Insights into the effects of ketamine on the opioid system

Dating to seminal studies of Krystal and colleagues in the 1990’s (1), racemic ketamine and subsequently esketamine have had a dramatic impact on psychiatric therapeutics as rapidly acting antidepressants in the management of patients with treatment resistant major depression (TRMD). The success of these agents has spawned numerous preclinical and clinical studies examining their efficacy, side effects and mechanisms of action. From a therapeutic perspective, these agents are considered mainstays in the management of TRMD and esketamine has an indication for acute suicidality. Other potential indications continue to be studied (2).

Ketamine is a dissociative anesthetic and uncompetitive inhibitor of the N-methyl-D-aspartate class of glutamate receptors (NMDARs) and produces a form of open channel block in which the ion channel closes around ketamine trapping the molecule in the channel (3). Relief from block requires channel opening at depolarized membrane potentials; thus, the block can be prolonged depending upon regional network activity. Acute ketamine effects may occur preferentially on certain inhibitory interneurons via specific subtypes of NMDARs and result in disinhibition of excitatory (glutamatergic) neurons. In turn, activation of downstream messengers including brain-derived neurotrophic factor (BDNF) result in longer-lived functional and structural synaptic plasticity. The net result is persisting and somewhat paradoxical enhancement of glutamate-mediated excitatory transmission in multiple brain circuits. These persistent changes are thought to underlie longer-lived behavioral and mood changes (4).

While NMDARs are a major site of action, ketamine has multiple other effects as well as active metabolites that could contribute to the anesthetic, analgesic and/or psychotropic actions along with adverse psychotomimetic, dissociative, cognitive and abuse liability effects. These additional, non-NMDAR actions include, but are not limited to, inhibition of hyperpolarization-activated nucleotide-gated cation-1 (HCN-1) channels and modulation of the endogenous opioid system including direct actions that stimulate certain opioid receptors (ORs). The roles of the various effects of ketamine on NMDARs and other targets in producing its diverse clinical effects have not been completely deciphered but have prompted substantial research and clinical interest. In particular, the role of ORs remains an area of controversy with some evidence indicating that antidepressant effects of ketamine in humans are prevented by naltrexone, a broad spectrum OR antagonist, while other studies support the critical importance of NMDARs (5,6). Thus, understanding how various ketamine targets contribute to cellular, synaptic, network and behavioral effects of ketamine remains an important topic of investigation. In this issue of the journal, Pomrenze and colleagues (7) investigate the role ORs in ketamine’s acute behavioral effects in mice using elegant behavioral, cellular and molecular methods. The study clearly demonstrates the power of reverse translation from clinical observations back to preclinical experiments in uncovering previously unrecognized mechanisms.

For their studies, Pomrenze et al (7) initially screened male and female wild type mice for behavioral changes produced by racemic ketamine and used inhibition of ketamine-induced changes by naltrexone to implicate ORs. This screening approach examined an array of standard tests that span multiple ketamine-sensitive behaviors including the open field test for locomotion, the forced swim test (FST) and tail suspension test (TST) for antidepressant-like activity in both naïve (unstressed) and mice stressed with exogenous corticosterone, tests of pain sensitivity for analgesic actions, and changes in gait using catwalk analysis (Figure 1). Surprisingly and perhaps contrary to expectations, they identified hyperlocomotion as the only ketamine-induced behavior that was naltrexone sensitive. Interestingly, inhibition of hyperlocomotion by naltrexone was only observed in male mice, and the effects of MK-801, another trapping NMDAR open channel blocker that lacks ketamine’s effects on ORs, were insensitive to naltrexone. Similarly, cocaine-induced hyperlocomotion was unaffected by naltrexone.

Figure 1.

The diagram depicts experimental approaches used in the Pomrenze et al study including initial screening for naltrexone-sensitive, ketamine-induced behavioral changes (top images) followed by studies of naltrexone-sensitive cFos expression and identification of key neurons in CeA mediating hyperlocomotion (bottom images). This figure was made in BioRender (Mennerick, S. (2025) https://BioRender.com/5qilg1p).

