The prior technology can make the tissue as clear and transparent as glass and expand it to several times its original size, which provides an unprecedented opportunity for the internal operation of the biopsy system. Recently, Japanese researchers mapped the cellular composition of the mouse brain with unprecedented precision.
System biologist Ueda Taiji and his team at the RIKEN Biosystems Dynamics Research Center in Japan used a technique called CUBIC-X to map a mouse brain. They chemically labeled each cell in the brain and then expanded its size tenfold while transparent. Then, they used a sophisticated imaging technique to reconstruct the neurons in three dimensions, including a total of about 72 million cells, according to Ueda Taiji.
The resulting map narrows the brain into a compact database of cell locations from which the research team can study the specific changes in different brain regions during development. In the future, this database can also promote a deeper exploration of specific brain structures, such as areas that control sleep-wake cycles. The toolbox for this type of method is expanding, and CUBIC-X is just one of them. It uses off-the-shelf chemicals to open a window for researchers to view not only the brain, but also almost any other organ.
Some of these methods remove tissue to make opaque tissue transparent; others scale the tissue on the basis of tissue removal, allowing more detail to be exposed to traditional microscopes. Which method to choose depends on the specific scientific issue. There are many different ways to achieve similar goals, and users should figure out their strengths and limitations before deciding which method to choose.
Brain Imaging – Tissue Clearance Technology
Neuroscientists first had a need for tissue clearance techniques; they felt frustrated because they could not track the axons and dendrites in the brain.
Typically, such studies continuously image the labeled brain tissue sections and then perform a three-dimensional reconstruction by computer. But the process is very slow: it takes weeks to use the microscope to image the circuits in the mouse brain. The constructed map is completely dependent on the quality of the input data. Viviana Gradinaru, a neuroscientist at the California Institute of Technology, said, ‘In most cases, you only sample a portion of the slice, so the reconstruction is not efficient. In addition, cutting can damage the surface and edges of the tissue, making it impossible to reconnect them.’
Another better way is to make the tissue transparent and then fully image it. However, it has only become possible in recent decades that molecular reagents, genetic strategies and imaging technologies have made great strides.
To illuminate the interior of the brain, lipids are a top priority issue need to be solved. When light passing through an aqueous solution encounters a lipid surface, a change in refractive index causes it to bend and scatter. ‘Just like the jelly, but the brain is mainly composed of protein and is translucent,’ said Gradinaru, ‘If you add cream to the jellyfish, it is no longer transparent.’ The cell and organelle membranes are mainly composed of lipids, as are the myelin sheathing axons. Clearing brain tissue means clearing these lipid molecules, but retaining the rest of the molecules.
In 1911, the German anatomist Werner Spalteholz first demonstrated the method of removing opaque tissue, using chemical solvents to eliminate biomolecules that cause light scattering. However, this method is not suitable for today’s fluorescent reagents, and it will cause damage to the structure of the tissue.
In 2011, Hans-Ulrich Dodt, a brain imaging specialist at the Vienna University of Technology, and Ali Bratke, now at the University of Munich in Germany, and Frank Bradke of the German Center for Neurodegenerative Diseases described one of the earliest modern tissue removal techniques. The method, called 3DISCO, is based on the principle of Spalteholz, but uses a milder chemical solvent mixture to dissolve the lipid while preserving the cell structure, dehydrating and hardening the specimen into a transparent framework, thereby preserving the original structure of the tissue. Alain Chédotal, a developmental neuroscientist at the Institute of Vision at the National Institute of Health and Medical Research (INSERM), said, ‘We believe that solvent-based methods are the most reliable in terms of repeatability and cost.’
In general, solvent-based methods are generally best suited for the use of fluorescently labeled antibodies as reporter molecules, as naturally expressed fluorescent proteins tend to produce weaker signals or undergo denaturation after treatment. But the Ertürk team has improved 3DISCO to overcome this problem. vDISCO uses dye-labeled ‘nanobodies’ to enhance the signaling of fluorescent proteins in tissues after solvent scavenging – the team cleaned and completely imaged the mouse brain in this way.
