Recent U-M BME Graduate Recognized for Leading Quiescence Research using AI
Quiescence has medical applications
During his time as a senior completing his undergraduate BME degree, Alec Eames led a study on using artificial intelligence (AI) to study cell quiescence (cells that are sleeping or dormant). Eames conducted his undergrad research under Dr. Sriram Chandrasekaran, Associate Professor, Biomedical Engineering. “Usually undergrads co-author papers as a team, but this study was fully carried out by Alec,” the professor noted. The AI tool that Eames developed has applications in drug discovery for cancer, regenerative medicine, and in reversing aging, according to Dr. Chandrasekaran.
Eames has since graduated and is working at Harvard Medical School. His paper, “Leveraging Metabolic Modeling and Machine Learning to Uncover Modulators of Quiescence Depth,” will be published soon by PNAS Nexus.
“This project started about a year and a half ago,” Eames said. “Dr. Chandrasekaran shared a very intriguing research article about quiescence, and that was the inspiration for my own research, which was on the same topic.”
Quiescence is traditionally thought of as a uniform cell state. But as we’re increasingly realizing, it exists on a spectrum from shallow to deep quiescence, Dr. Chandrasekaran noted.
“The analogy I like here is sleep,” Eames quipped. “There’s both light and deep sleep, and when you’re in a deeper sleep, it’s harder to be woken up. The same is true of quiescence. Cells that are in a deeper stage of quiescence are harder to activate.”
So making sure the depth of quiescence is tuned just right is critical for overall cellular health. For example, during an injury, quiescent stem cells need to be activated to regenerate lost tissue. However, stem cells enter deeper quiescence with age and lose regenerative capacity. “We’re starting to see that quiescence depth is often affected in diseases and during aging,” said Dr. Chandrasekaran.
“The first part of the research article was mainly about getting a better sense of what actually takes place in a cell when quiescence deepens. We used an approach called metabolic network modeling to simulate the metabolism of cells and subsequent epigenetic changes that take place as quiescence deepens,” Eames said.
As Eames explained, this is a prelude to the modulating part. “For this, we developed a machine learning tool (a type of AI algorithm) to predict quiescence depth,” he said.
“Essentially, this tool takes the activity levels of genes in a given cell to provide a score that quantifies the depth of quiescence for that cell. We found that there are certain genes that are able to predict the depth of quiescence remarkably well across different cell types and experimental conditions,” he added.
Eames used this tool for finding drugs and drug targets that might alter the depth of quiescence. He screened more than a thousand small molecules and genes for their ability to alter quiescence depth. Through this computational study, he found that a remarkable percentage of the predictions seem to align with experimental evidence. For example, many of the chemical compounds that were predicted to deepen quiescence are known to halt growth of cancer cells.
“We also found that a high proportion of the genes that were predicted to deepen quiescence are considered to be tumor suppressor genes, meaning they slow down cell growth and cancer proliferation,” Eames said. “On the flip side, we found that many of the genes that were predicted to move a cell to shallower quiescence are considered to be cancer oncogenes or cancer drivers–the kind of genes that tend to activate cell growth.”
Dr. Chandrasekaran said that this approach could have long-term implications for identifying therapeutics for tissue regeneration and treating diseases where quiescence depth is altered. “In some cancers, for example, there are cells that become quiescent and are difficult to kill with existing drugs, so this approach might help in finding new quiescence-lowering drugs to address these cancer types,” Eames added.
This study was funded by grants from the National Institute of General Medical Sciences, part of the U.S. National Institutes of Health, and the Camille and Henry Dreyfus Foundation to Dr. Chandrasekaran.