LightSheet Fluorescent Imaging at Expanded Space to Approach EM Resolution

Abstract

Electron microscopy (EM) has imaged densely labeled brain tissue at nanometer-level resolution over near-millimeter-level dimensions but lacks the contrast to distinguish specific proteins and the speed to readily image multiple specimens. Conversely, optical imaging techniques provide much important information in understanding life science especially cellular structure and morphology. However, the resolution of optical imaging is limited by the diffraction limit, which is discovered by Ernst Abbe, i.e. λ/2(NA), (NA is the numerical aperture of the objective lens). For the last 100 years, biologists and optical scientists were unable to obtain a clear optical image of biological entities down to molecular level that are smaller than the diffraction limit (around 200 nm at lateral resolution). The winners of Nobel Prize in Chemistry 2014, E. Betzig, W. E. Moerner and S. Hell, innovated super-resolution microscopic techniques. Such techniques enable biologists to visualize nano-sized fluorophores that are beyond the diffraction limit. These techniques do not physically violate the Abbe limit of resolution but exploit the photoluminescence properties and labelling specificity of fluorescence molecules to achieve super-resolution imaging ?? Instead of sweating on the super-resolution techniques to pursuit high spatial resolution, expansion microscopy (ExM) is invented to bypass the optical diffraction limit by physically expanding the samples to ~4 times larger than original with swellable polymers. Furthermore, applying the swellable hydrogel concept, with exchange of certain monomer component, we could also obtain ~10 times in size change of both cell culture and Drosophila brain. In order to image such expanded samples, we use lightsheet microscopy, a separate excitation lens perpendicular to the widefield detection lens to confine the illumination to the neighborhood of the focal plane. By combining intrinsic optical sectioning with widefield detection, lightsheet microscopy allows fast imaging speed to record multi-megapixel imaging of selected plane in a single exposure of the camera. By manipulating sample space, we wish to approach electron microscopic resolution with lightsheet fluorescent detection to investigate 3D subcellular morphologies and protein connectomics in several mm-scale samples. The challenges we are facing are as following, 1) Hydrogel expansion not only expands the sample itself, but also those unwanted blank regions. Current acquisition method requires manual intervention to identify and remove these chunks. A fast estimator that can filter out chunks that are unlikely to exhibit sample of interest. 2) Though swelled sample effectively lower the observable resolution, fluorophore spatial density will drop considerably in the process. Typically, experienced users can lengthen the exposure time to increase the signal-to-noise ratio, but inhomogeneous fluorophore distribution may require user to adjust the exposure duration on-the-fly for optimal result. A heuristic that can determine the exposure time for a local region, which can yield optimal (sufficient) SNR. 3) Large scale volumetric bio-images are often fragmented numerous (thousands) of sub-volumes, acquired serially. They are later stitched together using overlap similarities. Currently, similarity is determined by DFT-based correlation method, and later the visual discrepancies through minimizing the shifts. Context-oriented similarity comparison, that can prioritize sample continuity and take account their structural aspects, instead of relying on independent overlap regions, which is error prone. 4) Expanded hydrogel is composed of over 95% water, while favorable for acquisition optics, it introduces discernible non-repeated wiggling artifacts everywhere. Field homogeneity, camera shot noises, stripes and dot artifacts are distributions that we can identify and remove. However, non-repeatable pattern is still an unresolved challenge.