Another widely used method of tissue clearance is CLARITY, which was developed in 2013 by Gradinaru in the laboratory of neuroscientist Karl Deisseroth of Stanford University. Deisseroth Laboratories uses fluorescent proteins extensively in its neuroscience research and is constantly seeking more ‘natural’ clearance methods to minimize damage to target biomolecules. CLARITY uses detergents to eliminate lipids while forming a water-based hydrogel through the polymer to strengthen the tissue framework. Gradinaru explains, ‘All these monomers are connected to each other and fixed in the protein. Then you can use this mild detergent to remove the lipids.’
The early CLARITY was technically difficult and required the active cleaning of the lipids wrapped in the detergent by the electric field. Gradinaru then designed a simpler alternative to infusing the animal’s vasculature with a cleaning solution to achieve the same effect. While Ueda has developed another method named CUBIC, which uses ‘hydrophilic’ chemicals to draw water into a fixed sample while pushing out the dissolved lipids. Like CLARITY, this method preserves the structure and function of the fluorescent protein while clarifying the sample.
Extended application of tissue removal technology
Tissue removal methods allow researchers to explore neuroscience issues that were previously unsearchable. For example, Chédotal is using 3DISCO to explore the structure of the entire vision system. ‘From the depth of the eye, the signals output by the ganglion cells may map to 30 different brain parts,’ he said. ‘We don’t know how these axons find different targets.’ But tissue cleansing methods don’t just apply to the brain. ‘What surprised us is that most rodent organs become transparent within a few days,’ Gradado said when she mentioned the perfusion-based CLARITY method she developed. ‘In addition to skin and bones, the entire body of the animal has been cleaned up.’
Gradinaru says bone is a particular challenge in tissue clearance because the residual calcium in the bone still reflects light after routine removal. However, additional processing can solve this problem. For example, Ueda’s team has proven that the laboratory’s commonly used chemical EDTA is effective in removing calcium from bones. Gradinaru’s team also developed a CLARITY method that removes bone tissue. These methods have made it possible to perform whole-body imaging of intact specimens, which researchers have used to track rare stem cell populations or tumor metastases, even the developing vasculature and peripheral nervous system in early human embryos. However, as specimens get larger, microscopy can become a bottleneck. Researchers must strike a balance between research goals and operability.
Light microscopy is a widely used coping strategy. ‘You are scanning a light plane that is passing through the tissue rather than a light spot, which greatly speeds up the imaging,’ Deisseroth said. In addition, some tissue removal methods have their own physical shrinkage effects. In the 3DISCO method, dehydration can reduce the sample to 50%. ‘This means we can shoot whole human embryos at once,’ says Chédotal.
But the current method can only achieve this. After tissue removal of the entire human brain, Chédotal is able to image only a few cubic centimeters or about 1% of the entire brain.
Image enlargement – specimen extension technology
Clearance is a valuable starting point for tissue imaging, but in order to understand the finer molecular aspects of the content, researchers also need special ‘magnification’ means.
Traditional optical microscopy cannot distinguish between molecules with a spacing smaller than the “diffraction limit” of light, which is about 200 nanometers – that has driven the development of sophisticated super-resolution microscopes. Ed Boyden, a neurobiologist at the Massachusetts Institute of Technology, was originally joking when he proposed to ‘blowing’ the brain, but soon he saw the true potential behind the idea. ‘All the proteins in the cells are crowded together, and if we separate them from each other, maybe we can see them better,’ he said.
Boyden’s team spent decades studying hydrogel properties that scaled proportionally during hydration. In 2015, they published the first generation of ‘amplification microscopy’ methods in which samples were processed with specially designed fluorescent labels to identify the target molecule of interest and then incubated with a polymer solution capable of forming a hydrogel matrix. These tag molecules attach to the substrate and lock their relative positions.
Finally, the surrounding tissue is broken by chemical or enzymatic treatment, and the gel matrix is expanded by hydration. The relative position of the tag molecules remains fixed after expansion. But the distance may be four times the original distance. Thus, molecules that were too close together can now be distinguished using standard fluorescence microscopy.
For many biologists, this approach looks incredibly great. Joshua Vaughan, a bioimaging researcher at the University of Washington, recalled, ‘My first reaction was this is crazy, how can it be achieved?’ But Vaughan was very interested in it and made a related attempt – then designed an alternative approach that is to use traditional fluorescent proteins or antibodies rather than specially designed labeling reagents.
