Introduction

Esketamine, or (S)-ketamine, is one of the enantiomers of the racemic drug (R,S)-ketamine. Both esketamine and racemic ketamine constitute a pharmacologically novel class of medications for the treatment of mood disorders, especially those related to depression such as treatment resistant major depressive disorder and acute suicidal ideation and behavior [1, 2]. These medications act on different pharmacological targets compared to traditional monoaminergic-based drugs which act more slowly and, in some patients, fail to achieve full recovery [3]. Ketamine and esketamine target primarily the N-methyl-D-aspartate receptors (NMDAR), with esketamine being more potent than racemic ketamine as well as the other enantiomer arketamine, (R)-ketamine. However, the affinity of ketamine’s enantiomers for the phenylcyclidine (PCP) binding site of the NMDAR is relatively low (0.2–3 µM [4, 5]) compared to other canonical ligands such as PCP or dizocilpine that are on the low nanomolar range [5]. Moreover, preliminary clinical studies indicate that arketamine (the lower NMDAR affinity enantiomer) may have better antidepressant properties than esketamine, or at least have better safety and tolerability [6, 7]. Taken together, this challenges the prevailing concept that the NMDAR is the solely responsible for the antidepressant effects of ketamine and suggests that other pharmacological targets might explain some aspects of its therapeutic properties. In fact, many preclinical studies have suggested a wide array of molecular targets including sigma and opioid receptors [4, 8,9,10,11], or the brain derived neurotrophic factor (BDNF) receptor TRKB [12], among others (for review see [13]).

A key feature of ketamine is its ability to modulate dopaminergic transmission [14, 15]. Some studies have suggested direct interactions with monoamine transporters [16, 17] or pre-synaptic receptors [18, 19]. In a microdialysis study in mice, esketamine (20 mg/kg) produced a long and sustained increase in dopaminergic tone compared to arketamine [20] and ketamine restored the deficits in dopaminergic activity observed in learned helplessness Wistar-Kyoto rats [21]. Other studies have linked ketamine’s effects on dopamine to the NMDAR antagonism in select cell populations, especially in the medial prefrontal cortex (mPFC) [22]. In agreement with previous PET imaging studies in humans [23, 24], our PET imaging studies in rats using the dopamine D2 receptor antagonist [18F]fallypride and the glucose analog [18F]fluorodeoxyglucose ([18F]-FDG) reported a selective action of esketamine (compared to arketamine) to modulate both dopaminergic tone and metabolic activity in corticolimbic circuits (i.e. mPFC to nucleus accumbens (NAc)) [4], together with a reduction in binding potential of the opioid antagonist [18F]fluoroethyldiprenorpine ([18F]-FE-DPN) in mPFC after an acute esketamine administration [8]. The decrease in [18F]-FE-DPN binding potential could be explained by direct competition of esketamine to bind opioid receptors [4], but other studies also suggested that ketamine promotes β-endorphin release in the mPFC, which would also lead to a decrease in radiotracer uptake [25]. Finally, the involvement of opioid receptors was further suggested by functional ultrasound imaging (fUS) where the acute increase in blood flow in the NAc was partially blocked by naltrexone [8, 26]. However, all these prior studies studying neurotransmitter release relied on PET imaging or in vivo microdialysis, both of which have low temporal resolution and hence can only detect slow or persistent changes in dopamine tone. Therefore, in this study, we explore the mechanistic aspects of the dopaminergic effects of esketamine using behavioral assays and genetically encoded biosensors targeted to discrete cell types and circuits using fiber photometry in mice.

Results

Biphasic effects of esketamine administration on locomotor activation and sensitization

Increases in psychomotor activity and development of sensitization are behavioral phenomena caused by drugs of abuse and are mediated, at least in part, by alterations in dopaminergic neurotransmission [27, 28]. In naïve mice (N = 64, 30 male, 34 female), a single esketamine administration (10 mg/kg, i.p.) [4] produced a rapid and short-lasting increase in locomotion that was later followed by a decrease in activity below the levels of saline-treated mice (Fig. 1A–C, E). The statistical analysis of the distance traveled using a three-way ANOVA with repeated measures by subject with time bin, esketamine treatment and naloxone treatment as main factors showed significant effects of time bin (F(9.871, 641.6) = 34.81, p < 0.0001) and a triple interaction between time bin x esketamine x naloxone (F(29, 1885) = 2.241, p = 0.0002). At the same time, this effect was accompanied by a significant decrease in time spent in the center of the arena (Fig. 1A, B, F), a response indicative of increased anxiety-like behavior [29]. Three-way ANOVA with session, esketamine and naloxone as factors showed significant effects of esketamine (F(1, 60) = 4.716, p = 0.033), but no interaction between esketamine and naloxone. In summary, while the initial locomotor stimulating effect was partially blocked by the mu-opioid receptor (MOR) antagonist naloxone (3 mg/kg, s.c.), both the late locomotor depression and the center time were not (Fig. 1C, F).

Fig. 1: Some, but not all, psychomotor effects of esketamine require opioid receptor engagement.
figure 1

Representative trajectory of a representative mouse placed in an arena for 30 min after saline A or acute esketamine administration B, where the color of the line depicts the speed of the animal. C and D Time course of the elapsed distance before and after saline (grey), naloxone (blue), esketamine (orange) or naloxone and esketamine (green) administration, line and shade represent mean ± SEM of all animals. E Accumulated distance during the 10 min following drug (or saline) administration (each dot represents one animal, bars represent mean, summary statistics: post-hoc Tukey’s multiple comparisons test: *p < 0.05, ** p < 0.01). F Percentage of time the animals spent in the center of the arena (each dot represents one animal, summary statistics: post-hoc Tukey’s multiple comparisons test: *p < 0.05, ** p < 0.01) G Representative heatmap of the preferred locations of mouse placed in the Elevated Plus Maze for 5 min after saline or acute esketamine administration. H, I Time spent in the open arms, and total distance traveled by the animals (each dot represents one animal, statistical significance calculated by t-tests: ***p < 0.001 ****p < 0.0001.

