Ketamine’s opioid receptor connection has been argued from the outside for years: block opioid receptors with naltrexone in depressed patients, and ketamine’s antidepressant effect weakens. That behavioral evidence, first published in 2018 and replicated since, has never directly shown what ketamine is doing at the receptor itself. A team at Washington University School of Medicine, led by Tao Che, has now closed that gap. Their paper, published June 22, 2026 in Nature Structural & Molecular Biology, reports six cryo-electron microscopy structures showing ketamine and its parent compound phencyclidine physically occupying the orthosteric binding site, the same pocket the receptor’s natural activating molecules use, on both the kappa and mu opioid receptors.
What the structures show
This is direct, atomic-level evidence, not inference from blocking one drug’s effect with another. The team solved structures of kappa and mu opioid receptors bound to PCP, to S-ketamine, and to a related PCP metabolite, plus a fifth structure capturing the kappa receptor’s shape with nothing bound at all, a rare and technically difficult baseline that lets the researchers see exactly how the receptor’s pocket rearranges specifically to accommodate these two drugs. In parallel behavioral work, a subanesthetic dose of ketamine produced measurable pain relief in mice that was completely blocked by naloxone, the standard opioid antagonist, and separately by a kappa-receptor-selective antagonist, confirming the structural finding translates into a real, blockable physiological effect rather than existing only in isolated receptor preparations.
Why this revises rather than simply confirms prior work
A widely cited 2021 study had characterized ketamine as a full agonist at the kappa opioid receptor and a partial agonist at the mu receptor. This new structural data tells a more precise story: across every enantiomer of ketamine tested, and PCP itself, the compounds functioned as partial agonists at all three opioid receptor subtypes, not full agonists at any of them. That is a real, specific revision to the field’s understanding, not a confirmation with better resolution, and it matters because a full agonist and a partial agonist carry different implications for a compound’s therapeutic ceiling and its abuse liability profile.
Why the binding itself is unusually weak, and why that matters
Ketamine’s affinity for opioid receptors is roughly ten to a hundred times weaker than its affinity for the NMDA receptor, its primary known target. The researchers found this weak affinity stems specifically from an unusual binding mode: ketamine and PCP rely almost entirely on loose, hydrophobic contacts within the receptor pocket, without forming the tight salt-bridge interaction that most opioid drugs, including morphine and fentanyl, use to lock into place. That structural looseness likely explains why ketamine’s opioid receptor engagement has been so difficult to pin down pharmacologically for decades, and it directly supports a specific, testable clinical hypothesis: this pathway may only become relevant at higher drug concentrations, such as during anesthesia or overdose, rather than at the lower doses used for depression treatment, where NMDA receptor engagement would dominate.
Where this connects to the rest of this desk’s coverage
This is the same mechanistic question this desk examined directly in a recent piece on tianeptine, another atypical antidepressant whose behavioral effects were shown to depend specifically on opioid receptor activation. That piece used ketamine, tested in the same paradigm, as the clean contrast case, a naloxone-insensitive, non-opioid signature. This new structural paper adds real nuance to that contrast: ketamine’s opioid engagement is real and now structurally confirmed, but it appears to require a different exposure threshold than tianeptine’s, and the two compounds’ relationship to the opioid system may not be as simply opposed as that earlier framing suggested. Both pieces of evidence can be true at once: ketamine has confirmed, structurally direct opioid receptor engagement, and that engagement may still play a smaller role in its standard antidepressant dosing than it does at the higher concentrations reached during anesthesia.
The caveats
This is structural and rodent behavioral data, not a claim about what happens in human patients receiving standard antidepressant-dose ketamine. The paper’s own authors are explicit that opioid receptor engagement is more likely to be physiologically relevant at higher-exposure conditions, anesthesia, overdose, or peak concentrations immediately after rapid infusion, than at the lower concentrations used in psychiatric treatment, and they call directly for further work to determine which receptor subtypes and circuits matter in vivo. The clinical literature on whether naltrexone blocks ketamine’s antidepressant effect in humans remains substantially mixed, with some case reports showing preserved antidepressant response despite opioid blockade, an unresolved tension the structural evidence alone does not settle.
The frame
For prescribers, payers, and the regulatory bodies that oversee ketamine’s use, a structural confirmation that this receptor engagement is real, rather than a statistical artifact of blocking one drug with another, changes the weight the opioid-mechanism question deserves in how ketamine gets discussed, prescribed, and eventually scheduled or labeled. It does not resolve the practical question that matters most for patient care: at what dose, in what clinical context, and for which patients this mechanism becomes something worth managing rather than a laboratory curiosity. That question is now better posed than it was before this paper, structurally grounded rather than argued entirely from indirect behavioral evidence, but it remains open.