A New Study Found the Power Switch Behind Getting High. Turning It Off Is Still Years Away.

In a lab run out of Beijing’s Nanhu Laboratory, researchers gave mice a lever for methamphetamine and, separately, a lever for sugar water, then watched one narrow gate inside their dopamine-releasing brain cells to see which lever made it open. Only the drug did. That gate is a channel called the mitochondrial calcium uniporter, or MCU — a valve that controls how much calcium, a signal that tells a cell to ramp up energy production, gets let into the mitochondria, the compartments inside a cell that generate its power. Opioids and methamphetamine forced that valve open inside dopamine neurons in the nucleus accumbens, the brain’s reward hub; food, water, and sex did not. When the researchers deleted the MCU gene specifically in those neurons, drug-driven dopamine surges and drug-seeking behavior both dropped — while the mice’s appetite for sugar water and normal reward stayed intact.

This is a real biological finding, not a treatment, and treating the two as the same thing cheats people who need the second one.

The paper, published this week in Nature Neuroscience and first posted as a preprint in June 2025, is a mechanism study. It didn’t test a drug in humans. It found something more specific and, in its own way, more interesting: a metabolic step that drugs need and natural rewards don’t. Senior author Xin Pan put it plainly to Medical Xpress: “This mechanism is not a general regulator of all dopamine release, but a selective pathway that specifically fuels the pathological processes underlying addiction.” If you’ve ever been prescribed naltrexone or buprenorphine and still felt the pull, this is worth sitting with: those medications work by blocking or dampening a receptor. They don’t touch the energy math happening one step downstream, inside the cell doing the work. This study is the first serious look at that math.

It landed in a field of addiction science that badly needs new math. As the American Journal of Psychiatry’s own September issue put it in a piece on where addiction treatment goes next, there are FDA-approved medications for alcohol, nicotine, and opioid use disorder — and none for cocaine or methamphetamine, the exact two drug classes hardest hit by the current stimulant wave. The medications that do exist for opioid use disorder don’t get everyone across the finish line either: extended-release naltrexone still leaves a large share of the people who start it back in active use within a year, which is part of why a 12-trial, 2,000-person meta-analysis in JAMA Psychiatry argued this year that trials should stop treating anything short of total abstinence as failure — in that data, 31.2% of participants achieved meaningfully reduced use against just 13.3% who hit full abstinence, and the reduced-use signal was consistently the stronger one. Read together, the two papers make the same point from opposite ends: the field’s current tools were built around one narrow definition of success, on one narrow biological target, and both are cracking under the weight of stimulant use disorder specifically.

Here’s the catch nobody covering this study is saying loudly enough. This was mice, not people, and Pan said as much himself: further research is needed “to determine whether a comparable mechanism contributes to human drug addiction.” There is no MCU-blocking pill waiting in a phase 1 trial. A prior high-throughput screen turned up mitoxantrone as a selective MCU inhibitor, and it’s too toxic to give anyone for addiction — it’s a chemotherapy drug with its own cardiac risk profile. The research team’s own stated next step is developing a safer inhibitor and testing it in animals again before anything resembling a human trial. That’s not a knock on the science. It’s the honest distance between “we found a mechanism” and “we found a medicine,” a distance addiction research has crossed badly before — ibogaine, baclofen, and a long list of “promising target” compounds have stalled exactly here, in the gap between a clean mouse result and a safe human dose.

The research team’s own stated next step is developing a safer inhibitor and testing it in animals again before anything resembling a human trial.

None of that makes the finding small. It reframes what a future medication could even look like: not a broader dopamine blocker, which risks flattening motivation and pleasure across the board, but something that only interrupts the metabolic overdrive specific to compulsive drug use. That’s a meaningfully different design goal than anything currently on the market, and it’s worth watching whether NIDA’s grant portfolio shifts toward mitochondrial bioenergetics in the next funding cycle, the way it shifted toward glutamate systems a decade ago.

What it doesn’t do is give anyone in recovery right now a new tool. If you’re on medication-assisted treatment today, this study changes nothing about your regimen, your risk, or your timeline — the people who told you MAT works were right before this paper and are still right after it. The honest version of this story is smaller and slower than the headlines about a “cellular switch” suggest: a real gate, found in real mice, that someone now has to spend years learning how to close safely in a person. That’s science working the way it’s supposed to. It’s just not, yet, help.