Tissue clearing techniques make it possible to image three-dimensional structures deep inside a sample, which is particularly important for structures that are not confined within single imaging planes. Tissue clearing techniques have been available for a century, but the last 5 years have seen a spree of developments in this area. “People have wanted to be able to look at the whole, not just each of the parts,” says Marc Tessier-Lavigne from Rockefeller University. Both reducing the differences in refractive indices within the tissue and removing colored tissue components improve the optical transparency of biological samples. Ideally, either these techniques preserve the fluorescence of reporters such as GFP, or they are compatible with immunohistochemistry.

“After more than a decade since the completion of human genome projects, which enables systems biology at [the] molecular and cellular level, there are now strong needs toward organism-level systems biology,” says Hiroki Ueda from the University of Tokyo about the recent wealth of developments in tissue clearing. However, tissue clearing protocols are optimized for matching refractive indices within brain tissue. So far, these protocols have not attempted to reduce light absorption by endogenous colored substances such as hemoglobin.

Axonal projections in the forearm of a mouse embryo. Figure reproduced from Renier et al., Elsevier.

Ueda and his team have addressed tissue decolorization with their CUBIC (clear, unobstructed brain imaging cocktails and computational analysis) perfusion protocol, whereas Nicolas Renier and Zhuhao Wu in Tessier-Lavigne's lab have optimized the combination of immunohistochemistry with tissue clearing in their iDISCO (immunolabeling-enabled 3D imaging of solvent-cleared organs) technique. Furthermore, imaging internal plant structures is now possible with a protocol developed by Janine Sherrier's team at the University of Delaware.

Almost by accident, Kazuki Tainaka and Shimpei Kubota in Ueda's lab discovered that a compound in the CUBIC cocktail, an aminoalcohol, leads to tissue decolorization as it releases heme from blood within samples. The researchers found that this effect is due to the properties of this aminoalcohol and not because of an extreme pH. Importantly, the moderately basic pH of the decolorization cocktail preserves the fluorescence of proteins such as GFP and even allows for immunohistochemistry. Ueda and his colleagues applied the CUBIC procedure to a variety of different mouse tissues, such as muscle, kidney and lungs, to embryos and even to an adult mouse (Tainaka et al., 2014)2. They obtained transparent samples in all cases and were able to achieve single-cell resolution during imaging.

In plants, obtaining transparent tissue suitable for imaging is a challenge, as pigments absorb light, and cell walls and cytoplasm have highly divergent refractive indices. Sherrier and her colleagues used a urea and glycerol–based cocktail to clear tissue from plants such as tobacco, Arabidopsis and maize (Warner et al., 2014)3. They could perform immunohistochemistry and image fluorescent proteins in these cleared tissues at a depth three times greater than in untreated tissue.

Tessier-Lavigne and his team wanted to combine tissue clearing with immuno-labeling and therefore optimized different steps of a current clearing protocol, which resulted in iDISCO. In their hands, antibodies that perform well on tissue sections lead to equally good results in various organs and specimens with their iDISCO procedure, as they demonstrated with a panel of 28 antibodies. The clearing protocol is rather fast, taking between 8 and 18 days, depending on the sample. Tessier-Lavigne says that “we are trying to push the envelope and see if we can shorten the timelines even more.”

The researchers applied iDISCO to the adult or developing mouse brain, tongue, skin or an entire forearm, kidneys, muscle or vasculature (Renier et al., 2014)1. iDISCO does not preserve endogenous fluorescence well, but this disadvantage can be overcome by immunolabeling fluorescent proteins. Tessier-Lavigne does not consider the quenching of fluorescent proteins a limitation because immunolabeling allows them to use fluorophores in the red and far-red spectrum and avoid autofluorescence in the blue-green spectrum.

Both CUBIC perfusion and iDISCO should be amenable to use in other labs. “CUBIC is relatively easy to introduce and use,” says Ueda; Tessier-Lavigne thinks that “the activation energy to trying is very low” for iDISCO. Although researchers now have a choice between several protocols tailored to different needs, processing the large amounts of data that can be obtained from cleared samples may be the next bottleneck to overcome. “The framework to share huge three-dimensional image data is also one of the important challenges in the field,” says Ueda.