Letter in the July 12 online edition of Nature communication, Researchers at the University of California’s San Diego School of Medicine (UC) describe a new approach that uses machine learning to search for disease targets and then predict whether a drug is likely to gain approval from the U.S. Food and Drug Administration (FDA) receives.
The study results could measurably transform the way researchers sift through mountains of data to find meaningful information of significant benefit to patients, the pharmaceutical industry and the country’s health systems.
“Academic laboratories as well as pharmaceutical and biotech companies have access to unlimited amounts of ‘big data’ and better tools than ever to analyze such data. However, despite these incredible advances in technology, drug discovery success rates are lower today than they were in the 1970s, ”said Pradipta Ghosh, MD, senior study author and professor in the Departments of Medicine and Cellular and Molecular Medicine at UC San. Diego School of Medicine.
In the new study, Ghosh and colleagues replaced the first and last steps in preclinical drug discovery with two novel approaches developed at the UC San Diego Institute for Network Medicine. Researchers used the inflammatory bowel disease (IBD) disease model, a priority disease area for drug discovery and a difficult-to-treat disease because no two patients behave alike.
The first step, known as target identification, used an artificial intelligence (AI) method to model the disease using a map of the successive changes in gene expression as it began and as it progressed.
The final step, called target validation, was carried out in a first phase 0 clinical trial of its kind using a live organoid biobank created by IBD patients. The Phase 0 approach involves testing the effectiveness of the drugs identified using the AI model on human organoid models – cultured human cells in a 3D environment that mimic diseases outside the body. In this case, an IBD-infected intestine in a bowl.
“The concept of the phase 0 study was developed because most drugs fail somewhere between phases I and III. Before going to patients in the clinic, Phase 0 tests efficacy in the human disease models where ineffective compounds can be rejected early in the process, saving millions of dollars, ”said Soumita Das, PhD, co-senior author of the study and to associate professor in the Department of Pathology at the UC San Diego School of Medicine.
The drug identified by the AI model not only repaired the broken cell barriers that are characteristic of IBD in the organoid models, but also protected them from attack by pathogenic bacteria that were added to the intestinal model. “These results suggest that the drug could work for both acute flare-ups and maintenance therapy to prevent such flare-ups,” said Das.
The researchers found that the computational approach had a surprisingly high level of accuracy in different cohorts of IBD patients, and together with the Phase 0 approach, they developed a top-notch therapy for restoring and protecting the leaky gut barrier in IBD.
“Our study shows how the probability of success in Phase III clinical trials can be determined with mathematical precision for any target,” said Debashis Sahoo, PhD, co-senior author on the study and associate professor in the Pediatrics and Computer Science departments the UC San Diego School of Medicine and UC San Diego. “Our approach could provide the predictive power that will help us understand disease progression, assess the potential benefit of a drug, and develop a strategy for using a combination of therapies when the current treatment fails.”
The authors said the next steps include evaluating whether the drug that passed the Phase 0 human trial in a court can pass the Phase III trials in the clinic; and whether the same methods can be used for other diseases.
“Our design has the potential to shake the status quo and provide better drugs for chronic diseases for which there are still no good therapeutic solutions,” said Ghosh.
Edited by Gary Cramer