Ultramicroscopy of thy1 driven green fluorescent protein expression visualizes the three-dimensional network of pyramidal cell dendrites.
Satellite and aerial images in the mapping program Google Earth place your home in the context of the buildings that surround it. A birds-eye view would also be helpful when imaging cellular gene expression in the brain. However, brain imaging techniques (like magnetic resonance imaging) do not resolve individual neurons, and cell imaging techniques (like confocal microscopy) do not address the relationship between individual neurons and the rest of the brain. Now Dodt et al. and Verveer et al. report variations on a 100-year-old microscopy technique to image whole tissues with single-cell resolution in recent articles in Nature Methods.
In traditional ultramicroscopy, also called light sheet illumination, researchers shine a thin beam of light on transparent tissue in a darkened chamber. They can then view the illuminated cross section of tissue through an objective perpendicular to the light beam. Unlike confocal microscopy, in which light scatters throughout the tissue, ultramicroscopy illuminates only one plane of tissue is at a time, limiting photobleaching, or light-induced damage. Because the resolution of the two-dimensional image relates to the thickness of the plane of light, researchers can use small aperture objectives, allowing a large field of view. Both groups summed multiple image planes to achieve a three-dimensional image.
Unlike tissue traditionally imaged by ultramicroscopy, brains are opaque. So, Dodt et al. used a clearing technique to make them transparent. They immersed brains in solutions that seeped into intra and extracellular spaces, equalizing their refractive indexes, and imaged brains in glass chambers containing clearing medium. They focused two beams of blue laser light on opposite sides of the chamber into one horizontal sheet of light that illuminated only a thin slice of tissue. Then they moved the chamber vertically to take consecutive two-dimensional images. In contrast, Verveer et al. embedded tissue in agarose, which would sustain living tissue, and rotated it at set angles for consecutive two-dimensional images.
Dodt et al. imaged neurons expressing green fluorescent protein (GFP) under the control of the thy1 promoter in brains from ten-day-old mice. To generate a three-dimensional model of the brain, they summed images of autofluorescence that visualized the outline of each brain section. GFP expression localized to cell bodies and dendrites in the hippocampus. The authors increased resolution by imaging dissected hippocampi and visualized the three-dimensional dendritic network of pyramidal cells by summing individual images of GFP expression. With a more powerful objective, they identified GFP-labeled spines on pyramidal neuron dendrites.
The authors visualized three-dimensional models of intact mouse embryos and Drosophila. They had difficulty imaging brains from mice older than two weeks of age because increased myelination prevented them from clearing in solution, suggesting that new clearing strategies will be necessary to image the adult brain.
Verveer et al. found better resolution of three-dimensional subcellular structures by ultramicroscopy than confocal microscopy, which is more labor and data intensive. According to Dodt et al., ultramicroscopy will help map three-dimensional networks of related neurons and rapidly identify changes in these networks in mouse models of disease.