living samples. FIB-SEM volume imaging can further probe the multicellular environment at ultrastructural resolution. Cryo-electron tomography (cryoET) bridges the gap with atomic resolution to investigate macromolecular complexes in the cell. Further, deep-learning software is required to correlate and quantitatively analyse these data. Here, we present the state of the art and future direction of light and electron microscopy imaging of organoids both at room temperature and cryogenic regime.

Matrigel-embedded organoids were either cultured directly in HPF carriers1 or multi-well dishes using liquid medium. These carries were used to perform confocal live-cell imaging using cell-permeable fluorescence dyes. After high-pressure freezing, the fluorescence signal guided FIB-SEM volume imaging acquisitions at low temperature. Further, sample thinning for cryo-electron tomography was achieved by lift-out. Alternatively, organoids in HPF carriers were processed by freeze-substitution and resin embedding follow by laser branding to FIB-SEM volume imaging acquisitions2. FIB-SEM was done on a performed on a Zeiss CrossBeam XB550 and on Thermo Fisher Scientific Aquilos. Lastly, correlation of light and electron microscopy data as well as automatic image segmentation was done by ORS Dragonfly.

We show how to make organoid delicate specimens amenable to high-pressure freezing and fluorescence light microscopy both a room temperature and cryogenic regime. We compare the advantages of live dyes versus genetically encoded fluorescent reports and we show the limits of fluorescence light microscopy in imaging large (> 200µm) and deep (> 20µm) biological samples. Then, we illustrate correlative light and electron microscopy approaches for organoid samples. Further, we describe our results of FIB-SEM volume imaging at ultrastructural resolution of organoids at room temperature and the current limits of imaging at low-temperature. To provide a quantitative understanding of the complex and heterogenous multicellular environment of organoids, large FIB-SEM dataset need to be annotated. While manual annotation becomes unfeasible given the large amount of data, deep-learning algorithms offer a workable solution for automating this task. We used trainable convolutional neural networks provided in the software ORS Dragonfly for automatic segmentation of subcellular structures. We focused our attention especially on the interaction between cortical cytoskeleton and cell junctions in the context of patient-derived cancer organoids. Finally, we provide an overview of the actual possibilities to perform lift-out from high-pressure frozen samples for cryoET imaging.

Ronchi, P. et al. High-precision targeting workflow for volume electron microscopy. J. Cell Biol. 220, (2021). D"Imprima, E. et al. Integrated light and electron microscopy continuum resolution imaging of 3D cell cultures. bioRxiv 2021.07.02.450855 (2021) doi:10.1101/2021.07.02.450855.