High-Throughput In Vivo Screening
Many physiological processes cannot be replicated in vitro and require use of animal models such as zebrafish. Visualizing zebrafish organs often requires physically manipulating larvae for proper orientation in a viscous medium such as agar. This has been a major impediment to the development of high-throughput zebrafish screening assays.
Engineers at Massachusetts Institute of Technology have developed a new technique that can analyze larvae in seconds. “We flow embryos through our instrumentation by loading them from reservoirs to an imaging platform where they are automatically rotated to the optimal position for imaging. The animals are unharmed by the process,” explained Mehmet Fatih Yanik, Ph.D., associate professor of electrical engineering and computer science.
With a resolution that is high enough to visualize individual cells inside various organs, the entire process takes less than 10 seconds per animal, compared to about 10 minutes for manual manipulation. “We are able to perform drug screens with our own chemical libraries for several diseases.
“In Parkinson disease, specific dopamine neurons in the brain are lost. We have a toxin model that kills the same cells that deteriorate in the human brain due to the disease process. We are screening for drugs that protect cells. This is very exciting work because it is the first time a large-scale chemical screen can be done on a vertebrate model. We hope to provide leads that may be of therapeutic value.”
Dr. Yanik's team is also studying epilepsy as well as spinal wound repair. “For our epilepsy studies, we induce seizures and then look for chemicals that modulate the hyperactivity. Additionally, we employ high-throughput surgery for our studies of wound repair. In this case, lasers are utilized to induce spinal cord injury. This allows us to screen potential compounds for enhancing nerve regeneration.”
When high-throughput screening is coupled with imaging, serious data overload can occur. “This is a big issue and one that we are currently addressing. Typical runs can generate up to 1 gigabyte of data per minute. We are developing algorithms to allow us to crunch the data more efficiently, such as how to automatically detect and identify different organs and count remaining neurons after compound treatment.”
Other scientists may soon be able to benefit from Dr. Yanik's approach. He is planning to commercialize the technology. “Because it is a high-throughput vertebrate screening platform with the capability of cellular-resolution imaging and manipulation, we anticipate it will be used for large-scale in vivo studies of many different types of diseases. We look forward to making it available to the entire research community.”
Successful imaging requires overcoming three key challenges, notes Stuart Logan, imaging specialist at PerkinElmer, who helped conduct an imaging workshop at the conference.
“The first challenge is dealing with live specimens, where phototoxicity issues can have damaging effects on the organism. Short wavelengths may cause DNA damage. Exposure to fluorescent light may also cause damage through heating. Other concerns are photobleaching (due to the signal rapidly fading under illumination) and the ability to detect low level fluorescence. Finally, because living systems are dynamic, the speed of capture must be fast enough to capture biological changes.”
A second challenge relates to the amount of data generated. “When running long time-lapse experiments, one generates huge volumes of data. It is common to generate greater than 30 gigabytes of data for a single experiment.”
Analyzing the data presents the third major challenge to successful imaging. “Image analysis requires accuracy, reliability, and comparability of statistical information. Software must be able to automatically generate quantitative results for comparison and publication from a large body of image files. Further, because there may be many different users, the software needs to be highly intuitive and easy to use.”
Claire Stewart, associate product manager, described PerkinElmer's approach to solving these issues. “For live-cell imaging, we offer our UltraVIEW® VoX 3D system. It is the only 3-D spinning disk system for acquisition to analysis. Spinning disk microscopy features rotating disks that bear thousands of pinholes on their surfaces. This allows simultaneous collection of multiple data points. One can use a lower dose of laser light because there is faster tracking of the sample. This technology, controlled by software designed for imaging of live specimens, allows capture and storage of images at a very high frame rate minimizing phototoxicity and photobleaching.”
For reliable, statistically unbiased analysis of large numbers of images, particularly from high-content imaging systems, the company suggests the Columbus™ Image Data Management and Analysis System. “Multiple users can easily store, access, and evaluate their data. Further, data can be accessed, evaluated, and measured using a web browser.”
For a more in-depth analysis, PerkinElmer offers Volocity® 3D Image Analysis Software. “To truly understand the biology of your samples, you need to view and evaluate data in 3-D,” Stewart reported. “Only then can you obtain a more complete understanding of the complex interactions that occur among cellular structures.”
Zebrafish live-cell microscopy is gaining in popularity as a model system because it can provide unmatched views of the inner complexity and workings at the resolution of the cell. As further improvements in the instrumentation and ability to analyze complex data occur, researchers will continue to develop more insights into the complex living processes from the single-cell level to functions within the entire organism.