Full size image

We continued to inject the mice with esketamine (10 mg/kg, i.p., daily) for four more consecutive days and on the fifth day we re-evaluated the acute effects of the drug. In agreement with our previous results [4], we found that esketamine increased locomotor activation in male mice on Day 5 compared to Day 1, a response indicative of sensitization (Fig. 1E, S1) (main effect of session: F(1, 26) = 8.510, p = 0.0072), however female mice did not show increased locomotion after repeated esketamine administration (Fig. S1) (session: F(1, 28) = 0.8347, p = 0.37). Interestingly, on male mice while the statistical analysis of the time course (Fig. 1D, S1) showed a triple interaction between time bin, esketamine and naloxone on Day 1 (F(29, 1885) = 2.241, p = 0.0002), there was no interaction between these factors on Day 5 (F(29, 1766) = 0.5925, p = 0.95). On the other hand, and in contrast to the first esketamine administration, esketamine did not change the time spent on center after 5 days of repeated administration (Fig. 1F) and the statistical analysis showed an interaction between esketamine and experimental session (F(1, 55) = 22.32, p < 0.0001). Overall, these data suggest that single vs. repeated esketamine administration produces multifaceted effects on both locomotor activity and anxiety-like behavior, with the locomotor stimulating effects being dependent on opioid receptors.

To further confirm the putative anxiogenic-like effects of acute esketamine administration we performed the elevated plus maze (EPM) test [30] in naïve mice (N = 78, 42 male, 36 female). Mice received saline or esketamine (10 mg/kg, i.p.) and 15 min later they were placed in the center of an EPM and their trajectory was evaluated for 5 min (Fig. 1G). Male mice injected with esketamine spent significantly less time exploring the open arms of the maze (Fig. 1H) and this effect was not affected by pretreatment with naloxone. However, female mice -which on saline conditions spent less time exploring the open arms than male mice- were unaffected by esketamine administration. Three-way ANOVA with sex, esketamine, and naloxone administration as factors revealed significant main effects of sex (F(1, 69) = 4.260, p = 0.0428) and esketamine (F(1, 69) = 6.240, p = 0.0149) and an interaction between sex and esketamine (F(1, 69) = 18.50, p < 0.0001). However, consistently with the open field data, although both sexes showed similar overall activity upon vehicle administration, only male mice were less active when administered with esketamine and this effect was blocked by naloxone (Fig. 1I). Three-way ANOVA: main effect of esketamine (F(1, 72) = 12.97, p = 0.0006), interaction naloxone x esketamine (F(1, 72) = 2.947, p = 0.0904), sex x esketamine (F(1, 72) = 5.691, p = 0.0197). To try to correct for this putative biasing factor, we calculated the relative distance traveled in the open arms (Fig. S1). These data showed no significant effects on esketamine-treated mice (Fig. S1, F(1, 73) = 2.645, p = 0.1082) and hence indicated that the avoidance of the open arms was not due to an impairment of their locomotor performance or sedation, but active avoidance of the open arms, confirming that a subanesthetic dose of esketamine increases anxiety-like behavior in male mice.

Esketamine alters phasic dopamine frequency and re-uptake kinetics in mice

The behavioral data presented above indicated that esketamine produces behavioral effects that are typically modulated by mesolimbic dopaminergic neurotransmission. Therefore, to explore the potential involvement of dopamine in the pharmacological effects of esketamine, we used the genetically encoded dopamine biosensor dLight1.3b [31] to monitor extracellular dopamine concentrations in the NAc of mice (Fig. 2A). Three to four weeks after intracranial injection of a viral vector AAV-1/2-hSyn1-Dlight1.3b in the NAc of mice (N = 35, 18 male, 17 female), we inserted a fiber optic into the injection site and fixed it to the skull (Fig. 2A, B). After 4–5 days of recovery, we recorded the dopamine-elicited changes in fluorescence while the animals were in their home cage after an acute administration of saline, naloxone (3 mg/kg, s.c.), esketamine (10 or 30 mg/kg, i.p.) or naloxone plus esketamine (Fig. 2C–G, Fig. S2). Esketamine did not produce any observable change in dopamine tone (Fig. 2C, Fig S2), however the 30 mg/kg dose of esketamine decreased the frequency of dopamine transients (Fig. 2E; mixed-effects model with repeated measures (REML) with esketamine, naloxone and session as factors: main effect of esketamine (F(2, 36) = 11.09, p = 0.0002) and it was unaffected by naloxone (REML interaction esketamine x naloxone: F(2, 36) = 0.07,147, p = 0.93)). Esketamine administration also produced a dose-dependent change in the kinetics of the spontaneous dopamine transients compared to vehicle-treated mice and this effect was blunted upon naloxone pre-administration. Specifically, while the effects on dopamine peak amplitude were not altered (Fig. 2F, Fig. S2), the duration of the dopamine transients increased significantly from 0.96 ± 0.09 s in baseline conditions to 1.17 ± 0.11 and 1.30 ± 0.13 after 10 and 30 mg/kg esketamine, respectively (Fig. 2G, Fig. S2; REML esketamine: F(2, 34) = 25.41, p < 0.0001; esketamine x naloxone: F(2, 34) = 5.035, p = 0.0122). Treating mice with vehicle or naloxone at the same time intervals as esketamine did not alter the photometry signal during the course of the experiment, indicating that the observed effects were due to esketamine administration and not time spent in the testing arena (Fig. S2). These results indicate that the so-called increase in dopamine neurotransmission caused by esketamine is not due to an enhancement of dopamine release but to longer duration of dopamine presence in the postsynaptic area.

Fig. 2: Esketamine administration causes mixed effects on spontaneous dopaminergic transmission.
figure 2

A Schematic representation of dopamine sensor expression in the NAc. B Representative immunohistochemical image of dLight1.3b expression in the NAc. Scale bar: 1 mm. C Average trace detected from the fluorescent signal of dLight1.3b in the NAc after saline or esketamine (10 mg/kg and 30 mg/kg) administration. D Representative traces of the corrected dLight1.3b in the NAc after saline or esketamine (10 mg/kg and 30 mg/kg) administration. E–J Peaks per minute, dopamine transient amplitude and dopamine transient duration on day 1 and day 5 of esketamine administration with and without pretreatment with naloxone (3 mg/kg) (each dot represents one animal). K Schematic representation of the viral strategy for expression dopamine and calcium biosensors in the NAc and VTA-originated terminals, respectively, and representative traces of Dlight1.3b, RCaMP1.07 and isosbestic control signals. L–N Peaks per minute, transient amplitude, and transient duration of calcium in the NAc (each dot represents one animal). In all panels, statistical significance, evaluated by Tukey’s post-hoc multiple comparisons test, * p < 0.05 and ** p < 0.01). O, P Correlation between extracellular dopamine and intracellular calcium from terminals from VTA neurons in the NAc. All correlations were statistically significant with p < 0.001 and esketamine treatment did not alter Pearson’s r coefficients on amplitude (p = 0.52) nor duration (p = 0.58).

