Speaker
Description
We present a practical, high-quality workflow for large-volume plasma focused ion beam-scanning electron microscopy (PFIB-SEM) tomography on mast cells. Beyond their classic roles in allergy, mast cells are vital for tissue homeostasis, repair, and defence, yet aspects of their ontogeny and functions in cancer and cardiovascular health remain unresolved. Progress requires ultrastructural reconstructions and segmentation of the numerous vesicles and granules, a step that is still largely carried out manually. Given that robust AI model training demands extensive datasets, we employed large-volume Xenon PFIB-SEM tomography, to accelerate data acquisition and extend an established correlative light and electron microscopy (CLEM) workflow with AI based segmentation.
A key challenge in PFIB-SEM of high-pressure frozen, cryo-substituted mast cells embedded in epoxy is mitigating charging and thermal damage while preserving a smooth, artefact-minimized surface. We addressed this by adding conductive carbon nanoparticles to the resin; a 5% loading best suppressed charging while maintaining workable viscosity. Following optimization of the slicing conditions at 30 keV, combined with stage rocking, we achieved smooth sections with minimal curtaining. For SEM imaging at 5 keV, a low-energy backscattered electron detector delivered the highest signal‑to‑noise ratio and enhanced ultrastructural contrast. Using these settings, it is possible to generate both large datasets comprising several cells with a voxel size of 50 nm and high-resolution reconstructions with ca. 1000 slices and isotropic voxels of 7 nm.
We will detail the workflow and initial AI‑segmentation results and discuss how PFIB‑SEM accelerates the collection of large, high-fidelity datasets, thereby representing a promising method for further pathology-relevant cells and tissues.