Revealing Cellular Structures by Expansion Microscopy

Abstract

Expansion microscopy (ExM) enables nanoscale imaging of biological structures by physically enlarging specimens embedded in swellable hydrogels. In this study, we aimed to investigate the synaptonemal complex (SC) in mouse spermatocytes in a near-native state using quantitative ExM and to optimise hydrogel-based expansion conditions. To quantify the linear expansion factor (Q), we embedded fluorescent beads in the hydrogel and explored two approaches. A nearest-neighbour–based comparison of pre- and post-expansion bead distributions proved impractical, as standard image-registration algorithms failed on sparse, low-information bead images, likely due to convergence to local minima. We therefore used a second, sampling-based approach, in which bead positions were measured before and after expansion and Q was calculated under the assumption of isotropic swelling. Computer simulations were used to estimate the theoretical error in Q as a function of the number of sampled beads. With a quantitative readout of Q, we systematically examined hydrogel swelling in a saline series (0–1000 mM NaCl). We found that salt concentration modulates expansion in an overall inverse manner, with Q values ranging from ~2 at 0 mM to ~12 at 0.3 mM and ~3–5 at higher ionic strengths. These results demonstrate that ionic conditions can be tuned to achieve a desired expansion when maximal enlargement is not required. The pan-ExM protocol was first validated on U-2OS cells, where single and especially double expansions revealed ultrastructural detail not observable in unexpanded samples. NHS-ester pan-staining produced strong protein-dense signals, and compatibility with nucleic acid counterstaining was confirmed using propidium iodide, which identified bright nuclear foci as nucleoli. Isolated mouse spermatocytes were then expanded and immunostained for SYCP3 to visualise SC lateral elements (LEs). Single expansion improved SC visibility but was accompanied by ruptures along the SC axis. Increasing denaturation time and buffer salt concentration alone did not resolve this issue. However, denaturation at 95 °C with pre-heated buffer increased the signal ratio along the SC axis by 27% and preserved SYCP3 epitopes for post-expansion staining, yielding improved LE continuity and detail. Nonetheless, single expansion did not provide sufficient axial resolution to fully resolve individual LEs, indicating that double expansion, combined with optimised high-temperature, high-salt denaturation, will be required for a more complete, quantitative characterisation of SC architecture in future work.