A label-free approach to vascular imaging of living organisms

Confocal microscopy has fostered impressive advancements in cell biology in recent decades, thanks to the cornucopia of information achievable with 3D reconstructions of subcellular details. Now, new imaging techniques able to visualize entire organisms have made it possible to study functional and morphological processes in vivo. This could dramatically change developmental and systems biology.

The majority of 3D microscopy techniques are based on fluorescence contrast. Optical projection tomography (OPT) can use fluorescence, but does not have to: 3D reconstruction is obtained by acquiring a sequence of bright-field images of the sample.¹ Conceptually, bright-field OPT is the optical analog of x-ray computed tomography. The visible light, transmitted through the specimen, is recorded on a camera at several different orientations. Afterwards, the sequence of projections is elaborated with a back-projection algorithm to obtain virtual sections of the absorbing structures within the sample.

We have recently demonstrated a new approach to 3D vessel reconstruction by combining OPT with a digital motion-analysis algorithm. We called this technique flow OPT, because it allows vessel tomography by looking directly at the flow of blood cells.²

Our group applied flow OPT technology to the in vivo study of zebra fish (Danio rerio). D. rerio is increasingly used to model human diseases, to screen preclinical drugs and to study organogenesis and cellular functions. It is an ideal model for in vivo studies, presenting several advantages including amenability to in vitro manipulation, feasibility of reverse and forward genetic approaches, ease of drug administration and experimentation, and optical transparency at the embryo stage. Our focus has primarily been dedicated to the study of its complex circulatory system and its development from embryo to adult.

The flow OPT setup is equivalent to that of bright-field OPT. An LED sheds visible light onto the sample, which is mounted on a rotation stage and immersed in water. Transmitted light is collected on a camera using a microscope objective. We found that using telecentric lenses for illumination and detection was particularly convenient, as they help maintain constant magnification through the entire sample thickness. By recording successive short time frames for each projection, a flow contrast map can be extracted with the motion-analysis algorithm. The algorithm detects moving structures and highlights the traces of cells inside the bloodstream, making vessel distribution visible. Once contrast maps are extrapolated at different orientations around the sample, tomographic reconstruction is performed using back-projection algorithms. This procedure is shown in Figure 1, where contrast maps obtained at different angles and a reconstructed transverse section of the sample are presented.

Figure 1. Contrast maps obtained at different orientations are combined with a back-projection algorithm to obtain sections of the vascular system.

We visualized a detailed casting of the circulatory system of a zebra fish, from embryo to juvenile, using flow OPT. Figure 2 shows the reconstructed sections of the trunk of a zebra fish larva (four days post-fertilization). These images are merged with bright-field OPT sections, obtained simply by measuring the transmission of light through the sample.

Figure 2. Transverse (a), coronal (b), and sagittal (c) sections of the trunk of a zebrafish. 3D volume reconstruction of the whole region (d).

OPT is gaining more and more importance for its capability to quickly image large areas in a relatively short time. Flow OPT retains this feature, since several frames (each with a duration of a few milliseconds) can be acquired at 100Hz or beyond. The acquisition of the entire data set typically takes just a few minutes. Another key point of OPT is its low cost. Flow OPT takes this advantage even further, since it does not require fluorescence excitation and detection. The cost for such a system could be as low as $10,000. Finally, the advantage of not needing fluorescent labeling or markers is particularly important when wild-type organisms are under study and injection of fluorescent probes is inconvenient.

In summary, we have described a novel noninvasive technique that does not require fluorescent contrast for 3D vasculature imaging of small transparent animals. Time-lapse measurements could make flow OPT a valuable tool for research in developmental biology, vascular organogenesis, and normal and pathological angiogenesis. Our future work will develop similar technologies for 3D imaging of larger scattering samples that could be applied to the study of mouse organ development or artificial organs.