In the midst of a global opioid epidemic that claims over 350,000 lives each year, scientists have identified a previously underappreciated brain gene that may be driving heroin addiction. Using machine learning to probe the brains of people who had used heroin, a new study has found a unique molecular fingerprint associated with opioid use—and pinpointed a gene called Shisa7 as a key predictor of heroin-seeking behavior. The discovery, published in Biological Psychiatry, offers valuable insight into the neurobiology of addiction and potential new avenues for treatment.
Mining the Addicted Brain with Machine Learning
The research, led by Dr. Yasmin L. Hurd of the Icahn School of Medicine at Mount Sinai in New York, applied advanced machine-learning algorithms to postmortem human brain tissue. The team examined the orbitofrontal cortex (OFC) — a region of the frontal lobe just above the eyes that is critical for impulse control, decision-making, and drug-seeking behavior. By sequencing RNA from OFC samples of deceased heroin users and non-user control subjects, the scientists obtained a transcriptome (a readout of all active genes) for each individual. They then trained a machine learning model to distinguish the molecular signatures of heroin users’ brains from those of controls.The algorithm’s analysis converged on one gene in particular—Shisa7—as the single strongest marker of a heroin user’s brain. In other words, Shisa7’s expression in the OFC was a reliable indicator of whether a person had been a habitual heroin user. This gene had not been previously linked to addiction. Shisa7 encodes a protein that is known to interact with neurotransmitter receptors in the brain; notably, it serves as an auxiliary subunit influencing GABA and glutamate receptors, which govern the brain’s main inhibitory and excitatory signals. By helping control these chemical messaging systems, Shisa7 can modulate brain activity balance. The prominence of Shisa7 in the machine learning model was unexpected, revealing a new molecular player in the biology of opioid addiction.
From Prediction to Causation: Testing Shisa7 in Rats
Finding Shisa7 in human brains was only the first step—the researchers next wanted to know if this gene actually drives addictive behavior. To investigate causality, they turned to a preclinical rat model of heroin use. The team artificially boosted (overexpressed) the Shisa7 gene in the OFC of otherwise drug-naïve male rats—animals that had never been exposed to heroin.The result was striking: rats with elevated Shisa7 levels began behaving like rats addicted to opioids. They showed intensified drug-seeking behavior, eagerly pressing levers to obtain heroin in self-administration tests, and they struggled with cognitive flexibility (an ability to adapt to changing rules to obtain rewards). In essence, turning up Shisa7 in the rat brain made the animals more compulsive and inflexible, hallmarks of addiction. This provided strong evidence that Shisa7 is not merely a passive marker of heroin use, but an active driver of addictive behavior.
A Molecular Mirror of Heroin’s Effects
Beyond behavioral changes, overexpressing Shisa7 in the rats’ brains produced an eerie molecular parallel to human heroin use. The researchers found that ramping up Shisa7 triggered sweeping changes in gene expression in the rat OFC—changes that closely mirrored those seen in the brains of human heroin users. In other words, simply increasing this one gene’s activity caused the rats’ brain cells to behave, at a molecular level, as if the animals had been chronically taking heroin.The Shisa7-driven gene signature was strongly enriched for mechanisms seen in disorders like Parkinson’s, Alzheimer’s, and Huntington’s disease, highlighting an unexpected link between chronic opioid exposure and brain degenerative processes.
The convergence of evidence suggests that Shisa7 acts as a molecular switch that can reproduce the wide-ranging neural effects of heroin. In exploring how Shisa7 might exert such broad influence, the team also identified proteins that bind to Shisa7 in the brain. Many of these binding partners turned out to be components of GABA_A or glutamate receptors—the very neurotransmitter systems that Shisa7 helps regulate and that are known to be disrupted by long-term drug use. This finding aligns with the idea that altering Shisa7 levels can skew the brain’s excitatory/inhibitory balance, potentially leading to neural circuit dysfunction over time. It also underscores a connection between opioid addiction and neuroimmune and neurodegenerative pathways: the gene networks influenced by Shisa7 included those involved in inflammation and the health of brain cells.
Toward New Therapies for Opioid Addiction
By identifying Shisa7 as a central molecular player in heroin addiction, this research opens new doors for therapeutic development. The hope is that scientists can leverage these insights to design interventions that break the cycle of addiction at the level of the brain’s biology. Targeting Shisa7 or its downstream effects could become a strategy to reduce drug-seeking behavior or perhaps protect the brain from opioid-induced damage. For example, future medications might aim to dial down Shisa7 activity or modulate the GABA and glutamate receptor pathways it influences, thereby restoring healthier brain function in people with opioid use disorder. The innovative blend of human data, artificial intelligence, and animal experiments in this study exemplifies a translational approach to addiction research—one that starts with real-world human biology and then tests causality in the lab to find actionable targets.For further details, readers can refer to the original study by Ellis et al. in Biological Psychiatry: 10.1016/j.biopsych.2024.12.007
If you're interested in such publications, please share, react and leave comments.