A look inside stem cells helps create personalized regenerative medicine

In two papers from the lab of Ahmet F. Coskun, an early career Bernie Marcus Professor in the Coulter Department of Biomedical Engineering at the Georgia Institute of Technology and Emory University, researchers examined a specific type of stem cell with a box intracellular tools to determine which cells are most likely to create effective cell therapies.

“We study the placement of organelles in cells and how they communicate to help better treat disease,” Coskun said. “Our recent work proposes the use of an intracellular toolbox to map the biogeography of organelles in stem cells, which could lead to more precise therapies.”

Creating the Subcellular Omics Toolkit

The first study – published in Scientific Reports, A Nature portfolio diary — looked at mesenchymal stem cells (MSCs) which have historically offered promising treatments for repairing faulty cells or modulating the immune response in patients. In a series of experiments, researchers were able to create a data-driven, single-cell approach through rapid subcellular proteomic imaging that enabled personalized stem cell therapy.

The researchers then implemented a rapid multiplexed immunofluorescence technique in which they used antibodies designed to target specific organelles. By emitting fluorescent antibodies, they tracked wavelengths and signals to compile images of many different cells, creating maps. These maps then allowed researchers to see the spatial organization of organelle contacts and geographic distribution in similar cells to determine which cell types would best treat various diseases.

“Usually stem cells are used to repair faulty cells or treat immune diseases, but our micro-study of these specific cells has shown how different they can be from each other,” Coskun said. “This proved that the patient population being treated and the personalized isolation of stem cell identities and their bioenergetic organelle function must be considered when selecting the tissue source. In other words, in the treatment of specific disease, it might be better to harvest the same cell type from different locations depending on the patient’s needs.”

RNA-RNA proximity matters

In the next study published this week in Cell report methods, the researchers took the toolbox one step further, studying the spatial organization of several neighboring RNA molecules in individual cells, which are important for cellular function. The researchers evolved the tool by combining machine learning and spatial transcriptomics. They found that analyzing variations in gene proximity for classifying cell types was more accurate than analyzing gene expression alone.

“Physical interactions between molecules create life; therefore, the physical locations and proximity of these molecules play an important role,” Coskun said. “We created an intracellular toolbox of subcellular gene neighborhood networks in the different geographic parts of each cell to take a closer look at this.”

The experiment consisted of two parts: the development of computational methods and experiments on the laboratory bench. The researchers looked at published datasets and an algorithm to group RNA molecules based on their physical location. This “nearest neighbour” algorithm helped determine gene clusters. On the bench, the researchers then labeled the RNA molecules with fluorescents to easily locate them in individual cells. They then discovered many features of the distribution of RNA molecules, such as how genes are likely to be in similar subcellular locations.

Cell therapy requires many cells with very similar phenotypes, and while there are unknown cell subtypes in therapeutic cells, researchers cannot predict how these cells will behave when injected into patients. With these tools, more cells of the same type can be identified and distinct subsets of stem cells with uncommon genetic programs can be isolated.

“We’re expanding the toolkit for the subcellular spatial organization of molecules — a ‘Swiss army knife’ for the field of subcellular spatial omics, if you will,” Coskun said. “The goal is to measure, quantify and model several independent but also interrelated molecular events in each cell with multiple functionalities. The end goal is to define the function of a cell that can reach neighborhood networks of genes high-energy Lego-like modular units and various cellular decisions.”

This research is funded by Regenerative Engineering and Medicine at Georgia Tech, as well as the NSF Engineering Research Center for Cell Manufacturing Technologies (CMaT).