In diseased cells, genes are in a state of disarray. Some people receive signals that cause them to produce too much protein. Some people have reduced activity to abnormal levels. Up is down and down is up.
The right molecules can restore order and reverse the dysregulation of certain genes. But finding the ideal compound may require studying the effects of millions of chemicals on hundreds or thousands of genes.
An MSU-led research team has demonstrated a better way. Using machine learning trained on vast amounts of publicly available data, they were able to predict how chemicals affect gene expression based solely on their structure.
Their research, recently published in the journal Cell, discovered a compound that holds promise for treating two difficult diseases: liver cancer, the most aggressive form of cancer, and chronic lung disease, for which there is no cure.
This discovery, which has implications for faster drug discovery, is the result of years of research across multiple disciplines and institutions, said one of the senior authors, Bing Chen, associate professor in the Department of Pediatrics, Human Development, Pharmacology, and Toxicology at the College of Human Medicine.
“So many people worked on this concept. We had over 20 researchers involved, and it’s been a long journey,” Dr. Chen said. His research focuses on collaborating with computer scientists, bench scientists, and clinicians to develop computational methods and tools for drug discovery.
This interdisciplinary approach was key to this project. It started by training a “gene expression profile predictor for chemical structure” (GPS) on millions of experimental measurements. Chen collaborated at this stage with another senior author, Dr. Jiayu Zhou, formerly of MSU and now at the University of Michigan.
Chen likened the process to training a neural network to classify images as people, cats, and dogs.
“With our approach, rather than looking at cats or dogs, we want to know whether the compound up-regulates or down-regulates the expression of a particular gene,” Chen said. “It’s still a classification issue, but it’s more of a biological cause.”
“But the biological data is rarely clean,” Zhou says. “Imagine trying to learn from a large number of examples, some clear, some ambiguous, and some misleading. Our approach helps models distinguish between stronger and weaker signals, allowing them to learn from the data without getting confused by all the noise.”
After evaluating the data for theoretical application to multiple diseases, the team selected two for real-world testing. Hepatocellular carcinoma (HCC) is the third most common cause of cancer-related death worldwide. Idiopathic pulmonary fibrosis (IPF) is a chronic lung disease with a median survival of 3 years after diagnosis.
Both diseases require new treatments, and both are of intense interest to researchers.
Chen previously discovered that anthelmintic drugs could be used to treat HCC. He and contributors Samuel Soh, MD, PhD, Louis Hak Min Professor and Senior Investigator Maysee Chua, PhD, of Stanford University’s Asian Liver Center, have been collaborating for many years with the goal of developing compounds that benefit HCC patients.
“Our efforts so far have been limited to repurposing FDA-approved drugs,” Chua said. “This new approach significantly expands the pool of novel compounds with potential therapeutic activity in HCC.”
“As the incidence of HCC continues to increase in the United States, novel and more effective compounds that can target the molecular heterogeneity of HCC could directly address an unmet clinical need,” So said.
MSU’s other senior author, Dr. Xiaopeng Li, an associate professor in the Department of Pediatrics and Human Development in the College of Human Medicine, whose research focuses on lung diseases such as IPF.
“We know this disease is difficult to tackle,” Lee said. “There have been so many failures in identifying new drugs over the last 20 years, and I think the AI component has allowed us to explore the problem in a different way and more systematically.”
Discovering compounds theoretically is another thing. They still need to be tested in the real world, said Dr. Edmund Ellsworth, director of the MSU Medicinal Chemistry Facility and professor in the Department of Pharmacology and Toxicology.
As contributors to the research, Ellsworth and his team were responsible for creating related compounds discovered by the platform and optimizing them into safe and effective drugs. This important step is just the beginning of a complex process, he said.
“To move forward, we need to recognize that drug discovery is a team sport and not for the faint of heart,” Ellsworth said. “It’s complex and all kinds of things can happen. It takes a diverse set of experts to navigate it and be successful.”
These compounds were tested in laboratory cell lines to confirm their effects on genes and identify potential candidates for in vivo testing.
When testing anti-HCC compounds in mice, the team discovered two new compounds that reduced tumor size. In the case of IPF, the team identified one repurposed drug and two promising new compounds.
Testing for compounds in IPF also began in mice, but has been expanded to samples of human lung tissue, thanks to a clinical research collaboration with Corewell Health’s lung transplant program in Grand Rapids.
This program is the busiest in Michigan. And because pulmonary fibrosis is a key indicator for lung transplantation, the program had enough explants to share with researchers to test as live cultures, said pulmonologist Reda Guirgis, MD, medical director of the transplant program and study contributor.
Guirgis, who is also a professor in the College of Human Medicine, said this study shows that progress is possible through collaboration between Corewell and MSU.
“I think the best way to advance medical knowledge is for clinicians to work in collaboration with biologists and now computational experts,” Gargis said. “That’s very important for advancing research.”
The team shared code and developed a web portal for researchers to use GPS for virtual compound screening.
“This is kind of a paradigm-shifting approach for people to drive discovery,” Chen says. “And we want more people to try this approach. But most importantly, we want people to actually use this approach to discover new treatments.”
Lee shared that ambition.
“I think we have already proven that this platform can be applied to two very different diseases,” he said. “This platform can therefore be used for other diseases as well, unlocking potential.”
This research was supported by the National Institutes of Health, the National Science Foundation, the Michigan State University Strategic Partnership Grant, the Corewell Health Michigan State University Alliance Corporation, the CJ Huang and Har Lin Yip Foundation to the Stanford University Asian Liver Center, and the Louis Hak Min Liver Cancer Research Foundation.
/Open to the public. This material from the original organization/author may be of a contemporary nature and has been edited for clarity, style, and length. Mirage.News does not take any institutional position or position, and all views, positions, and conclusions expressed herein are those of the authors alone. Read the full text here.
