A team of researchers from Cleveland Clinic's Genome Center and IBM are using artificial intelligence (AI) for drug discovery in advanced pain management.

According to the researchers, their deep-learning framework identified several gut microbiome-derived metabolites along with U.S. Food and Drug Administration (FDA)-approved drugs that can be repurposed to select non-addictive, non-opioid options for treating chronic pain.

Source: DeepMind from Pexels

While treating chronic pain with opioids is still challenging thanks to the risk of severe side effects as well as dependency, the research team noted that evidence has demonstrated that drugging a specific subset of pain receptors in a protein class called G protein-coupled receptors (GPCRs) can offer non-addictive, non-opioid pain relief.

To determine a path for targeting those receptors, the researchers wondered if they could apply research methods they have already created for finding pre-existing FDA-approved drugs for potential pain indication rather than inventing new molecules from scratch. Part of accomplishing this, the team noted, involves mapping out gut metabolites to identify drug targets.

To identify these molecules, the research team updated an earlier drug discovery AI algorithm while collaborators from IBM helped write and edit the manuscript.

The team explained that to ascertain if a molecule will work as a drug, they need to forecast how it will physically interact with and subsequently influence proteins in the human body — specifically, the pain receptors. To make such predictions, the team needed a 3D understanding of both molecules built upon 2D data about their physical, structural and chemical properties.

"Even with the help of current computational methods, combining the amount of data we need for our predictive analyses is extremely complex and time-consuming," the researchers explained. "AI can rapidly make full use of both compound and protein data gained from imaging, evolutionary and chemical experiments to predict which compound has the best chance of influencing our pain receptors in the right way."

The new tool dubbed LISA-CPI (Ligand Image- and receptor's three-dimensional (3D) Structures-Aware framework to predict Compound-Protein Interactions), which relies on deep learning, reportedly enabled the researchers to make predictions including: whether or not a molecule binds to a specific pain receptor; where on that receptor the molecule will physically bind; how strongly the molecule will bind to that receptor; and if binding a molecule to a receptor will turn signaling effects on or off.

LISA-CPI made predictions about how roughly 369 gut microbial metabolites and 2,308 FDA-approved drugs would interact with 13 pain-associated receptors. The AI framework identified multiple compounds that could be potentially repurposed to treat pain, thereby minimizing the experimental burden researchers must overcome to identify a list of candidate drugs for further testing.

Additionally, the team also suggests that LISA-CPI could be used to test even more drugs, metabolites, GPCRs and other receptors to identify therapeutics for treating diseases beyond pain — such as Alzheimer's disease, for example.

An article detailing the team’s findings, “A deep learning framework combining molecular image and protein structural representations identifies candidate drugs for pain,” appears in the journal Cell Reports Methods.