Full size image

We then repeated the same experiment after four days of daily esketamine injections (Fig. 2H–J). Before esketamine administration, dopamine transient frequency and their amplitude and duration did not differ from naïve mice (Fig. 2H–J). However, esketamine administration after four days had weaker effects on frequency and transient duration, but increased the dopamine transient amplitude, an effect that was not observed on day 1 (Fig. 2H–J). Statistical REML analysis showed a significant interaction between esketamine and session on transient duration F(2, 15) = 3.921, p = 0.042) and no significant effects of interaction between naloxone and session were observed in any of the three parameters.

We aimed to evaluate whether the esketamine-induced extracellular dopamine increases were due to pre- or postsynaptic actions of esketamine. In a subset of mice (N = 5 male, 1 female), we injected a viral vector encoding the red fluorescent calcium biosensor RCaMP1.07 (AAV-1/2-hSyn1-RCaMP1.07) into the VTA together with the dLight1.3b sensor in the NAc. In these mice, NAc dLight1.3b (from postsynaptic cell bodies and dendrites) and RCaMP1.07 fluorescence (from dopaminergic presynaptic terminals) were recorded simultaneously to establish a correlation between the two signals (Fig. 2K). We observed a high coherence between calcium and dopamine signals (Fig. 2K, O, P) and, in agreement with the dLight1.3b data, esketamine increased the duration of the spontaneous calcium transients in NAc dopaminergic terminals (mixed-effects model with repeated measures: F(1.733, 6.066) = 13.68, p = 0.0063) without affecting their amplitude (Fig. 2L–N).

Esketamine increase in electrically evoked dopamine is dependent on MOR engagement

To explore the pharmacological mechanisms underlying esketamine’s effects on dopamine neurotransmission we used local electrical stimulation to induce neurotransmitter release in anesthetized mice [32] (N = 32, 20 male, 12 female). This protocol allowed repeated and controlled release events and hence it allowed us to test drug actions over time after establishing a baseline for each mouse (Fig. 3A, B). In concordance with the awake photometry data (Fig. 2), esketamine increased electrically evoked dopamine peak duration (Fig. 3C, D; two-way ANOVA with esketamine dose and presence or absence of naloxone (see below), main effect of esketamine: F(2, 36) = 68.06, p < 0.0001) without altering peak amplitude (Fig. 3E; two-way ANOVA (esketamine’s main effect): F(2, 37) = 0.07,698, p = 0.9261)), and resulting in an overall increase in extracellular dopamine concentration as quantified as the area under the curve (AUC) of each peak (Fig. 3F; two-way ANOVA (esketamine’s main effect): F(2, 32) = 81.23, p < 0.0001).

Fig. 3: Multiple molecular mechanisms account for the dopaminergic effects of esketamine.
figure 3

A Schematic representation of AAV injection site and fiber/electrode placement in the NAc. B Representative heat map of extracellular dopamine concentrations before and after each electrical stimulation event. C Representative dopamine traces before (black) or after esketamine administration (10 mg/kg, orange, 30 mg/kg purple). D–F Duration, amplitude, and AUC of extracellular dopamine transients. G Schematic representation of AAV-hSyn-GCaMP7s injection site and fiber/electrode placement. H Representative calcium (GCaMP7s) traces after esketamine administration (10 and 30 mg/kg). I–K Duration, amplitude, and AUC of the electrically-evoked intracellular calcium transients. L Representative traces of extracellular dopamine (RdLight1) after optogenetic stimulation of the ChR2-expressing terminals (red) or non-ChR2 expressing terminals (black) in the NAc. Cyan line represents the stimulation period (400 ms). M Schematic representation of AAV-RdLight1 injection in the NAc, Channelrhodopsin-2 (ChR2) injection in the VTA and fiber in the NAc that was used for both recording and stimulation. N Representative RdLight1 trace after several optogenetic (cyan) and electrical (black) stimulation events. O Averaged traces (3–6 per animal) of electrical or optogenetic stimulation events before or after esketamine (30 mg/kg) administration. In all panels, each dot represents the average of 3 to 5 trials per animal and the bars represent the mean of all dots; statistical significance, evaluated by Tukey’s post-hoc multiple comparisons test, * p < 0.05, ** p < 0.01, **** p < 0.0001).

Full size image

We used the mu-opioid receptor (MOR) antagonist naloxone (3 mg/kg, s.c.) to evaluate whether the dopaminergic effects of esketamine were dependent on MOR. Naloxone partially blocked esketamine-induced increases in peak duration and AUC (Fig. 3D–F). Two-way ANOVAs with esketamine dose and presence or absence of naloxone showed a significant interaction between esketamine and naloxone for peak duration (F(2, 37) = 0.4700, p = 0.6287) and AUC (F(2, 32) = 13.10, p < 0.0001). However, the opioid agonist morphine (10 mg/kg, i.p.) did not produce any dopamine effects using this preparation (Fig. 3D–F), indicating that MOR engagement alone is not sufficient to cause an increase in electrically-evoked dopamine release. We also compared ketamine to the high affinity and selective NMDAR non-competitive antagonist phencyclidine (PCP), which has 10-times higher affinity for NMDAR and >1000-times selectivity over opioid receptors [4, 5, 11, 33]. An equipotent dose of PCP (3 mg/kg, i.p.) produced a similar effect to esketamine (Fig. 3D–F), however this effect was not blocked by naloxone. To test whether these effects were due to prolonged activation of the dopaminergic terminals, we used the green calcium sensor GCaMP7s injected in the VTA and we placed a fiber optic/electrode in the NAc (Fig. 3G, H). Upon esketamine administration, we did not observe any quantitative difference in the main intraterminal Ca2+ increase (Fig. 3I–K), however we observed that the calcium dynamics traces have two components, one main increase event (caused by the electrical stimulus) followed by a secondary minor increase (Fig. 3H). Unfortunately, the location of the secondary increase was inconsistent between animals and oftentimes merged with the main release event and could not be quantified.

Overall, these data indicate that the electrically-evoked increases in dopamine are not mediated by a single pharmacological mechanism but that several neurochemical systems are involved. The method used to electrically evoke dopamine release necessarily produces local release of other neurotransmitters, which might in turn locally affect dopaminergic cell activity [34, 35]. Therefore, we used an optogenetic approach to selectively activate NAc dopaminergic terminals. We injected Channelrhodopsin-2 (ChR2) in the VTA, and the red fluorescent dopamine biosensor RdLight1 in the NAc. Then, we lowered an optic fiber into the NAc that was used both to stimulate ChR2 (470 nm, 50 Hz during 400 ms) and simultaneously record red fluorescence from RdLight1 (Fig. 3L, M). A bipolar electrode was also attached to the optic fiber, allowing us to compare electrical vs optogenetic stimulation (Fig. 3N) with alternating pulses in the same mouse. Blue-light stimulation produced transient increases in RdLight1 signal only in animals transduced with ChR2, but not in control mice (Fig. 3L), confirming that the signal represented dopamine release rather than artefactual blue-light excitation of this biosensor as recently described [36]. While the amplitude of the peaks generated by either electrical or optical stimulation was comparable in scale factors, the duration of the ChR2-induced peaks was markedly reduced (Fig. 3N, O), suggesting that the release of other neurotransmitters caused by electrical stimulation sustains the duration of the dopamine release. Esketamine (30 mg/kg) did not alter the dynamics of the optogenetic dopamine transients (Fig. 3O). Taken together, these results suggest that esketamine modulation of dopamine neurotransmission does not occur by direct modulation of the NAc dopaminergic neurons but rather by affecting other neurotransmitters locally in the NAc.

Esketamine rapidly decreases glutamatergic transmission in the NAc

To test whether alterations in glutamate release might contribute to the dopaminergic effects of esketamine, we injected mice with a mix of the green fluorescent glutamate sensor iGluSnFR and the red dopamine sensor RdLight1 (Fig. 4A, Fig. S3) in the NAc. In one set of mice (N = 6, 4 male and 2 female), we performed electrical stimulation under anesthesia using the same conditions described previously (Fig. 4B). Acute esketamine administration (10 or 30 mg/kg, i.p.) did not produce a significant dose-dependent effect on glutamate release (Fig. 4C–E). We then performed fiber photometry on awake free behaving mice. Mice (N = 5, male) were allowed to explore their home cage for 20 min and, after that, 10 and 30 mg/kg of esketamine was injected at 30 min intervals. In contrast to the data using the green dopamine biosensor dLight1.3b (Fig. 2C), we observed a fast increase in dopaminergic tone and a decrease in glutamatergic tone after esketamine injection (Fig. 4F, G). In addition, the peak analysis found a reduction on glutamate transient frequency (RM one-way ANOVA, (F(1.753, 7.010) = 6.654, p = 0.026; Fig. 4H) but no effect on the duration of individual glutamate transients (RM one-way ANOVA: F(1.353, 5.411) = 0.4576, p = 0.58; Fig. 4I) nor their amplitude (RM one-way ANOVA:; Fig. 4J). Taken together, these results suggest that the increase in esketamine-induced increase in dopamine activity in the NAc is accompanied by a decrease in glutamatergic activity.

Fig. 4: Esketamine has opposite effects on dopaminergic and glutamatergic transmission in the NAc.
figure 4

A Schematic representation of dopamine and glutamate sensors injected in the NAc. B Representative traces of electrically evoked extracellular dopamine (red) and glutamate (green) in the NAc. C–E Amplitude, duration and AUC of glutamate release upon electrical stimulation after administration of saline (black) or esketamine 10 mg/kg (orange) or 30 mg/kg (purple). F Average of extracellular glutamate (green) and dopamine (red) traces before and after esketamine administration (upper panel), or after saline administration (bottom panel). Each line is the average of 4–5 animals. G AUC of the tonic changes in dopamine (red) or glutamate (green) during 20 min after drug administration. Each dot represents one animal, and the bars represent mean ± SEM. H–J Quantification of the glutamate transients (frequency, amplitude, and duration). In all panels except (F) and (G), each dot represents the average of 3 to 5 trials per animal and the bars represent the mean of all dots; statistical significance, evaluated by Tukey’s post-hoc multiple comparisons test, * p < 0.05, ** p < 0.01, *** p < 0.001).

Full size image

Reward-evoked dopamine release is impaired by acute esketamine administration

The data presented above suggests that esketamine alters basal and stimulated dopaminergic activity. Dopamine is a key modulator of motivated behaviors [37]. Therefore, we examined the relationship between esketamine-induced increases in dopamine activity and motivated behavior. Mice (N = 24: 12 male, 12 female) were trained to nose poke on a port to obtain a delivery of a 20 mg sucrose pellet under increasing fixed-ratio (FR) schedules (1, 5 and 10). Nose poking on an inactive second port did not result in any pellet delivery. Once their responding on FR was stable (see Methods) they were moved to a progressive-ratio (PR) schedule, in which the number of pokes required to obtain a pellet exponentially increased after each pellet delivery within a session (Fig. S4). After 3 days on PR schedule to establish a baseline response, animals were administered with esketamine (10 mg/kg, i.p.; Fig. 5A, B) five minutes before the start of the session. One subset of mice was also pretreated with naloxone (3 mg/kg, i.p.) 10 min before esketamine administration. Female mice (Fig. A, C) reached a higher break point than male mice (Fig. 5B, D), but when the data were normalized to the rewards obtained under baseline conditions, no statistical differences by sex were observed and, therefore sex was not considered a statistical variable. Esketamine treatment reduced the total number of rewards obtained (Fig. 5E) and increased the time it required to achieve them (Fig. 5F), however naloxone did not block these effects. Esketamine effects were most pronounced during the first 20 min of the session, a time point corresponding to esketamine-induced hyperlocomotion (Fig. 1) and tonic changes in glutamate and dopamine (Fig. 4). To test whether their ability to respond correctly or retrieve pellets was impaired, we measured the ratio of correct responses and the time elapsed between pellet delivery and retrieval. Esketamine reduced the ratio of correct responses (Fig. 5G; REML showed a positive interaction between time bin and treatment: F(10,84) = 3.73, p = 0.0004), and also increased the pellet retrieval time (Fig. 5H; Brown-Forsythe ANOVA 12.21 F(4.000, 417.7) = 12.21, p < 0.0001).

Fig. 5: Esketamine administration impairs reward-evoked dopamine release.
figure 5

A–D Cumulative sucrose rewards obtained on a PR schedule after saline (black), esketamine (orange) or esketamine and naloxone (green) administration on days 0 (black) 1 and 5 of the procedure 5 days. E Total rewards obtained in a 1-h session for each depicted condition (each dot represents one animal, bars represent mean ± SEM of all animals). F Time in minutes the animals need to obtain half of the total amount of pellets (each dot represents one animal; floating bars represent 95% confidence intervals). G Time course of the percent of correct pokes in 10-min time bins. Each dot represents mean ± SEM of 12 animals per group. H Delay (s) between pellet delivery and pellet retrieval by the animal. Each dot represents one animal. I Representative trace of dLight1.3b and isosbestic control signals in the NAc during sucrose self-administration on FR5. Triangles represent animal actions: rewarded active poke (green), inactive poke (red) or pellet retrieval (black). J Heatmap representing extracellular dopamine variation on each pellet delivery event (aligned to the onset of the cue -black triangle-). Each row is one individual trial from 9 different animals (5–10 trials per animal). K Average of the corrected dLight1.3b signal aligned to the cue (top) or pellet retrieval (bottom) in control (black) or esketamine (10 mg/kg)- treated animals. Each trace is the average of 9 animals. L–N Amplitude, duration and AUC of the dLight1.3b signal increase during cue presentation, each dot represents the average of all trials (5–10) per animal (9) on saline of esketamine sessions. In all panels statistical significance, evaluated by the corresponding test (see main text): * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001).

Full size image

To evaluate dopamine dynamics during sucrose self-administration, we injected mice (N = 9 male) with the dopamine biosensor dLight1.3b in the NAc and performed fiber photometry while they performed sucrose self-administration on an FR5 schedule (Fig. 5I). The conditioned stimulus that signaled reward delivery elicited a transient dopamine increase (Fig. 5J, K) that lasted 1.1 ± 0.5 s (Fig. 5K, L). In the following session the mice were pretreated with esketamine (10 mg/kg, i.p.) five minutes before the start of the session. The mice that were administered esketamine also showed a transient increase in extracellular dopamine upon achievement of a reward (1.2 ± 0.5 s, paired t(8) = 0.81, p = 0.44, vs Saline condition). However, the amplitude of the dopamine transient was reduced by ~50% (ratio paired t(8) = 4.54, p = 0.0019, Fig. 5M). Consequently, the AUC of this dopamine transient in the animals treated with esketamine was also reduced by ~65% (ratio paired t(8) = 4.88, p = 0.0012, Fig. 5N). Overall, this data suggests that despite the higher dopaminergic activity elicited by esketamine in other experimental procedures, esketamine impairs reward-evoked dopaminergic activity.

Discussion

In this study we evaluated the dopaminergic effects of esketamine and their behavioral correlates in naïve male and female mice. Our experiments suggest that subanesthetic doses of esketamine (10 and 30 mg/kg) increase dopaminergic activity through a pharmacological mechanism of action in which MOR are necessary but not sufficient. The multifaceted molecular mechanisms of esketamine administration are evidenced by our behavioral results (Figs. 1 and 5), in which mice showed a sensitization to the locomotor effects (psychostimulant-like) of esketamine that is consistent with previous reports using esketamine or racemic ketamine [4, 38, 39]. However, the psychomotor effects of esketamine cannot be defined solely as an increase in distance traveled but also as an increase in speed and an active avoidance of the center of the arena. The over-avoidance of open areas in an open field has been classically attributed to increased anxiety [29] and this was further confirmed by the EPM [30] experiment, in which esketamine treated animals spent less time in the open arms of the maze. Although these effects seem contradictory with the expected antidepressant effects of esketamine that, among others, should reduce anxiety-like behaviors, they could also be explained by an alternative hypothesis: the psychoactive effects of the drug. Indeed, while in males, the locomotor effects of the drug were sensitized over 5 days of repeated esketamine injections, both male and female mice developed tolerance to the anxiogenic-like effects at day 5 (Fig. 1F, S1). In a similar scenario, esketamine had stronger effects during its first administration in animals performing the sucrose self-administration task, and a milder effect during its fifth consecutive administration in both male and female mice (Fig. 5). Overall, these results align with the clinical data, which show that patients develop tolerance to esketamine when assessed using the Clinical-Administered Dissociative Sates Scale (CADSS [40]) score, without a reduction in antidepressant efficacy [41]. Interestingly, pretreatment with the MOR selective antagonist naloxone partially blocked acute locomotor effects, but it did not affect the time spent exploring the center of the arena, the EPM score and sucrose self-administration, nor the tolerance that mice developed to both behaviors. Our data agree with recent data from Pomrenze and Vaillancourt et al. [42], who show that the locomotor effects of ketamine were opioid receptor dependent. However, opioid receptors were not involved in other behaviors involving analgesia or antidepressant-like efficacy on the forced swim test. Taken together these results agree with prior work evaluating the putative role of opioid receptors in the behavioral actions of esketamine [4, 8, 25, 43,44,45,46]. Nonetheless, while increasing clinical evidence suggests the opioid system is a key component of the antidepressant mechanisms of ketamine and esketamine [47, 48], it remains unclear whether its engagement is directly related, or inseparably tied to their abuse liability.

The dopaminergic system has been also widely implicated in both the pathophysiology and pharmacology of mood disorders and has been linked to both the antidepressant and reinforcing effects of ketamine [4, 14, 15, 49]. Hence, we aimed at unraveling the dopaminergic effects of esketamine on dopamine and dopamine-related behavior using fiber photometry, as well as potential neurochemical mechanisms to explain it. Our data using the fluorescent dopamine biosensor dLight1.3b [50] (Figs. 2 and 3) could not reproduce data from a previous report showing that racemic ketamine increased dopaminergic tone [49]. However, that study used another dopamine sensor (dLight1.1) which as a high affinity for dopamine (Kd = 330 nM), while dLight1.3b range has ~5 times lower affinity for dopamine (Kd = 1.6 µM) despite its higher dynamic range. When we used the red-shifted dopamine sensor RdLight [31] (which despite being structurally related to dLight1.3b shows higher affinity for dopamine Kd = 860 nM) we observed a tonic increase in extracellular dopamine (Fig. 4) like that reported by Simmler et al. [49] for racemic ketamine. Therefore, our results suggest that esketamine and racemic ketamine produce comparable dopaminergic tone effects. Here, the use of dLight1.3b allowed to extend these insights by measuring phasic dopamine [50, 51], and hence evaluate dopamine transients after esketamine administration in freely behaving mice and electrically or optogenetically evoked dopaminergic release during anesthesia (Figs. 2 and 3). Our results revealed that esketamine exerts a bimodal effect on dopaminergic activity by reducing the frequency of dopaminergic transients but at the same time increasing transient duration without changes in amplitude. Alternatively, the reduction in the frequency of dopamine transients might be explained by the combination of the low sensitivity of dLight1.3b for low concentrations of dopamine and the extension of the duration of dopamine transients, that might lead to the apparent suppression of smaller transients with small prominence over a slightly increased baseline signal.

The pharmacological effects of esketamine on both transient duration and dopaminergic tone were similar to those elicited by psychostimulants that act on dopamine re-uptake mechanisms, such as cocaine [31, 32, 52]. Prior reports suggested a direct interaction of ketamine with monoamine transporters or presynaptic receptors that directly control terminal excitability [17,18,19], although our own previous studies could not replicate these direct interactions [4, 53]. Our results measuring electrically evoked dopaminergic release in anesthetized mice (Fig. 3) agree with our awake data, as in both cases esketamine increased the half-life of extracellular dopamine after each release event. However, if ketamine was acting directly on receptors controlling the excitability of dopaminergic neurons (i.e. dopamine D2 autoreceptors) or the reuptake mechanisms (i.e. DAT), we would expect to see changes in intracellular calcium in the dopaminergic neurons along an extended half-life of the extracellular dopamine after optogenetic-evoked dopamine release, respectively. However, neither of these effects were observed (Fig. 3). Importantly, optogenetic stimulation with ChR2 and concurrent RdLight1 photometry allowed us to selectively stimulate the NAc that originate in the VTA (where the opsin was transduced) while measuring dopamine in the NAc. While esketamine had no effect on optogenetic stimulation, in the same animals, interspersed electrical stimulation events showed increases in the duration and AUC of the evoked dopamine transients, the same effect as described previously with the green sensor dLight1.3b.

Morphine would be expected to increase dopaminergic activity in the NAc by inhibiting MORs on GABAergic neurons of the rostromedial tegmental nucleus, effectively disinhibiting their inhibitory control onto dopaminergic neurons [54], suggesting that this experimental set up may not be able to detect changes in excitability of dopaminergic neurons at the VTA, perhaps because mechanism requires intact glutamate tone in the VTA [54]. Instead, local electrical stimulation necessarily causes local release of other neurotransmitters in the NAc. Activation of extrasynaptic NMDAR on glutamatergic terminals has been reported to inhibit electrically stimulated dopamine release [55]. Interestingly, a recent report also suggested that ketamine modulates one of the main glutamatergic inputs in the NAc, the mPFC, by the release of β-endorphin [25]. Therefore, the interplay between glutamate and opioid system might converge in a positive feedback mechanism by which spillover glutamate can act on inhibitory metabotropic receptors (i.e. mGluR2) on presynaptic dopaminergic terminals to inhibit dopamine release [55,56,57]. Therefore, we hypothesized that inhibition of presynaptic NMDAR by esketamine or PCP might reduce glutamate release and hence potentiate dopaminergic transmission. We indeed observed pro-dopaminergic effects of PCP and esketamine (Figs. 2, 3) and a concomitant decrease in extracellular glutamate by esketamine (Fig. 4). These results conflict with one of the leading hypotheses of esketamine’s antidepressant action, the so-called ‘glutamatergic burst’ [22, 58] which claims that low doses of esketamine (or ketamine) selectively target NMDAR receptors on GABAergic interneurons leading to overall disinhibition of pyramidal neurons and increases in glutamatergic neuron firing, or that the increase in glutamatergic transmission is due to other molecular targets such as AMPAR or increased BDNF. Nevertheless, the duration of the glutamate depression observed here was shorter than the sampling rate of the microdialysis studies [22] so glutamate could be increased at longer time scales and therefore be responsible for the sustained effects of the drug.

Our study had other limitations. The effects on dopaminergic and glutamatergic transmission were evaluated in the NAc of naïve mice and we do not know how these effects would be manifested in mice exposed to stress or other depression-inducing conditions. Taken together, here we report a complex interaction between neurotransmitter systems, therefore -as the sucrose self-administration data highlights- there is the possibility that the neurochemical effects of esketamine are both circuit- and state-dependent. Thus, further work is required to evaluate the neurochemical effects of esketamine under pathological conditions.

Materials and methods

Animals

Male and female adult wild-type (C57BL/6) mice (Charles River Laboratories, MA) (20–30 g and ~ 8–12 weeks old) were used for this study. Mice were housed in rooms with controlled temperature and humidity conditions in a 12 h light-dark cycle with access to food and water ad libitum. Experiments were performed during the light cycle. Mice that underwent viral injections were completely recovered from intracranial surgery before the experiments. Sample size was determined based on preliminary data and previous experience and no statistical methods were used to estimate sample size. All animals were allocated randomly to the experimental groups. Experimenters were not blind to the experimental group when performing the experiment but analyses were performed using assigned names to each sample which did not state the experimental group. All the experiments were approved by the Comitè d’Experimentació Animal from the Generalitat de Catalunya (CEEA40/22 and CEEA167/22) and conducted in accordance with the guidelines described in the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, 1996) and the European Communities Council Directive 2010/63/EU.

Drugs

(S)-(+)-ketamine hydrochloride was from Biosynth, Switzerland; Naloxone hydrochloride was from Sigma, MO; R(+)-SCH23390 hydrochloride was from Sigma-Aldrich, Merck Life Science S.L., Spain; D-Amphetamine sulfate and Phencyclidine hydrochloride were from Tocris Bioscience, UK; Morphine hydrochloride was from Alcaliber, S.A., Spain. All vehicle controls received saline (0.9% NaCl). All drugs were dissolved in sterile saline and administered intraperitoneally (i.p.) at 10 ml/kg except naloxone hydrochloride that was injected subcutaneously (s.c.) for the locomotor activity and EPM experiments.

AAV injections

Mice were anesthetized with isoflurane (4.5% induction, 1.5% maintenance) and placed in a stereotaxic apparatus (RWD Life Science, China). A small hole was drilled above the injection site for viral vector delivery and consequent optic fiber implantation. Mice were unilaterally injected with 500 nL of dLight1.3b [ssAAV-1/2-hSyn1-chI-dLight1.3b-WPRE-bGHp(A)] at 7.6 × 1012 vg/ml, or a mix of iGluSnFR(A184S) [pssAAV-5/2-hSyn1-SF_iGluSnFR(A184S)-WPRE-SV40p(A)] at 5.30 × 1012 (250 nL) and rdLight1 [ssAAV-1/2-hSyn1-chI-RdLight1-WPRE-bGHp(A)] at 6.7 × 1012 (250 nL), and 300 nL of RCaMP1.07 [ssAAV-1/2-hSyn1-chI-RCaMP1.07-WPRE-SV40p(A)] at 6.4 × 1012 or 300 nL of GCaMP7s [pssAAV-2-hSyn1- chI-jGCaMP7s-WPRE-SV40p(A)] at 7.7 × 1012 or 300 nL of ChR2(H134R) [ssAAV-8/2-hSyn1-hChR2(H134R)_EYFP-WPRE-hGHp(A)] at 6.4 × 1012. All viral vectors were purchased from ETH Viral Vector Facility, University of Zurich). Injections were done in the nucleus accumbens (NAc) shell based on the Paxino’s atlas stereotaxic coordinates (AP: 1.1, ML: 0.7, from bregma DV: −4.2 from dura mater) and in the Ventral Tegmental Area (VTA) (AP: −3.0, ML: −0.4, DV: −4.7 from bregma), respectively, using a 2 µL microsyringe (Hamilton Neuros) and a syringe pump (Stoelting, USA) at a flow rate of 50 nl/min. After virus infusion, the microsyringe was left in place for 5 min. In animals with permanent fiber implants, a fiber optic cannula (6 mm, 400 µm core diameter and 0.5 NA) (RWD Life Science) was lowered to the NAc shell and implanted using dental cement (Dentalon, Kulzer GmbH, Germany), the scalp was sutured and the animal was allowed to recover for at least 3 days before the start of the food self-administration training or 3–4 weeks before the fiber photometry recordings.

Locomotor activity

Before the experiment, mice were habituated to a 20 × 15 cm plexiglass arena with bedding material during two 20-min sessions. On the first experimental day (Day 1) mice were placed in the arena for 20 min prior to saline or esketamine (10 mg/kg, i.p.) administration and recorded with a top-view camera at 30 fps. Subsequently, locomotor activity was monitored for 30 min. Where indicated, mice received naloxone (3 mg/kg, s.c.) 10 min before esketamine or saline injection. Video recordings were analyzed using DeepLabCut™ [59], which allows to track the precise coordinates of multiple body parts (nose, right and left ear, centroid and tail base) independently and simultaneously, and follow the motion of these specific anatomical features across a series of video frames, to study patterns of movement. Total ambulatory distance and time spent in the center of the arena (defined as the middle one-third of the height and width of the arena) were measured using Matlab.

Elevated Plus Maze

Anxiety-like behavior was measured using the elevated plus maze (EPM), which consisted of two open arms (65 × 6 cm) perpendicular to two enclosed arms (66 × 6 cm) at a height of 50 cm above the ground. A recording camera was placed on the top of the maze. 15 min after injection of esketamine 10 mg/kg (i.p.), or saline, mice were placed individually in the center of the maze and allowed to explore it for 5 min. Where indicated, naloxone (3 mg/kg, s.c.) was administered 10 min before esketamine. The amount of time spent in the open and enclosed arms as well as overall distance traveled by the mouse was analyzed using DeepLabCut™ and Matlab.

Fiber photometry experiments

Three to four weeks after AAV injections and optic fiber implantation, to allow for recovery and sensor expression, fiber photometry experiments were performed. Extracellular dopamine (dLight1.3b or RdLight1) and glutamate (iGluSnFR) or intracellular calcium (RCaMP1.07) from VTA terminals were recorded while mice were exploring a 20 × 15 cm cage for 30 min after a single saline or esketamine (10 and 30 mg/kg, i.p.) administration, or during sucrose self-administration (see below). Where indicated, mice received naloxone (3 mg/kg, s.c.) 10 min before esketamine or saline administration. A cohort of mice was then treated for 4 more consecutive days (daily single esketamine administration, 10 mg/kg, i.p.) and the photometry experiment was repeated on the last day of treatment.

Another cohort of mice was placed in a stereotactic apparatus, an optic fiber was lowered to the NAc shell during anesthesia and release of dopamine (dLight1.3b or RdLight1) or glutamate (iGluSnFR) or dopaminergic intracellular calcium (GCaMP7s) were elicited by electrical stimulation in the same brain region, using a bipolar electrode (Plastics One) customly attached to either side of the optic fiber stud. Each electrical stimulus (400 ms sequence of 2 ms pulses at 50 Hz and 0.4 mA) [32, 52] was repeated every 3 min. After baseline collection (10 min or until stable pulses), mice received saline, naloxone (3 mg/kg), esketamine (10 or 30 mg/kg), morphine (10 mg/kg) or PCP (3 mg/kg) injections and data were acquired for another 30 min. The indicated animals were pretreated with naloxone (3 mg/kg i.p.) 10 min before esketamine or PCP administration. Electrically evoked dopamine transients were compared across experimental groups and averaged at equivalent time windows (20 to 30 min from saline or drug administration) to account for putative non-specific changes over time.

In another set of animals, extracellular dopamine was measured with the red-shifted dopamine sensor RdLight1 and release was elicited by optogenetic stimulation (470 nm light pulses of 10 ms at 50 Hz during 400 ms) of the terminals of VTA cells expressing ChR2 or controls without ChR2 expression.

In all cases, excitatory light was delivered using T-Cube™ LED drivers (Thorlabs) coupled with the integrated LEDs of Doric Minicubes, the emission light was optically filtered and registered using the built-in photodetectors of the Minicubes. All signals were controlled and recorded using a Fiber Photometry Console (Doric) and Neuroscience Studio V6 (Doric).

Changes in dopamine, glutamate or calcium were recorded as changes in fluorescence emission, based on excitation by light source at 470 nm and 535 nm for green and red biosensors respectively, and 405 nm for excitation of an isosbestic control for decay and movement artifact correction. All data was processed using custom Matlab® code. Data was downsampled (10x) and low-pass (10 Hz) filtered. The isosbestic signal was rescaled to the biosensor signal levels by a polynomial fit to the minima of the biosensor signal. Then, the biosensor signals were corrected using the formula dFF = (F-F0/F0), where F is the fluorescence of the signal at a given time point and F0 is the corresponding rescaled isosbestic signal. To identify transients, a prominence over baseline threshold was set for every animal on a baseline recording and it remained constant across all experimental conditions for that animal. Then the signal was analyzed to isolate and quantify the properties (amplitude, duration, frequency, decay and AUC) of the spontaneous fluorescence transients or elicited peaks. To measure slow changes in dopamine concentration, data was further low pass filtered at 0.5 Hz to eliminate transients and the traces from all animals were aligned and averaged.

At the end of each experiment, mice were sacrificed, brains were dissected, and immunohistochemistry (see below) was performed to confirm the expression of the viral vectors and the correct location of the optic fiber. Animals without proper viral expression were excluded from the study.

Sucrose self-administration

Mice were trained to self-administer sucrose pellets (20 mg/pellet) using the Arduino-based Feeding Experimentation Device version 3.1 (FED3.1) (Open Ephys Production Site, Portugal) [60], under fixed ratio (FR) and progressive ratio (PR) schedules. Mice were trained to nose poke on the left port to obtain a delivery of a 20 mg sucrose pellet under increasing fixed-ratio (FR) schedules (FR1: 2 days, FR5: days, FR10 until meeting training criteria). Nose poking on the right port was recorded but did not result in any pellet delivery or penalty. Once their responding on FR was stable (less than 20% variation between consecutive sessions) and their responses on the inactive port were below 15% they were moved to a progressive-ratio (PR) schedule. No animals were excluded based on these criteria. On the PR schedule the amount of pokes required to obtain a pellet exponentially increased after each pellet delivery within a session according to the formula: 6 · e(X + 1) * 0.2–5 where X is the nth reward obtained; After 3 days on PR schedule to establish a baseline responding, animals were administered with esketamine five minutes before the start of the session. After training, mice underwent 5 daily PR sessions, being the first and the fifth session preceded by a single injection of esketamine 10 mg/kg (i.p), to test the acute effect of esketamine before and after the repeated exposure to the drug, while on the second to fourth sessions, mice received esketamine 10 mg/kg (i.p) after each session. A subgroup of mice also received naloxone (3 mg/kg, i.p.) 10 min prior to esketamine injection on the first and fifth day.

Another cohort of mice was injected with dLight1.3b in the NAc and implanted with a fiber optic in the same region as described previously and was trained to self-administer sucrose pellets under an FR5 schedule. The FED3.1 BNC output and the optic fiber from the mouse brain were connected to the Doric console, allowing to record synchronously photometry signals and FED3.1 actions. Mice were connected and habituated to the cable during the last two training sessions prior to the experiment. On the first day of the experiment, dopamine was recorded using fiber photometry during the task for 60 min, while the second day the mice received a single injection of esketamine 10 mg/kg (i.p.), 5 min prior to the task. The photometry data was processed as described previously and the peri-event signals (active or inactive nosepoke, and pellet retrieval) were isolated, aligned, averaged and analyzed using custom Matlab scripts.

Histology

Animals were anesthetized and perfused with 4% PFA solution, followed by brain collection. Brains were fixed in 4% PFA solution at 4 °C overnight and then placed in sucrose 30% in PBS for 72 hr at 4 °C. Then they were frozen at −16 °C and cut on a cryostat (Leica Biosystems) at 40 µm-thick slices. Brain slices were further processed for immunohistochemistry. For immunohistochemistry, 40 µm slices were washed in phosphate buffered saline (PBS, pH = 7.4), then in PBS with 0.03% Triton X-100, PBS with 0.05% triton X-100 and incubated in blocking buffer - PBS, pH = 7.4 with 0.05% triton X-100 and 2% Albumin Bovine Serum Cohn Fraction V (Life Science) - for 1 h. Slices were then incubated with chicken anti-GFP polyclonal antibody (1:500, Invitrogen), rabbit anti-tyrosine hydroxylase polyclonal antibody (1:1000, Sigma-Aldrich) and anti-mApple (1:250, MyBioSource) in blocking buffer overnight. Then, slices were washed in PBS with 0.05% Triton X-100 and incubated for 2 h in blocking buffer containing goat anti-chicken IgY conjugated with Alexa Fluor 488 (1:1000, ThermoFisher), donkey anti-rabbit IgG conjugated with Alexa Fluor 647 (1:1000, Fisher Scientific) and donkey anti-goat conjugated with Alexa Fluor 555 (1:1000, ThermoFisher) secondary antibodies. After three five-minute washes at RT using PBS with Triton X-100 (0.05%), slices were mounted on a microscope slide with Ever Brite Hardset Mounting Medium with DAPI (Biotium) and imaged using Zeiss Apotome 3 microscope.

Statistics

All statistical analyses were conducted using Prism 10 (GraphPad Software, La Jolla, CA, USA). Unpaired or paired Student’s t-tests were used to compare means between groups of independent or repeated measures, respectively. Repeated measures one-, two-way or three ANOVA were performed to assess differences among three or more groups. Mixed models were used when there were missing values. Brown-Forsythe ANOVA tests and Kruskall-Wallis tests were used when the data showed different SD between groups or non-gaussian distribution, respectively. No sphericity was assumed, and Geisser-Greenhouse correction was used. Results were considered statistically significant when p < 0.05. All statistical tests performed for each experiment and the relevant parameters are described in the Results section or Figure legends.