Whole-animal imaging, gene function, and the Zebrafish Phenome Project

Abstract

Imaging can potentially make a major contribution to the Zebrafish Phenome Project, which will probe the functions of vertebrate genes through the generation and phenotyping of mutants. Imaging of whole animals at different developmental stages through adulthood will be used to infer biological function. Cell resolutions will be required to identify cellular mechanism and to detect a full range of organ effects. Light-based imaging of live zebrafish embryos is practical only up to ∼2 days of development, owing to increasing pigmentation and diminishing tissue lucency with age. The small size of the zebrafish makes possible whole-animal imaging at cell resolutions by histology and micron-scale tomography (microCT). The histological study of larvae is facilitated by the use of arrays, and histology's standard use in the study of human disease enhances its translational value. Synchrotron microCT with X-rays of moderate energy (10-25 keV) is unimpeded by pigmentation or the tissue thicknesses encountered in zebrafish of larval stages and beyond, and is well-suited to detecting phenotypes that may require 3D modeling. The throughput required for this project will require robotic sample preparation and loading, increases in the dimensions and sensitivity of scintillator and CCD chips, increases in computer power, and the development of new approaches to image processing, segmentation, and quantification.

Figure 1. Comparison of Synchrotron Parallel Beam vs. Commercial Cone Beam MicroCT geometry

In cone beam microCT, the X-ray comes from a focal spot, for which diameter is inversely proportional to resolution and directly proportional to flux. The flux (brightness) is low. The cone shape of the beam allows magnification of the desired area by adjusting the relative position of the sample and scintillator to the focal spot, and can be used to focus a subarea of the specimen across the full area of the scintillator. These scintillators may be coupled by fiber optics to the cooled CCD. In synchrotron microCT, X-rays are of parallel geometry, monochromatic, phased, and of high flux. There is no geometric specimen magnification by the X-ray. Edge enhancement by phase contrast is made possible by phased monochromatic X-ray, and is adjusted by changing the sample-to-scintillator distance. The transmitted X-ray induces light in the scintillator, which is projected through optical lenses (e.g. 5x or 10x microscope objectives) whose magnification onto the cooled CCD determines optical magnification of the scintillator surface. For both types of microCT, the specimen is rotated over at least 180 degrees, commonly with one projection image taken every 0.1 to 0.3 degrees of rotation (yielding 1800 and 600 images, respectively).

Figure 2. Micron-Scale Computed Tomography (microCT)

A series of processing steps allows 3D models to be generated from microCT data. Demonstrated here is the application of these steps to reconstruct a zebrafish’s skull and inner ear. (A) One of 600 x-ray projection images taken, over ~4 hours, through 180° of the head of an unstained, 60 dpf juvenile zebrafish wrapped in parafilm (one image every 0.3°). (B) One of 1500 digital, coronal 2D slices generated by applying the Feldkamp cone-beam reconstruction algorithm to these 600 images. (C) A screen-capture of part of a 3D model generated from these 1500 2D slices in a Volume Graphics software package known as VGStudio Max (Heidelberg, Germany). The arrows labeled L, S, and A, are the lapillus, sagitta, and asteriscus, which are found in the utricl, saccule, and lagena, respectively [67]. A similar comparison can be accessed at http://www.zfatlas.psu.edu/comparison.php?s[]=262&s[]=267&s[]=268. Scale bar, 1mm.

Figure 3. Zebrafish Phenome Project Paradigm

For the phenome project, mutant and control animals will be generated for live and fixed animal studies. The imaging data will benefit from segmentation software, as well as software to facilitate analysis, visualization, and integrated interpretation (black boxes). The output, as well as raw data, would then become available to the zebrafish and other research communities.