Tissue processing advancements now allow for the labeling of proteins at the individual cell level across entire, intact rodent brains and other large samples quickly and consistently, just as effectively as in isolated single cells.
A groundbreaking technology from MIT empowers researchers to label proteins in millions of individual cells within fully intact 3D tissues with exceptional speed, consistency, and adaptability. Thanks to this innovation, the team could extensively label complete rodent brains and other substantial tissue samples in just one day. Their recent publication in Nature Biotechnology confirms that labeling proteins with antibodies at a single-cell level throughout whole brains can unveil insights obscured by conventional labeling techniques.
Analyzing the proteins produced by cells is vital in the fields of biology, neuroscience, and related areas, as the proteins a cell expresses at any moment can indicate its functions or reactions to conditions such as diseases or treatments. Despite significant advancements in microscopy and labeling technologies that have led to countless discoveries, researchers have struggled to effectively track protein expression among millions of densely packed individual cells within whole, 3D intact tissues, like a full mouse brain or a complete area of the human brain. Typically restricted to thin tissue sections viewed under slides, scientists have lacked the necessary tools to thoroughly understand protein expression across unified and connected biological systems.
“Traditionally, studying molecules within cells necessitates breaking down tissue into single cells or slicing it into thin sections because the light and chemicals required for analysis cannot penetrate very far into tissues. Our lab developed techniques like CLARITY and SHIELD to examine whole organs by making them transparent, but we now needed a method to chemically label entire organs for valuable scientific insights,” explained Kwanghun Chung, the senior author of the study and an associate professor at The Picower Institute for Learning and Memory, along with the Departments of Chemical Engineering and Brain and Cognitive Sciences, and the Institute for Medical Engineering and Science at MIT. “Think of marinating a thick steak by briefly dipping it into sauce. The outer layers absorb the marinade quickly and intensely while the inner layers remain mostly untouched unless soaked for a longer time. The same concept applies to the chemical treatment of tissues: if cells within a tissue aren’t consistently processed, they can’t be quantitatively compared. The challenge is amplified for protein labeling, as the chemicals we use for labeling are significantly larger than those in marinades. Consequently, it can take weeks for these molecules to diffuse into intact organs, rendering uniform chemical processing of large tissue samples nearly impossible and exceedingly slow.”
The newly developed “CuRVE” technique marks a substantial step forward, years in the making, by demonstrating a fundamentally different approach to uniformly processing large and dense tissues in their entirety. In their study, the researchers describe how they overcame various technical challenges using a version of CuRVE named “eFLASH” and showcase eye-opening demonstrations of the technology, highlighting its potential for novel neuroscience findings.
“This represents a remarkable advancement, particularly regarding the actual performance of the technology,” commented co-lead author Dae Hee Yun, a former MIT graduate student now serving as a senior application engineer at LifeCanvas Technologies, a startup founded by Chung to spread the innovations developed in his lab. The other lead author is Young-Gyun Park, a former postdoc at MIT who is now an assistant professor at KAIST in South Korea.
Clever chemistry
The primary challenge in uniformly labeling large 3D tissue samples lies in the slow seepage of antibodies into tissues, while they bind rapidly to their target proteins. This speed mismatch leads to a scenario where simply soaking a brain in antibody solution will result in strong labeling of proteins on the outer layer, but minimal labeling of deeper cells and proteins.
To enhance labeling, the team devised a solution that encapsulates the essence of CuRVE: a strategy to continuously manage the speed of antibody binding while simultaneously expediting their penetration into the tissue. They crafted and executed an advanced computational simulation to test various parameters, including binding speeds and tissue compositions.
Next, they moved forward to apply their method to actual tissues, building on a previous technology called “SWITCH,” where Chung’s lab found a way to temporarily halt antibody binding, allowing antibodies to permeate the tissue, followed by restarting the binding process. Although successful, Yun noted that they recognized potential for considerable improvement if they could control the binding speed continuously. However, the chemicals used in SWITCH were too harsh for ongoing application. The team then screened a library of similar substances to discover one capable of subtly and continuously regulating the binding speed, ultimately selecting deoxycholic acid as the optimal candidate. By using this chemical, the team could adjust antibody binding through variations in its concentration and the acidity (pH) of the labeling solution.
To facilitate faster movement of antibodies through the tissues, they also used another earlier innovation from Chung’s lab: stochastic electrotransport, which accelerates the spread of antibodies within the tissue by applying an electric field.
Applying the eFLASH system for rapid dispersion alongside continually adjustable binding speed led to the multitude of labeling successes reported in their study. In total, the researchers utilized over 60 different antibodies to label proteins in cells throughout the complete brains of mice and rats, entire mouse embryos, various mouse organs like the lung and heart, and brain tissue blocks from larger species, including humans.
Remarkably, all these specimens were labeled within a single day, which the authors described as “ultra-fast” for entire organs. Additionally, each preparation did not require new optimization steps.
Valuable visualizations
Among the tests conducted with eFLASH was a comparison against another common technique: genetically engineering cells to emit fluorescence when a gene for a target protein is being transcribed. While genetic methods do not require antibody dispersal throughout the tissue, they can be prone to discrepancies because gene transcription does not always directly correlate with actual protein production. As Yun noted, antibody labeling provides a reliable and immediate indication of target protein presence, while the genetic approach may be less timely and can continue to fluoresce even after the protein is no longer present.
In their study, the team utilized both labeling methods simultaneously. Through this dual visualization, they identified numerous instances where the antibody and genetic labeling results varied significantly. In certain areas of mouse brains, they discovered that two-thirds of neurons that showed PV (a protein prominent in certain inhibitory neurons) expression via antibody labeling did not exhibit any fluorescence from the genetic method. In another case, only a minuscule fraction of cells that demonstrated expression through the genetic method for a protein called ChAT were also recognized through antibody labeling. This indicates cases where genetic labeling either drastically underreported or overstated the actual protein expression compared to antibody labeling.
The researchers do not intend to undermine the utility of genetic reporting techniques; rather, they advocate for the inclusion of organ-wide antibody labeling, enabled by eFLASH, to provide broader context for the data. For example, since they noted considerable over-reporting of PV expression in mice (with significant variation among individuals), researchers utilizing whole-brain antibody labeling of PV could obtain a more thorough understanding for analyzing genetically indicated changes in the protein.
“Our findings regarding large regional losses of PV-immunoreactive neurons in healthy adult mice with high individual variability highlight the necessity of comprehensive and unbiased phenotyping,” the authors stated.
As Yun expressed it, the two forms of labeling serve as “different tools for the job.”
Along with Yun, Park, and Chung, the paper lists additional authors: Jae Hun Cho, Lee Kamentsky, Nicholas Evans, Nicholas DiNapoli, Katherine Xie, Seo Woo Choi, Alexandre Albanese, Yuxuan Tian, Chang Ho Sohn, Qiangge Zhang, Minyoung Kim, Justin Swaney, Webster Guan, Juhyuk Park, Gabi Drummond, Heejin Choi, Luzdary Ruelas, and Guoping Feng.
The research received funding from various organizations, including the Burroughs Wellcome Fund, the Searle Scholars Program, a Packard Award in Science and Engineering, a NARSAD Young Investigator Award, the McKnight Foundation, the Freedom Together Foundation, The Picower Institute for Learning and Memory, the NCSOFT Cultural Foundation, and the National Institutes of Health.
