0
0
February 5, 2024
Scientists at the University of Virginia (UVA) have unveiled a groundbreaking approach to drug discovery using machine learning to identify drugs that could minimize harmful scarring following heart attacks and other injuries. This new technique has already identified a promising candidate to prevent harmful heart scarring, marking a significant departure from conventional drug discovery methods.
The UVA researchers, led by computational biologist Dr. Anders R. Nelson and Dr. Jeffrey J. Saucerman of UVA’s Department of Biomedical Engineering, developed a cutting-edge computer model that combines decades of human knowledge with machine learning capabilities. This innovative approach has the potential not only to predict the effects of drugs on fibroblasts but also to explain how they work, providing crucial insights for designing clinical trials and identifying potential side effects.
“Many common diseases such as heart disease, metabolic disease, and cancer are complex and hard to treat,” said Dr. Nelson. “Machine learning helps us reduce this complexity, identify the most important factors that contribute to disease, and better understand how drugs can modify diseased cells.”
The researchers focused on fibroblasts, cells responsible for repairing the heart after injury by producing collagen and contracting the wound. However, these cells can also contribute to harmful scarring known as fibrosis during the repair process. Previous attempts to identify drugs targeting fibroblasts have faced challenges in understanding how these drugs work and have often focused on selected aspects of fibroblast behavior.
To address these challenges, Saucerman and his team employed a novel approach called “logic-based mechanistic machine learning.” This approach not only predicts the effects of drugs but also explains how they influence fibroblast behaviors. The researchers tested the effects of 13 promising drugs on human fibroblasts, training the machine learning model to predict their impacts on cells and behavior.
The model successfully provided a new explanation of how the drug pirfenidone, already FDA-approved for idiopathic pulmonary fibrosis, suppresses contractile fibers within fibroblasts, ultimately preventing heart stiffening. Another experimental drug, the Src inhibitor WH4023, showed promise in targeting a different type of contractile fiber.
While further research is needed to validate these findings in animal models and human patients, the UVA researchers are optimistic about the potential of mechanistic machine learning in drug discovery. The team emphasizes the technology’s versatility, foreseeing its application in the development of treatments for various diseases beyond heart injuries.
“We’re looking forward to testing whether pirfenidone and WH4023 also suppress the fibroblast contraction of scars in preclinical animal models,” said Dr. Saucerman. “We hope this provides an example of how machine learning and human learning can work together to not only discover but also understand how new drugs work.”
The research received support from the National Institutes of Health through grants HL137755, HL007284, HL160665, HL162925, and 1S10OD021723-01A1. The findings open new avenues for drug discovery, offering hope for more effective treatments across a range of challenging medical conditions.
