AI-Designed Drug Cuts Fentanyl Intake and Reshapes Brain Circuits in Preclinical Study
The compound reduced opioid intake within hours in rats and was linked to changes in synaptic organization and gene expression in reward-related brain regions. In a rat model of opioid addiction, the compound reduced fentanyl intake and altered brain circuits linked to reward.
The Growing Crisis of Opioid Use Disorder
Opioid use disorder (OUD) is now widely understood as a chronic brain disease — one defined not simply by drug use, but by a maladaptive pattern of behavior that persists despite devastating consequences. In the US, nearly 82,000 people died from opioid overdoses in 2022 alone, a tenfold increase since the late 1990s. It is difficult to accurately estimate the total economic burden associated with substance use, encompassing factors from the cost of treatment as well as reduced productivity, loss of life, and the emotional toll on those left behind.
Stephen Loyd, Chief Medical Officer at Cedar Recovery, told DDN, “If you take every person in the US living with addiction and consider that each one affects seven others — family, friends, coworkers — you’re talking about roughly three-quarters of the country being touched, directly or indirectly.”
He went on to explain that failing to manage these conditions is costing the US healthcare system. “We spend three to four trillion dollars a year on healthcare in this country, and about 80 percent of that goes toward chronic disease,” Loyd said. “When you look at what’s actually driving those costs, substance use disorder is one of the biggest drains — not heart disease, not stroke, not hypertension.”
Significant Treatment Gaps and Stigma
Despite this growing crisis, there are still significant treatment gaps. Only an estimated 28 percent of people eligible for medications for OUD receive them, even though medication-assisted treatment is strongly associated with reduced opioid use, improved retention in care, and substantially lower mortality. However, access is limited by low clinic uptake and persistent stigma surrounding both addiction itself and the medications used to treat it.
“We have two or three effective medications for OUD, but two of them are controlled substances themselves. Buprenorphine and methadone both reduce mortality by as much as 50 to 70 percent, but access and uptake are limited,” Loyd said. “The bigger issue is stigma. People hear ‘you’re treating opioid use disorder with an opioid’ and think it’s just swapping one drug for another. But these medications save lives. What we still need are treatments that work without carrying that stigma, so more people are willing to seek care.”
Looking Directly at the Addicted Brain
This problem has driven researchers to look for entirely new approaches to OUD treatment. One such effort, now published in Proceedings of the National Academy of Sciences, combines human brain tissue analysis with AI-guided drug design to identify a compound that targets the underlying neurobiology of addiction rather than opioid signaling alone.
Instead of starting with animal models or screening existing drug libraries, the research team focussed on human biology. Researchers at the University of California, Irvine, partnered with GATC Health to examine postmortem brain tissue from individuals with OUD, using those samples to identify molecular signatures associated with addiction. Those datasets did not immediately yield a single therapeutic target. Instead, they served as a starting point for interpretation.
Through GATC’s Operon AI platform, the team analyzed patterns across the data to infer which combinations of biological pathways might be driving the persistent dysregulation seen in addiction. “Our initial goal was simply to validate whether the biomarker signals they had identified were meaningful,” said Jayson Uffens, Chief Technology Officer at GATC Health. “But when we applied our system, we realized we could go further — not just confirming biomarkers, but identifying potential therapeutic targets, which in this case pointed toward serotonin receptor pathways, and ultimately toward drug design itself.”
From the tissue-derived signals, the system generated and ranked multiple candidate target combinations. Among the most consistent were serotonergic pathways — particularly serotonin 2A and 6 receptors — suggesting that modulation of these systems, rather than direct opioid receptor targeting, might influence craving and decision-making circuits implicated in relapse.
