Dual action of ketamine confines addiction liability

This study used male and female mice (C57BL/6J and several transgenic lines) maintained on a controlled light–dark cycle. Behavioural assays included open-field locomotion, repeated injection paradigms for sensitization, and intravenous self-administration with implanted jugular catheters. Key drug doses reported were intraperitoneal (i.p.) ketamine 30 mg kg-1 and cocaine 15 mg kg-1 for acute tests; intravenous (i.v.) self-administration infusions were ketamine 1 mg kg-1 per infusion and cocaine 0.75 mg kg-1 per infusion (some cocaine experiments used 0.5 mg kg-1). Long-access self-administration sessions lasted up to 240 min in some cohorts; punishment sessions delivered 1-s air puffs after every three infusions. For circuit and activity measurements, the investigators used fibre photometry with genetically encoded sensors: dLight1.1 for dopamine in the NAc and GCaMP6m for Ca2+ activity in VTA dopamine and GABA neurons. Optogenetic tools included inhibitory eArch3.0 expression in VTA GABA neurons and ChR2-based stimulation of dopamine neurons. A Cre-dependent CRISPR–SaCas9 approach was used to delete the obligatory NMDAR subunit (GluN1/Grin1) selectively in VTA GABA neurons; validation included sequencing and loss of NMDA currents in infected cells. The team also applied pharmacology in vivo, including an irreversible D2 receptor antagonist (fluphenazine-N-mustard, FNM) and fentanyl for comparison. Ex vivo approaches comprised whole-cell patch-clamp recordings from VTA and NAc slices to measure AMPA/NMDA ratios, rectification indices (used to infer the presence of Ca2+-permeable AMPA receptors; inward rectification indicates such receptors), and NMDAR-mediated currents. The investigators tested induction of long-term potentiation (LTP) in slices and measured whether ketamine (50 µM in vitro, an estimated brain-relevant concentration) blocked NMDAR-mediated currents and LTP. Additional methods included immunohistochemical detection of FOS in D1-expressing medium spiny neurons (D1-MSNs) and hotplate analgesia testing. Experimental design elements reported include random assignment, blinding where possible, replication of experiments at least twice, normality testing and appropriate parametric or non-parametric statistics, and sample-size estimation.

Acute behaviour and dopamine signals: A single i.p. injection of ketamine (30 mg kg-1) increased locomotion in an open field to a degree similar to cocaine (15 mg kg-1). Using the dLight1.1 sensor, ketamine produced robust dopamine transients in the NAc that were similar in peak magnitude to those elicited by cocaine but were substantially shorter in duration; area-under-the-curve (AUC) of the transients increased with ketamine dose. VTA activity and circuit mechanism: Fibre photometry of VTA cells showed that ketamine increased activity in dopamine neurons for roughly 5 min while producing a strong and sustained inhibition of VTA GABA neurons. By contrast, cocaine decreased VTA dopamine neuron activity (consistent with D2 autoreceptor activation) and did not suppress VTA GABA neuron activity. Selective deletion of GluN1 from VTA GABA neurons abolished ketamine’s inhibitory effect on those GABA cells and reduced ketamine-evoked dopamine release in the NAc, supporting NMDAR antagonism on GABA neurons as the primary disinhibitory mechanism. Optogenetic and pharmacological tests: Optogenetic inhibition of VTA GABA neurons was itself reinforcing and evoked NAc dopamine transients. When optogenetic disinhibition was combined with ketamine, laser-induced dopamine transients were partially occluded by a non-saturating ketamine dose, consistent with overlapping mechanisms. The rapid termination of ketamine-evoked dopamine signals despite continued GABA inhibition prompted tests of D2 receptor involvement: pre-treatment with the irreversible D2 antagonist FNM prolonged ketamine-induced dopamine neuron activity, and similarly attenuated cocaine-induced auto-inhibition, indicating that D2 receptor-mediated feedback contributes to the fast off-kinetics of the dopamine signal; the extension was incomplete and variable between individuals. Synaptic plasticity assays: Ketamine did not induce the characteristic inward rectification at excitatory synapses onto VTA dopamine neurons that signals insertion of Ca2+-permeable AMPA receptors, whereas cocaine did. Optogenetic activation experiments established a temporal threshold: 15 min of in vivo dopamine neuron stimulation did not change the rectification index, whereas 60 min did, indicating that duration of dopamine neuron activation predicts induction of this early plasticity. To test whether prolonged exposure to ketamine could produce plasticity, the investigators administered repeated i.v. infusions every 2–4 min for 2 h to match the amplitude and duration of dopamine elevation produced by cocaine; this cocaine regimen increased the rectification index at VTA excitatory synapses but the matched ketamine regimen did not. NMDAR antagonism and blockade of plasticity: In vitro, ketamine at 50 µM strongly inhibited NMDAR-mediated synaptic currents and blocked induction of LTP in NAc slices. Thus, even when dopaminergic activation was prolonged experimentally, ketamine’s NMDAR antagonism prevented the NMDAR-dependent induction mechanisms required for synaptic potentiation. Downstream and behavioural plasticity: Unlike cocaine, a single ketamine injection did not increase FOS expression in NAc D1-MSNs. Repeated daily injections produced short-term locomotor sensitization for both drugs across the first four to five days, but after 7 or 30 days of withdrawal ketamine-treated mice returned to baseline locomotor responses whereas cocaine-treated mice showed persistent sensitization. Ex vivo measures after withdrawal showed no ketamine-evoked changes in rectification index or AMPA/NMDA ratio at NAc synapses, whereas cocaine produced expected changes at mPFC-to-NAc D1-MSN synapses. Self-administration and motivation: In short-access self-administration (2-h sessions) mice self-administered ketamine (1 mg kg-1 per infusion) at rates comparable to cocaine in initial tests. However, in extended-access protocols with escalating fixed ratios, mice reduced ketamine intake when effort was increased (decline at FR2) and overall infusions and active lever presses were lower for ketamine than for cocaine. Progressive-ratio break points were lower for ketamine than cocaine, indicating reduced motivation to obtain ketamine. Introducing mild aversive stimuli (air puffs) readily suppressed residual lever pressing for ketamine. Across measures, ketamine produced reinforcing initial behaviours but did not lead to uncontrolled self-administration.

