MIT Develops Ultrafast Protein Labeling Technique for Millions of Densely Packed Cells | MIT News

A new technology developed at MIT allows scientists to label proteins in millions of individual cells in completely intact 3D tissues with unprecedented speed, uniformity, and versatility. Using this technology, the team was able to label large tissue samples in a single day. In their new study in   Nature Biotechnology,   they also  show that the ability to label proteins with antibodies at the level of individual cells in large tissue samples can reveal insights that are obscured by other widely used labeling methods. 

The profile of proteins that cells produce is one of the most fundamental studies in biology, neuroscience, and related fields, because the proteins that a cell expresses at a given time can reflect the functions that the cell is trying to perform or its response to its conditions, such as disease or treatment. While microscopy and labeling technologies have advanced and enabled countless discoveries, scientists still lack a reliable and practical method for monitoring protein expression at the level of millions of individual cells densely packed into intact 3D tissues. As a result, scientists are often limited to thin tissue sections under glass slides, lacking the tools to fully assess cellular protein expression in the interconnected systems in which it occurs.

Examining intracellular molecules typically requires separating tissue into individual cells or cutting it into thin slices because the light and chemicals needed for analysis cannot penetrate deep into the tissues. “Our lab has developed technologies like CLARITY and SHIELD that allow the examination of entire organs by making them transparent, but now we need a way to chemically label entire organs to gain useful scientific insights,” says Kwanghun Chung, lead author of the study and an assistant professor in the Picower Institute for Learning and Memory and Science at MIT. “If cells in a tissue are not processed uniformly, they cannot be compared quantitatively. With conventional protein labeling, it can take weeks for these molecules to diffuse into intact organs, making uniform chemical processing of tissues on an organ scale virtually impossible and extremely slow.

A new approach called “CuRVE” represents a major step forward—years in the making—toward this goal by demonstrating a fundamentally new approach to uniformly processing large, dense textures. In this paper, the researchers explain how they overcame technical hurdles through an implementation of Curve, called “eFLASH,” and provide numerous clear demonstrations of the technology, including how it is bringing new insights to neuroscience.

“This is a significant leap, especially in terms of the actual functionality of this technology,” says Dae Hee Yun PhD ’24, an MIT graduate student who is now a senior software engineer at LifeCanvas Technologies, a startup Chang founded to publish tools invented in his lab. The paper’s other lead author is Young-Gyun Park, a former MIT postdoc who is now an assistant professor at KAIST in South Korea.

Smart chemistry

The main reason why it is difficult to uniformly label large, three-dimensional tissue samples is that antibodies penetrate tissue very slowly but bind rapidly to their target proteins. The practical effect of this speed mismatch is that immersing the brain in a bath of antibodies means that proteins are well labeled at the outer edge of the tissue, but virtually none of the antibodies find the cells and proteins at greater depths.

To improve labeling, the team considered a way—the conceptual nature of Curve—to resolve the speed mismatch. The strategy was to continuously monitor the rate of antibody binding while simultaneously increasing the rate of antibody penetration through the tissue. To understand how this might work and optimize the approach, they created and ran a complex computational simulation that allowed them to experiment with different settings and parameters, including different attachment speeds, densities, and texture combinations.

Then they decided to implement their approach in a real-world context. Their starting point was an earlier technology called “SWITCH,” in which Chang’s lab had devised a way to temporarily turn off antibody binding, allow the antibodies to penetrate tissue, and then turn the binding back on. As it worked, the team realized that significant improvements could be made if the rate of antibody binding could be continuously controlled, Yoon says, but the chemicals used in SWITCH were too harsh for such continuous treatment. So the team scoured a library of similar chemicals to find ones that could more precisely and consistently slow the rate of antibody binding. They found that deoxycholic acid was an ideal candidate. Using this chemical, the team could not only modulate antibody binding by changing the concentration of the chemical, but also by changing the pH (or acidity) of the labeling bath.

Meanwhile, to speed up the movement of the antibody through tissues, the team used another previous technology invented in Chang’s lab: random electrical conduction. This technology accelerates the dispersion of antibodies in tissues by applying electric fields.

Implementing this accelerated eFLASH dispersion system with a continuously variable binding rate yielded a wide range of labeling successes demonstrated in the paper. In total, the team reported using more than 60 different antibodies to label proteins in cells from large tissue samples.

Remarkably, each of these samples was labeled within a day, the authors say, an “incredibly fast” speed for intact, complete organs. Furthermore, the different preparations did not require new optimization steps.

Valuable visualizations

One way the team tested eFLASH was to compare their labeling with another common method: genetically engineering cells to fluoresce when the gene for the protein of interest is transcribed. The genetic method doesn’t require dispersing antibodies throughout the tissue, but it can be prone to inconsistencies because the message about gene transcription and the actual protein production aren’t exactly the same. Yoon added that while antibody labeling reliably and immediately reports the presence of a target protein, the genetic method can be much less immediate and persistent, and can continue to fluoresce even after the actual protein is no longer present.

In this study, the team used both types of labeling simultaneously on samples. By visualizing the labels in this way, they saw many examples in which antibody labeling and genetic labeling differed significantly. In some regions of the mouse brain, they found that two-thirds of neurons expressing PV (a protein prominent in certain inhibitory neurons) based on antibody labeling showed no genetic fluorescence. In another example, only a small fraction of cells that reported expression through genetic means of a protein called ChAT also reported it through antibody labeling. In other words, there were cases where genetic labeling significantly under- or over-reported protein expression compared to antibody labeling.

The researchers do not aim to question the obvious value of using genetic reporting methods, but instead suggest that the use of organ-level antibody labeling, as enabled by eFLASH, could help place these data in a richer and more complete context. “Our discovery of a large regional loss of PV-responsive neurons in healthy adult mice with high individual variability highlights the importance of holistic and unbiased phenotyping,” the authors write.

Or as Yun puts it, the two different types of labels are “two different tools for the job.”

In addition to Yoon, Park, and Chang, other authors of the paper include Jae-Hoon Cho, Lee Kamensky, Nicholas Evans, Nicholas DiNapoli, Catherine Zee, Seo-Woo Choi, Alexander Albanese, Yuksuan Tian, ​​​​Chang-Ho Sun, Qianghe Zhang, Minyoung Kim, Young-Ganyon Dee-Webster, Roengster Gonizda,

Financování této studie poskytly Burroughs Welcome Fund, Searle Scholars Program, Packard Award in Science and Engineering, NARSAD Young Investigator Award, McKnight Foundation, Freedom Together Foundation, Picower Institute for Learning and Memory, NCSOFT Cultural Foundation a National Institutes of Health.