To identify major brain regions contributing to OR-dependent effects of ketamine, the authors conducted an unbiased whole-brain screening assay using expression of the immediate early gene, cFos, to identify OR-sensitive ketamine effects in the absence and presence of naltrexone (Figure 1). In this assay, cFos expression served as an imaging-based proxy for neural activation. This screening assay identified clusters of activity of which the most prominent was in the central nucleus of the amygdala (CeA) and involved a specific subtype of neurons that expresses both μ-opioid receptors (MORs) and protein kinase C-delta (PKCδ). A subclass of CeA interneurons that expresses somatostatin (SST) and MORs was not involved in the ketamine effects. Because naltrexone non-selectively inhibits both MORs and κ-opioid receptors (KORs), the authors used selective pharmacological manipulations along with conditional MOR knockdown in CeA to demonstrate that MORs are key receptors mediating ketamine-induced locomotor changes.

The results of this study are significant in characterizing ketamine-induced behaviors in mice that are linked to activation of MORs in a specific class of neurons in the CeA. MOR- and PKCδ-expressing CeA neurons also express the endogenous opioid, enkephalin, and appear to participate in several potentially clinically relevant behaviors beyond hyperlocomotion including fear learning and anxiety-like changes. As the authors describe, these findings have potential translational relevance linking ketamine-induced MOR activation to a relatively simple behavioral assay. Less clear is whether these effects and underlying mechanisms relate specifically to psychotropic actions of ketamine in humans, including therapeutic and adverse effects. ORs did not appear to be involved in ketamine-induced changes in the FST assay for antidepressant-like effects in rodents (and ketamine was inactive in the TST). The authors argue that despite past predictive validity and some face validity of FST and TST, the more mundane locomotor activity assay may hold more translational relevance to human neuropsychiatric disorders. The present results should spur systematic evaluations of how these findings relate to ketamine-induced behaviors in humans and larger-scale studies of the possible role of MORs in the antidepressant effects of the drug in specific populations of patients.

Beyond the considerations above, there are multiple scientific questions stemming from these results. First, future studies should explore the differences between male and female mice found in this study and whether these observations translate to humans. This is particularly important if the results on hyperlocomotion relate to antidepressant effects in humans given the greater incidence of depression in women. Second, the findings about the key role of CeA MOR-PKCδ-expressing neurons prompt further studies of direct and indirect effects of ketamine and other antidepressants on these neurons, their connections, and other behavioral roles. Of immediate interest are the effects of S-ketamine (and the R-ketamine enantiomer) and metabolites of racemic ketamine, some of which have antidepressant-like actions. Third, these studies highlight unique mechanisms of ketamine in specific circuits. In other circuits other mechanisms of ketamine are being uncovered, including the role of prolonged inhibition of NMDARs and effects on neuronal bursting within the lateral habenula (3), and the role of specific messenger systems and downstream mechanisms in the persisting synaptic enhancement produced by ketamine in prefrontal cortex and hippocampus. Fourth, these results and the ensuing discussion highlight some of the limitations of translating studies in mice to humans. As noted by the authors, ketamine appears to have limited abuse liability in mice, based on tests of drug preferences. In contrast, ketamine is widely known for its abuse potential in humans, which is an important limiting variable in clinical use. Interestingly, the ability of abused drugs to promote hyperlocomotion in mice is associated with tendencies for drug self-administration; thus, the role of MORs in the hyperlocomotion effects of ketamine in male mice may be a predictor of drug misuse liability in humans. It will also be important to determine whether ORs contribute to behavioral and network effects of ketamine that persist over longer-periods of time, beyond the acute and sub-acute time frames of the present studies.

This study has important implications for the development of other antidepressants deriving from the success of ketamine and esketamine. These implications include opportunities for designing new molecular entities that increase or decrease various effects and mechanisms observed with ketamine. Additionally, opportunities also exist for repurposing other clinical drugs with mechanisms akin to but differing from ketamine. An example of the latter is nitrous oxide, an inhaled anesthetic that shares some but not all ketamine effects (8) and has shown beneficial effects in early phase clinical trials as a potential treatment for TRMD (9); nitrous oxide, like ketamine, also influences ORs. A critical need going forward is to find ways to prolong the clinical antidepressant effects of ketamine and other rapid antidepressants beyond the several weeks or less durations observed to date, perhaps building upon important mechanistic information such as described here and in other recent studies (10).