Simmler and colleagues interpret their data as showing that sub-anaesthetic ketamine produces a distinctive, dual action in the mesolimbic system that limits addiction liability. Behaviourally, ketamine is reinforcing and elicits dopamine transients by disinhibiting VTA dopamine neurons through antagonism of NMDARs on VTA GABA neurons. Electrophysiological and photometry data indicate that these dopamine signals are rapid and short-lived because feedback through D2 receptors on dopamine neurons terminates the activity. Importantly, the investigators found no evidence that ketamine induces the early synaptic potentiation in the VTA and downstream NAc that is typically evoked by addictive drugs: ketamine did not produce Ca2+-permeable AMPAR insertion in VTA dopamine neurons, did not increase FOS in D1-MSNs, did not support persistent locomotor sensitization and did not produce uncontrolled self-administration. The authors propose two complementary mechanisms to explain the absence of drug-evoked synaptic plasticity. First, the fast off-kinetics of ketamine-evoked dopamine release produce only brief activation of dopamine neurons, below the temporal threshold required to induce the permissive plasticity. Second, ketamine’s NMDAR antagonism blocks the NMDAR-dependent induction pathways needed for potentiation even when dopamine elevation is prolonged experimentally. Together, these mechanisms—disinhibition-induced transient dopamine plus NMDAR blockade—constrain the cascade of plasticity that normally progresses from VTA to NAc and eventually to dorsal striatum and compulsive behaviours. The discussion places ketamine’s circuit action alongside other drugs that inhibit VTA interneurons (opioids, benzodiazepines, GHB, cannabinoids) but highlights ketamine’s uniqueness in reducing interneuron activity specifically via loss of NMDAR excitation. The authors acknowledge uncertainty about whether a direct action of ketamine at D2 receptors contributes to the rapid termination of dopamine signals, noting conflicting evidence in the literature. They also note that their findings align with reports of short-term but not long-term sensitization and that absence of drug-evoked mesolimbic plasticity and lack of uncontrolled self-administration argue for limited addiction liability. Finally, they suggest the dual-action mechanism is relevant to clinical use: typical sub-anaesthetic regimens for depression (40-min i.v. infusion of 0.5 mg kg-1 or nasal esketamine doses) produce pharmacokinetics that, together with NMDAR blockade, may similarly preclude induction of reward-related synaptic plasticity. The authors indicate that these insights could inform consensus and policy around therapeutic access to ketamine for depression.