Sarah Cohen and her collaborators received a 2024 Allen Distinguished Investigator Award to develop a game-changing microscopy tool for visualizing multiple organelles in live cells.
Cells are anything but stagnant. From wound healing to metastasis, diverse cell types shuttle around the body to complete vastly different tasks. What few people tend to notice, though, is that within each cell exists a bustling, dynamic metropolis of organelles that is also constantly moving to manipulate cell behavior and coordinate cell function.
Typically, scientists can only capture a snippet of this metropolis by staining a couple of organelles at a time and visualizing them using common microscopy techniques that often affect cell viability. Electron microscopy, for example, requires cells to be dead, fixed, and sectioned, which prevents the ability to draw conclusions on complex organelle behavior in live cells and fully capture 3D organelle structure. Similarly, fluorescence microscopy uses lasers that can damage cells, preventing the long-term visualization of organelle changes.
Now Sarah Cohen, an assistant professor in the Department of Cell Biology and Physiology at the University of North Carolina at Chapel Hill, and her collaborators Assaf Zaritsky from Ben Gurion University and Shalin Mehta from the Chan Zuckerburg Biohub, San Francisco, are pioneering a new method to visualize and perform multiparameter analysis on eight organelles at a time in live cells by combining label-free imaging, fluorescence microscopy, and machine learning. The trio recently received a 2024 Allen Distinguished Investigator Award, a $1.5 million award that funds the development of paradigm-shifting tools for the research community.
Answering, “What happens to organelles during differentiation?”
“I’m very excited about this award because it’s a collaboration with two of my colleagues. It’s an interdisciplinary project with the high-level goal of understanding how cells reorganize during development to support different physiologies,” said Cohen. During development, naïve cells differentiate into specialized cell types. When this happens, their organelles also change in shape, frequency, and organelle-to-organelle contact.
Previously, Cohen’s team observed several changes in organelle contacts during stem cell differentiation into neurons. Their mitochondria began touching the endoplasmic reticulum (ER) less and lysosomes more. Stem cells use glycolysis for metabolism, which enlists the mitochondria to make building blocks for actively dividing cells, but specialized cells no longer grow or divide and switch their metabolism to oxidative phosphorylation. “This requires a shift in the morphology and protein composition of the mitochondria. Increased lysosome contact may be involved in removing the parts of the mitochondria that are no longer needed,” said Cohen. As the cells began to transform into neurons, her team also saw an increase in peroxisome organelle contact with the ER, which is important for synapse formation.
These results suggest that changes in organelle contacts could play key roles in supporting cell differentiation and function. “We’re proposing to follow up on these observations but then also extend them to different cell types, including heart muscle cells and liver cells,” said Cohen.
A new tool for systems-level microscopy data analysis
To capture changes in organelle contacts during cell fate specification, the trio will develop a label-free machine learning approach. This approach will be gentler on cells and unlike existing technologies allow them to visualize the entire differentiation process.
They will use multispectral imaging, a technique developed by Cohen in 2017, to fluorescently label eight organelles at a time. They will also capture label-free brightfield images of the same cells. They will then input both the multispectral and label-free images into a machine learning algorithm.
“We can train the computer to know in label-free images what different structures are so that in the longer term, we could do this kind of microscopy without all the fluorescent labels but still be able to tell which structures are which,” said Cohen. “We can then use these machine learning approaches to tell us what’s important and what is going on in a cell in an unbiased way.”
Driving change
Once the team has identified what organelle contacts are important for a specific cell type, they plan to disrupt those contacts to see if they can prevent differentiation. Conversely, they will also try to drive changes in cell function and differentiation by forcing different organelles to make contact using artificial tether proteins. “A year ago, I would have said that driving a single organelle contact would not be enough to affect cell physiology, but we’re starting to see indications of that,” said Cohen.
Recent research suggests that these small changes can play big roles in human health and disease. Many mutations that cause neurodegenerative diseases are in proteins that interface between organelles and regulate organelle connections. Anecdotal findings from other research teams also suggest that slight changes in ER organelle morphology can drastically affect T cell function and exhaustion.
Whether changes in organelle contacts can cause meaningful consequences to human health remains unclear. For now, the first step in answering this question is developing better tools for organelle visualization in live cells. “I’m really excited to establish these collaborations and develop new methods for studying organelle communication that I hope will be useful for other groups beyond ours,” said Cohen. “And I’m excited to test this crazy idea of whether driving organelle contacts can change cell function.”
Reference
Valm, A.M., Cohen, S., Legant, W.R. et al. Applying systems-level spectral imaging and analysis to reveal the organelle interactome. Nature 546, 162-167 (2017).
Written by Tiffany Garbutt, PhD
Read the press release from the Allen Institute here