Scattered against a black background, vivid blue, beige, and orange dots—32 of them to start—commence an amazing dance. They quickly double in number, shrink, and double again. They fill one pole of a slowly spinning, invisible globe. As they continue to multiply, becoming hundreds, then thousands, of little points, they swarm, covering the globe evenly in brilliant blue with specks of orange flashing in and out of existence. The dots then crowd the equator and meld into the shape of a fish embryo.
The video is like computer-animated pointillism. But rather than a piece of digital art, it represents a scientific feat—a stunning series of images that reveals the development of a live zebrafish embryo over 24 hours. Each cell in the embryo is represented by a single dot, colored blue when still or moving slowly and shifting to beige and then orange when migrating more quickly.
This detailed way of watching development in action is the result of the latest version of the light-sheet microscope, developed by a team of scientists at the European Molecular Biology Laboratory.
(EMBL) in Heidelberg, Germany. The new tool has the potential, says team member Philipp Keller, now a fellow at HHMI’s Janelia Farm Research Campus, to achieve a goal coveted by developmental biologists: the generation of comprehensive computer models of embryogenesis in complex vertebrates.
Before this innovation, Keller notes that scientists had been able to reconstruct, in a comprehensive way, the development of only simple animals such as Caenorhabditis elegans—a tiny worm that hatches when it reaches just 500 to 600 cells, within 12 hours after fertilization.
|100 minutes after fertilization, a zebrafish has 64 cells, arranged in a seemingly random pattern. But over the next 24 hours, they morph into deliberate shapes. By the end of the first day of development, the embryo is in the Phyrangula stage, characterized by the emergence of a beating heart. Watch the transformation here. Video courtesy of Philipp J. Keller, EMBL Heidelberg|
Scientists studying the zebrafish, which grows to tens of thousands of cells on day one of its three-day embryonic development, had captured images of the embryo’s transformation into a juvenile. But they could describe the stages of development only in broad strokes. They had no tool to explore in detail the mysteries of gene expression, morphogenesis, and cell movement and division patterns. Existing microscopes often damaged embryos.
In addition, the strategy of patching together multiple images from different specimens left sizable holes in the resulting information. Part of the problem is that every embryo develops slightly differently. “If you stitch together data from different animals,” says Keller, “you don’t get the same coherent data set that you’d get by looking at one animal and observing it over time. Live microscopy was the only option.”
But in the early 2000s, microscopes fell short of the task.
“Neither confocal nor two-photon microscopes were fast enough,” says Keller. “The limitations in imaging speed do not allow following cell behavior for the entire organism, and, in the confocal microscope, the fluorescent markers would bleach very quickly and the embryo would be alive only for a short time.”
Joining Team EMBL
By 2005, Keller was working on his doctorate, becoming interested in studying life at its inception, and had already joined a team of scientists at EMBL led by Ernst Stelzer. The group had developed a new type of microscope for observing the previously invisible processes of embryonic development.
The scientists used a relatively simple trick from 100 years back. In 1903, chemist Richard Zsigmondy had invented the “ultramicroscope,” which illuminated a sample through a slit at a right angle to the viewing angle. Using the same principle, the EMBL researchers engineered a more sophisticated version for the 21st century, taking full advantage of the modern computing power needed to process and analyze massive amounts of data. They published the innovation in 2004 in Science, calling it selective plane illumination microscopy, or SPIM.
“Instead of collecting data point by point, using the same objective for illumination and fluorescence detection as the confocal microscope, we illuminate an entire plane [of the specimen] from the side,” Keller says. The only thing you have to do is collect the fluorescence emitted at a right angle from this plane using a camera with a conventional detection system. It’s relatively simple.”
Page 2 of 2
The inaugural version of the microscope yielded the first video of an embryo’s beating heart—that of a Medaka, or Japanese rice fish. After refining the instrument for version 2.0—dubbed “digital scanned laser light-sheet fluorescence microscopy,” or DSLM—Keller and his EMBL colleagues captured the first 24 hours of a zebrafish embryo’s development with unprecedented resolution and speed. They published their results in Science in 2008.
Progressing beyond day one of development became the challenge. As the embryo grew more complex, images blurred. The increasingly complicated and numerous structures in the rapidly developing fish scattered the light from the scope. To maintain high resolution of these complex images, Keller and his colleagues tweaked the scope so they had rapid electronic control over the pattern of light passing through the specimen. Increased control meant the instrument could accommodate the denser embryos of other lab animals, such as fruit flies and mice.
The scientists took nearly one million images over three days to follow neural development of a zebrafish embryo into its juvenile stage. They also created a “digital fly embryo,” a three-dimensional reconstruction of early Drosophila development with single-cell resolution. The group published their results in the August 2010 issue of Nature Methods.
The Next Generation
Since moving to Janelia in May 2010, Keller has been constructing the next generation of the light-sheet microscope. Using the new design and latest technology, he expects this iteration of the microscope to perform 40 times faster than the previous version. This faster speed and many additional capabilities will give scientists even more detailed information about the physical and chemical choreography that occurs during development.
Keller plans to use the revamped instrument to continue his embryology research, with a special focus on neural development in Drosophila. He also hopes to expand the technology’s reach to studying early development in other model organisms, including mice, the gold standard lab model for early investigations into human disorders.
In addition to the lab’s core research projects, Keller will have ample opportunities for collaboration. His work dovetails with that of his new Janelia Farm colleagues, including Hanchuan Peng and Gene Myers as well as Julie Simpson and Jim Truman. It was precisely for this kind of interaction that Keller came to Janelia, where integrated teams with very diverse backgrounds aim to break through existing barriers and solve problems.
Keller is working with Peng, a senior computer scientist at Janelia Farm, and Myers, a Janelia group leader, to implement computational solutions to managing the enormous amount of data the microscope will collect. Peng has created a three-dimensional digital map of the fruit fly brain, and his lab is developing a “smart” image acquisition method that can zero in on specific areas of the brain for analysis.
“These techniques,” Peng says, “may well fit with Philipp’s imaging pipeline—to reduce the data volume and produce quantitative analysis at the moment of [image] acquisition.”
Keller is looking to Janelia group leaders Simpson and Truman for their expertise in fly neuroscience, genetics, and novel labeling strategies. Simpson is investigating how genes, neurons, and neural circuits affect fruit fly behavior. “My optical microscopy expertise is limited to commercial confocals,” she says, “so I am eager to see what Philipp’s microscope can do with our specimens.”
The promise of live embryo imaging is unquestionable. Light-sheet microscopy will allow scientists for the first time to describe in detail the processes of development in complex vertebrates; to map the fates of cells as they become specialized; to track the effects of genetic mutations in a living embryo as it develops; and, of most interest to nonscientists, to witness and come to understand how developmental disorders arise.
“There is so much to be done, such an enormous potential,” Keller says. He’s not the only one to say so. In a review published in the same issue of Nature Methods as Keller’s paper, one imaging expert describes light-sheet microscopy and several other new imaging methods as part of a new frontier.
For the moment, using live imaging to answer developmental biology’s numerous lingering questions remains a sluggish endeavor. Keller says only a few dozen labs around the world have built versions of the microscope. But with a commercial version of the light-sheet microscope in the works, Keller hopes he’ll soon have plenty of company within developmental biology to help fulfill its potential.