Subsequent improvements include the Boyden group’s iterative approach: up to 20x magnification through two rounds of processing; and another method developed by MIT biomedical engineer Kwanghun Chung: denature biomolecules instead of digesting them, which will better protect the integrity of endogenous proteins and tissue structures. ‘We use this method to describe neural connections because in this case, we need to retain nerve fibers: once they are cut, they lose information,’ Chung said. Most importantly, these different methods can also induce tissue clearance, allowing researchers to drill down into oversized samples.
Vaughan’s team has used extended microscopes for the observation of multiple specimens, such as fly larvae, human kidneys, etc., while other researchers are applying them to clinical scenarios. However, specimens need to be ‘tendered’ before expansion, and this appropriate ‘tenderness’ requires trial and error. ‘There are some tough connective tissues in the human kidney,’ Vaughan said. ‘So we have to keep trying to treat it with enzymes. It is also crucial to maintain ‘isotropy’ in expansion or to be identical in all directions.’ But with careful operation and trial and error, even expansion is achievable. ‘Our results show that the distortion is less than 5%, which is similar to or even lower than the distortion rate when placing the sample,’ Chung said.
With the constant updating of immunological and genetic understanding, cancer medicine is rapidly developing. However, cancer pathology remains the same as in the past, relying on staining and microscopic observation of individual tumor tissue sections. Ueda Taiji, a systems biologist at the RIKEN Biosystems Dynamics Research Center in Osaka, Japan, said, ‘This method is very old and young clinical scientists are a bit frustrated.’
Organizational cleanup provides another option. For example, in 2017, Ueda and his colleagues showed that the tissue removal technology CUBIC has a higher sensitivity than traditional methods of preparation, allowing clinicians to observe larger tissue fragments and discover features that might otherwise be overlooked. ‘Sometimes early cancer is misdiagnosed,’ Ueda said. ‘But if you observe three-dimensional and tissue removal, you will rarely miss the diagnosis.’
Tissue clearance methods can also be used for tumor biology research. Karl Deisseroth, a neurobiologist at Stanford University, and PerUhlén at the Karolinska Institute in Sweden, have used an improved 3DISCO approach to gain an unprecedented understanding of tumor heterogeneity – tumor heterogeneity can profoundly affect treatment response. Ed Boyden, a neurobiologist at the Massachusetts Institute of Technology, and his team have demonstrated that tissue clearance techniques and high-resolution amplification microscopy help to more accurately detect early breast cancer lesions.
Both Boyden and Deisseroth founded the company to promote the clinical development of its technology. Some practitioners of organizational cleanup have foreseen that traditional methods are coming to an end.
Future prospects
For under-funded biologists, the extended microscope is the gospel, enabling ultra-high-resolution imaging without the need for expensive hardware. ‘Basically anyone can operate it,’ says Helge Ewers, a cell biologist at the Free University of Berlin. ‘Any ordinary microscope can be an ultra-high resolution microscope.’ Most importantly, the reagents you need don’t have to be specially designed, just need to try again and again to find the right ratio.
Chédotal pointed out that tissue cleaning of the entire mouse costs about $1. The equipment requirements are not high. ‘You only need to put the organs in the chemical. The only equipment needed is the incubator and the shaker – most laboratories have it.’ Ueda said. From a technical point of view, specimen expansion is more demanding than pure tissue removal, but Boyden and others have published ‘best practices guidelines’ that reduce the amount of unnecessary exploration by researchers when using amplified microscopes. “This is not exactly a ‘operational step book’, but we are working hard to get close,” Boyden said.
Researchers have begun to explore how to effectively use these methods. Multiple brain mapping projects are underway, with the goal not only for individual cells, but also for connections between cells. Other researchers are trying to clear and expand DNA and RNA to delve into gene expression and protein.
A biophysicist at Harvard University, Zhuang Xiaowei, designed an extended microscope-based method to quantify the expression of more than 100 genes at a single cellular RNA level. At the same time, Deisseroth and colleagues developed a technique called STARmap that allows direct sequencing of 1,020 different genes in the brain after tissue clearance.
The data obtained by these methods can be used to identify individual cells. But it can also be combined with loop diagrams and live animal experiments to reveal the interrelationships between the internal structure and function of the brain and those previously undiscovered connections.
‘Because the three-dimensional position of the cells in the tissue is preserved, we should be able to record transcriptomic data at the cell level, connect omics data, and live patterns of live animal cell levels collected in real time,’ Deisseroth said. ‘Organic clearance can make these completely different data combined.’