Using digital scanned laser light sheet fluorescence microscopy (DSLM) and structured illumination microscopy (SI), researchers have created 3-D models of embryonic development. The images used to make the models capture both rapidly occurring cellular processes and long-term developmental changes at a high resolution.
Studying the development of whole embryos has been a challenge for researchers due to technology constraints. Existing imaging techniques, such as confocal microscopy, use light sources that can damage an embryo before it has developed, and are prone to light scattering that blurs images. To work around this problem, researchers developed protocols like only reviewing a specific organ in an embryo at a time, or capturing images of several developing embryos at different stages and then combining them to get a clearer picture of total development. However, these protocols have been disappointing for documenting development, since the embryos cannot be completely imaged over longer periods of time.
However, Philip Keller—at the time a researcher at the European Molecular Biology Laboratory (EMBL)—created a new type of microscope in 2008 that could be used to evaluate whole embryos to study both cell processes that occur quickly (within seconds) and the entire embryo over the course of several days. The device works by illuminating only a thin plane of the embryo (called a light sheet) and recording the development of only that illuminated piece. The images are combined to create a whole image of the developing embryo.
Despite the breakthrough, the technique has one major flaw: it only works in transparent tissue. It worked with transparent specimens like zebrafish, but in organisms like fruit flies (Drosophila), the images would blur. “We found it difficult to obtain data of high quality at later stages of development, or for other model organisms that are less transparent than zebrafish,” Keller, now a fellow at the Howard Hughes Medical Institute (HHMI), told BioTechniques. “The main reason for this is that in these scenarios, the light is scattered on its way through the specimen. Unfortunately, conventional light sheet–based microscopy does not allow separating scattered from nonscattered light in the acquired images, which leads to blurry images with a low contrast.”
To correct the blurriness, the researchers turned to the University of Oxford where SI microscopy was being developed in the laboratory of Tony Wilson. SI microscopy helps generate clear, 3-D images of a specimen using stripes of light, rather than the constant sheet of light used by the original DSLM. Modulating the intensity of the laser light source in the microscope, the researchers shine a striped pattern on a single plane of an embryo. The researchers image each plane three times with different striped patterns, which are then combined to create a 3-D image; an algorithm is used to subtract any blurriness from the overlapping illumination patterns.
“Due to the enormous size of these data sets, it is impossible for a human to extract this information – we have to rely on computers running highly automated, robust image processing algorithms to extract this information for us,” said Keller. “For this reason, the microscopy images have to be of the highest possible quality. Processing low-quality images typically results in reconstruction artifacts that contaminate the database, which limits the usefulness of this database for further analyses.”
The new DSLM-SI imaging system provides optical sectioning (which is required for 3-D imaging) high speeds, low photobleaching, and low phototoxicity. Using images from a single day of development, Keller and colleagues created digital development models for zebrafish and Drosophila. “DSLM-SI allowed us for the first time to record detailed data sets showing zebrafish development from 9 to 67 hours post-fertilization, as well as early Drosophila embryonic development with subcellular resolution,” said Keller.
Keller’s laboratory is currently using the DSLM-SI technique to analyze the characteristics of neural development in the two species. “From these analyses we hope to be able to learn how a complete functional organism forms from a single cell,” said Keller. “In the long-term perspective, this may allow us to create and test computer models of development.”
The research was a collaborative effort of HHMI, EMBL, the University of Heidelberg, and the Sloan Kettering Cancer Center. The paper, “Fast, high-contrast imaging of animal development with scanned light sheet–based structured illumination microscopy,” was published July 4, 2010 in Nature Methods.