Research uses machine learning and imaging to gain insight into stem cell behavior

Machine Learning

Research uses machine learning and imaging to gain insight into stem cell behavior

A single hematopoietic stem cell dividing metabolically symmetrically and asymmetrically, as measured by the researchers' imaging technique. Credit: Hao Zhou and Keyue Shen / University of Southern California

Stem cells are like the human body's emergency toolkit. They have the unique ability to transform into other types of specialized cells, from immune cells to brain cells. They can divide and regenerate endlessly on command, repairing and replenishing the body's systems.

The ability to grow stem cells in the lab and grow them into any type of cell we need is the Holy Grail of medicine. This would allow clinicians to create an unlimited number of new cells to repair damaged tissues and organs, for example. But unearthing that Holy Grail requires a comprehensive understanding of how stem cells replicate and transform into different cell types.

New research from USC's Alfred E. Mann School of Biomedical Engineering brings us one step closer to unlocking these important cellular mysteries. Using machine learning, biomedical engineering associate professor Keyue Shen and his team have developed a noninvasive system that offers unseen insights into how stem cells grow and regenerate into specialized cells.

This work, Scientific advances.

Shen said the behavior of stem cells remains very mysterious, and the processes to understand how they divide and change are often invasive, requiring the stem cells to be extracted and ultimately destroyed in the laboratory.

The new study looked at hematopoietic stem cells, which reside in the bone marrow and give rise to all the cells in the blood, including red blood cells and immune cells. Stem cells need to divide symmetrically to proliferate, but they need to divide asymmetrically to produce new, different cell types (such as red and white blood cells) as they regenerate, Shen said.

“With bone marrow transplants, we want the stem cells to divide symmetrically so we have as many stem cells as possible that can be used across a range of patients, but right now we can't grow blood stem cells outside the body in the clinic,” Shen said. “If we could do that – if we could create a large stockpile of hematopoietic stem cells for bone marrow transplants – that would solve a huge problem for a lot of patients.”

Shen's team used a real-time imaging technique called fluorescence lifetime imaging microscopy to examine the metabolic behavior of stem cells – how they break down glucose for energy.

Stem cells produce their own fluorescent substance, called autofluorescence, which allows imaging to track the cell's metabolism, which is closely related to the cell's functions and changes.

“For example, NADH is one of the molecules that autofluoresces, and when bound to metabolic enzymes, it also exhibits different measurable optical fluorescence properties. This method allows us to measure it non-invasively without killing cells,” Shen said.

Shen and his team obtained this information using mouse models, extracting fluorescent properties from stem cell images and developing a library of 205 metabolic optical biomarker properties from individual stem cells, 56 of which were associated with hematopoietic stem cell differentiation.

The machine learning approach enabled the team to create a clustering map of stem and non-stem cells and track their behavior and differentiation over time, assigning a score to determine whether a daughter cell was likely to be a stem cell or whether the stem cell divided asymmetrically or symmetrically.

“It's really interesting because we're not killing the cells; we're just taking pictures of the cells and extracting their features. This gives us a whole lot of information about the cells.”

The team's real-time approach to understanding the metabolic state of stem cells will provide further fundamental knowledge to inform drug discovery and cutting-edge stem cell therapies, as well as regenerative medicine treatments to cultivate and replace human cells, tissues and organs.

“Nowadays there are other applications, such as cell therapy. People are trying to make T cells, macrophages, and other types of cells that have their own specific usefulness in different disease situations,” Shen said.

“For stem cell researchers, this is an exciting technology because it allows us to observe the state of stem cells in real time and then track individual cells over time, which is not possible right now.”

For more information:
Hao Zhou et al., Label-free metabolic optical biomarkers track stem cell fate transitions in real time, Scientific advances (2024). Source: sciadv.adi6770

Provided by University of Southern California

Quote: Study Uses Machine Learning and Imaging to Provide Insights into Stem Cell Behavior (July 4, 2024) Retrieved July 4, 2024 from

This document is subject to copyright. It may not be reproduced without written permission, except for fair dealing for the purposes of personal study or research. The content is provided for informational purposes only.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *