Dexmedetomidine Activation of Dopamine Neurons in the Ventral Tegmental Area Attenuates the Depth of Sedation in Mice

Only male mice were used in these experiments. The procedures of the animal experiments in this study strictly followed institutional guidelines and governmental regulations. To ensure the welfare of the animals, we tried our best to replace, refine, and reduce the mice in our experiments, and we will follow this principle in future research. All experiments were approved by the Institutional Animal Care and Use Committee at ShanghaiTech University, Shanghai, China, and the Chinese Academy of Sciences, Shanghai, China. Adult male DAT-Cre knock-in mice (Stock No. 006660, Jackson Laboratory, China) and male C57BL/6J mice (Shanghai Model Organism, China) at 8 to 12 weeks of age were used in our experiments. Our experiments did not examine the effect of dexmedetomidine-induced sedation in mice of different sexes since we did not use female mice. Before all experiments were performed, all mice were in good health at a normal weight (22 to 28 g). After all experiments were completed, the remaining mice were euthanized by isoflurane. All mice were maintained under a reversed 12/12 h day/night cycle (light at night and dark at day) at 22 to 25°C with ad libitum access to rodent food. All experiments were performed between 9 am and 9 pm.

See Supplemental Digital Content, table S1 (https://links.lww.com/ALN/C382), listing baseline characteristics of the experimental mice.

AAV vectors containing the DIO-GFP, DIO-GCaMP6m (fluorescent calcium indicators), DIO-mCherry, DIO-hM4Di (Gi-coupled human M4 muscarinic receptor), DIO-hM3Dq (Gq-coupled human M3 muscarinic receptor), DA1m (G protein-coupled receptor-activation-based dopamine sensor) and DA1m-mut (mutant G protein-coupled receptor-activation-based dopamine sensor) constructs were packaged into the AAV2/9 serotype with titers of 1 to 5 × 10¹² viral particles per milliliter. The surgical procedures generally followed our previous studies.²⁵  Briefly, after being deeply anesthetized with isoflurane in a small box, the mice were placed on the stereoscopic positioning instrument and were kept warm (37°C) with an electric heating pad (BrainKing Biotech, China). Anesthesia was kept constant with 1 to 1.5% isoflurane mixed with pure oxygen at a 1 l/min flow rate. After the scalp was cut open, 3% hydrogen peroxide was used to remove the fascia from the skull surface. The bregma and lambda points were used to adjust the mouse head to the horizontal position. A small window with a diameter of 300 to 500 μm was opened at the location of the point of virus injection and fiber implantation. The virus was injected at 100 nl/min for a total of 300 nl using a microsyringe pump (Nanoject III #3-000-207, Drummond, USA). After the injection, the injector was allowed to remain in place for 10 min. The injection points were as follows: ventral tegmental area (coordinates from bregma: –3.10 mm [anteroposterior], ±0.45 mm [mediolateral], –4.25 mm [dorsoventral]), medial prefrontal cortex (+1.60 mm [anteroposterior], ±0.30 mm [mediolateral], –1.90 mm [dorsoventral]) and nucleus accumbens (coordinates from bregma: +1.18 mm [anteroposterior], ±1.3 mm [mediolateral], –3.9 mm [dorsoventral]). Unilateral injections were performed for fiber photometry, but bilateral injections were performed for chemogenetic experiments. For the recording fiber implantations, the fiber (200-μm OD, 0.37 numerical aperture, Newdoon, China) was slowly inserted and implanted above the targeted brain areas at a distance of 0.15 mm. Then, the optical fiber and skull were fixed with dental cement. After the dental cement was completely dried, the mice were removed from the stereoscopic positioning instrument and placed on an electric blanket to keep warm (37°C). After the mice had fully recovered, they were returned to their home cage.

Fiber recording also followed the procedures described in our previous studies.²⁵  We used a fiber photometry system including one light-emitting diode at 488 nm as an excitation laser, a dichroic mirror with a 505 to 544 emission filter, and a photomultiplier tube (R3896, Hamamatsu, Japan). The analog voltage signals were converted to digital signals by a Power 1401 digitizer and recorded by Spike2 software (CED, United Kingdom) at 100 Hz. An optical fiber (RJPSF2, Thorlabs, USA) with an integrated rotary joint that prevented fiber damage from the animal movement was used to guide the light between the fiber photometry system and the implanted optical fiber. The laser power was adjusted at the tip of the optical fiber to a low level of 20 to 40 µW to minimize bleaching. The recordings were made after the animals had recovered from surgery for 2 weeks. While recording, the mice were allowed to move freely in a white chamber (40 × 40 × 40 cm, length × width × height) for at least 1,000 s. Then, the mice were intraperitoneally injected with saline (10 μl/g), and the recording continued for another 1,200 s. After that, the mice were administered dexmedetomidine (10, 40, 100, or 400 μg/kg; 10 µl/g) and recorded for at least 6,000 s. The data were extracted and analyzed by MATLAB R2018b (MathWorks, USA). To normalize the data, the ΔF/F was calculated by (F – F0) / F0, where F0 is the baseline fluorescence signal averaged over a window of at least 20 s, and F is the real-time fluorescence signal. ΔF/F values were expressed as percentages (i.e., multiplied by 100) and presented with average plots with a shaded area indicating the SD. To calculate the average response and increased duration in ΔF/F values, we first segmented the data based on behavioral events. Then, we calculated the average calcium signals after both saline administration and dexmedetomidine administration.

Immunostaining of ventral tegmental area dopamine neurons was performed after all the experiments had been completed. The mice were anesthetized by intraperitoneal injection of pentobarbital (100 mg/kg), and then normal saline was perfused through the heart. After most of the blood was drained, 4% paraformaldehyde was used for fixation. The perfusion was stopped when the limbs became stiff. To further fix the brain tissue, the mice were decapitated, and the heads were soaked in 4% paraformaldehyde at room temperature overnight. The brain tissue was removed the next day and then soaked in phosphate-buffered saline containing 30% sucrose for dehydration for 24 to 48 h. We used a freezing microtome (CM3050S, Leica, Germany) to collect 40-μm-thick coronary brain slices. The slices were washed with phosphate-buffered saline and blocked with a blocking solution (phosphate-buffered saline containing 5% bovine serum albumin and 0.3% Triton X-100 [Acros Organics, Belgium]) for 2 h. After that, primary antibodies (rabbit polyclonal anti– tyrosine hydroxylase; 1:1,000; Catalog no. p21962, Invitrogen, USA) were added to the blocking solution, and slices were incubated for 24 h. The next day, the solution was washed away, and the slices were washed at least three times with phosphate-buffered saline. The secondary antibody used was Alexa Fluor 594 donkey anti-rabbit IgG (1:1,000; Jackson ImmunoResearch, USA) diluted in phosphate-buffered saline and was allowed to coat the brain slices overnight at room temperature. 4’,6’-Diamidino-2-phenylindole dihydrochloride staining was used to identify cell bodies, and 10% glycerine was used to mount the slides. Finally, the slices were imaged with a 20× objective on an Olympus VS120 microscope (Olympus, Japan) and a Nikon A1R confocal microscope (Nikon, Japan) using a 40× objective.

The medial prefrontal cortex and nucleus accumbens slices were prepared in the same manner as the ventral tegmental area slices, but immunohistochemistry staining was with 4’,6’-diamidino-2-phenylindole dihydrochloride staining only. They were imaged with a 20× objective on an Olympus VS120 microscope and a Nikon A1R confocal microscope using a 40× objective.

Adult (8 to 12 weeks old) male DAT-Cre mice injected with AAV2/9-hsyn-DIO-mCherry in the ventral tegmental area were anesthetized with pentobarbital (100 mg/kg, intraperitoneal injection) and then transcardially perfused with ice-cold oxygenated (95% O₂, 5% CO₂) N-methyl-D-glucamine, a solution that included 93 mM N-methyl-D-glucamine, 93 mM hydrogen chloride, 2.5 mM potassium chloride, 1.25 mM monosodium phosphate, 10 mM magnesium sulphate, 30 mM sodium bicarbonate, 25 mM glucose, 20 mM HEPES, 5 mM sodium ascorbate, 3 mM sodium pyruvate, and 2 mM thiourea. After perfusion, the brain was quickly dissected out and immediately transferred into an ice-cold oxygenated N-methyl-D-glucamine artificial cerebrospinal fluid solution. Then, we sectioned the brain tissue in the coronal plane at a 300-µm thickness in the same buffer using a vibratome (VT1200 S, Leica, Germany). The brain slices containing the ventral tegmental area were incubated in oxygenated N-methyl-D-glucamine artificial cerebrospinal fluid at 32°C for 10 to 15 min and then transferred to a normal oxygenated solution of artificial cerebrospinal fluid (126 mM sodium chloride, 2.5 mM potassium chloride, 1.25 mM monosodium phosphate, 2 mM magnesium sulphate, 10 mM glucose, 26 mM sodium bicarbonate, and 2 mM calcium chloride) at room temperature for 1 h. All chemicals used in the slice preparation were purchased from Sigma-Aldrich (USA).

Slices were transferred to the recording chamber that was submerged and superfused with artificial cerebrospinal fluid at a rate of 3 ml/min at 28°C. Dopamine-positive neurons were identified with differential interference contrast optics (Olympus BX61WI, Japan). The recording pipettes (3 to 4 MΩ) were prepared with a micropipette puller (P2000, Sutter Instrument, USA). For whole cell recording, the pipettes were filled with artificial cerebrospinal fluid solution containing 133 mM potassium ion gluconate, 18 mM sodium chloride, 0.6 mM EGTA, 10 mM HEPES, 2 mM magnesium–adenosine triphosphate, and 0.3 mM sodium guanosine 5′-triphosphate sodium salt hydrate (pH: 7.2, 280 milliosmole). After the whole cell recording was established, the neurons were held at 0 pA under a current-clamp mode to record spontaneous action potentials. For action potentials evoked by current injections, a current-step protocol (from −50 to +250 pA, with 50 pA increments) was run and repeated. Dexmedetomidine (2 μM) was locally perfused to the recording cell through a drug perfusion system. RS79948 (Tocris Bioscience, United Kingdom; 1 mM) was diluted in fresh artificial cerebrospinal fluid until completely mixed and then transferred to separate graduated reservoirs connected to the chamber (1 to 2 ml/min). During the recordings, baseline levels of spontaneous action potentials were recorded for at least 3 min before the drugs were perfused. To isolate potassium ion currents during the whole cell recordings, tetrodotoxin (1 µM), cadmium chloride (200 µM), 6,7-dinitroquinoxaline-2,3-dione (20 µM), D(-)-2-amino-5-phosphonovaleric acid (20 µM), and picrotoxin (100 µM) were added to the artificial cerebrospinal fluid bath solution to block sodium channels, calcium channels, glutamate receptors, and γ-aminobutyric acid type A receptors, respectively. Potassium ion currents in the ventral tegmental area dopamine cells, in response to depolarizing voltage steps ranging from 0 to +140 mV (in increments of 10 mV), were acquired using a Multiclamp 700B amplifier, and signals were low-pass filtered at 3 kHz and digitized at 10 kHz (DigiData 1550, Molecular Devices, USA). The physiologic data were analyzed using Clampfit 10 software (Molecular Devices, USA) and Mini Analysis Program (Synaptosoft, USA).

We implanted EEG electrodes following previously described procedures.²⁶  Briefly, after being deeply anesthetized by intraperitoneal injections of pentobarbital (50 mg/kg), the mice were placed on the stereoscopic positioning instrument and were kept warm (37°C) with an electric heating pad. Three holes were drilled with a carbide bit: one over the left frontal cortical area (1 mm anterior to bregma, 1.5 mm lateral to midline), another over the parietal area (1 mm anterior to lambda, 1.5 mm lateral to midline) of the right hemisphere, and a third over the parietal area (1 mm anterior to lambda, 1.5 mm lateral to midline) of the left hemisphere. Then, three electrodes were implanted in the holes and fixed in place with dental cement. After the dental cement was completely dried, the mice were removed from the stereoscopic positioning instrument and placed on an electric blanket. When the mice fully recovered, they were returned to their home cage. EEG recordings were conducted 14 days after surgery. EEG data were digitized at 1 kHz by the PowerLab and LabChart system (AD Instruments, New Zealand), low-pass filtered at 0.3 Hz, and high-pass filtered at 50 Hz. The power spectrum between 0.1 and 30 Hz was transformed using the “mtspecgramc” function in the MATLAB signal processing toolbox chronux_2_12 (MathWorks, USA).²⁷  The EEG changes between frequency bands were used to estimate the sedation states: delta (δ: 0.5 to 4 Hz), theta (θ: 4 to 7 Hz), alpha (α: 8 to 15 Hz), and beta (β: 16 to 30 Hz).²⁸ 

Before the testing day, the mice were handled for 30 min on each of 5 days and adapted to having the EEG device connected to their head for 30 min in the testing room on each of 3 days. They were put into the testing room 1 h before the behavioral test for acclimation. The mice were then intraperitoneally injected with clozapine N-oxide (1 mg/kg). After 30 min, they were injected with dexmedetomidine (40 μg/kg), and EEG was immediately recorded for at least 1 h.

We randomized animals to the different treatment groups using the random number table method, and mice in each group were tested in ascending order. We also blinded the experimenters to the drug treatment to reduce selection and observation bias. None of the variables had missing data. Outliers were evaluated, and there were no outliers in our data. Graph Pad Prism 7.0 (GraphPad Software, USA), MATLAB R2018b, and IBM SPSS Statistics 19 software (IBM, USA) were used for statistical evaluation. Quantile–quantile plots were used to assess whether the sample conformed to a normal distribution. A paired Student’s t test was used for the fiber recording and patch-clamp recording. Two-way repeated-measures ANOVA with Bonferroni posttest was used for the patch-clamp recording and EEG recording. The Kruskal–Wallis one-way ANOVA test was used to detect the difference in EEG waveform distributions among the different groups over time. No statistical methods were used to predetermine the sample size. All tests were two-tailed, and P < 0.05 was considered significant in all tests. Data were accurately reported according to the original data to avoid false precision. Unless otherwise indicated, values were reported as the mean ± SD.

To study the effect of systemic administration of dexmedetomidine on ventral tegmental area dopamine neurons, we injected the Cre-dependent AAV-EF1α-DIO-GCaMP6m virus into the ventral tegmental area of DAT-Cre mice and used fiber photometry to record changes in Ca²⁺ signals in vivo (fig. 1A). The virus was expressed as in figure 1B, and the calcium indicator GCaMP6m was expressed in dopamine neurons (fig. 1C). In freely moving mice, the GCaMP6m signals ranged from –20 to 60% (ΔF/F), and the control GFP signal showed little fluctuation, indicating that the fluctuations were caused by Ca²⁺ binding (Supplemental Digital Content, fig. S1, https://links.lww.com/ALN/C383).²⁵  In response to peer review, additional dose-response experiments with dexmedetomidine were performed. The results showed that after intraperitoneal injection of saline or dexmedetomidine (10 µg/kg), the Ca²⁺ signals exhibited almost no change (fig. 1D). Statistical analysis showed that the average Ca²⁺ signals after administration of dexmedetomidine were approximately equal to those after administration of saline (dexmedetomidine, 0.460 [–1.283; 1.734], median [25%; 75%] vs. saline, –0.591 [–1.153; 0.640], normalized data, P = 0.751; n = 6 mice; fig. 1E). However, after intraperitoneal injection of dexmedetomidine (40, 100, and 400 µg/kg), the Ca²⁺ signals significantly increased (fig. 1, F, H, and J). Statistical analyses showed that the average Ca²⁺ signals after administration of dexmedetomidine were significantly higher than those after administration of saline (dexmedetomidine, 16.917 [14.882; 21.748], median [25%; 75%] vs. saline, –0.745 [–1.547; 0.359], normalized data, P = 0.001; n = 6 mice; fig. 1G; dexmedetomidine, 17.383 [12.325; 20.994], median [25%; 75%] vs. saline, –0.400 [–1.821; 0.048], normalized data, P < 0.0001; n = 6 mice; fig. 1I; dexmedetomidine, 21.738 [20.403; 27.561], median [25%; 75%] vs. saline, 0.087 [–1.722; 0.544], normalized data, P < 0.0001; n = 6 mice; fig. 1K).

To investigate whether dexmedetomidine had an influence on the change in dopamine neurotransmitter in the related forebrain projection areas of ventral tegmental area dopamine neurons, we injected AAV-hsyn-DA1m virus into the medial prefrontal cortex and nucleus accumbens of C57BL/6J mice.²⁹  Virus expression and fiber position are shown in figure 2, A and B, and figure 3, A and B. DA1m signals ranged from –5 to 15% (ΔF/F) when the test mouse was freely moving. There was almost no change in the DA1m-mut signals in the control group (Supplemental Digital Content, fig. S1, https://links.lww.com/ALN/C383). This implied that the signal fluctuations were caused by changes in dopamine neurotransmitter.²⁹  In response to peer review, additional dose-response experiments for dexmedetomidine were performed. The results showed that after the injection of normal saline and dexmedetomidine (10 μg/kg), the DA1m signals exhibited almost no change (figs. 2C and 3C). Statistical analyses showed that the average DA1m signals after administration of dexmedetomidine were approximately equal to those after administration of saline (dexmedetomidine, –0.708 [–1.118; 0.376], median [25%; 75%] vs. saline, 0.093 [–0.003; 0.249], normalized data, P = 0.241; n = 6 mice; fig. 2D; dexmedetomidine, 0.714 [–0.430; 1.382], median [25%; 75%] vs. saline, 0.096 [0.029; 0.396], normalized data, P = 0.326; n = 6 mice; fig. 3D). However, after the injection of dexmedetomidine (40, 100, and 400 μg/kg), the DA1m signals markedly increased (figs. 2E, 2G, 2I, 3E, 3G, and 3I). Statistical analyses showed that the DA1m signals in the medial prefrontal cortex and nucleus accumbens were significantly increased after systemic administration of dexmedetomidine (dexmedetomidine, 10.815 [9.713; 15.104], median [25%; 75%] vs. saline, –0.498 [–0.664; -0.355], normalized data, P = 0.001; n = 6 mice; fig. 2F; dexmedetomidine, 7.521 [5.615; 12.809], median [25%; 75%] vs. saline, –0.274 [–0.428; –0.136], normalized data, P = 0.003; n = 6 mice; fig. 2H; dexmedetomidine, 15.312 [12.330; 16.337], median [25%; 75%] vs. saline, –0.181 [–0.878; 0.016], normalized data, P < 0.0001; n = 6 mice; fig. 2J; dexmedetomidine, 8.543 [7.135; 11.828], median [25%; 75%] vs. saline, –0.329 [–1.220; –0.047], normalized data, P = 0.001; n = 6 mice; fig. 3F; dexmedetomidine, 11.309 [10.557; 14.164], median [25%; 75%] vs. saline, –0.706 [–1.226; –0.251], normalized data, P < 0.0001; n = 6 mice; fig. 3H; dexmedetomidine, 11.296 [8.199; 12.092], median [25%; 75%] vs. saline, –0.339 [–0.561; 0.089], normalized data, P = 0.002; n = 6 mice; fig. 3J).

To further investigate the mechanism by which dexmedetomidine activated ventral tegmental area dopamine neurons, we used slice electrophysiology technology. Whole cell recording showed that ventral tegmental area dopamine neurons were intrinsically active and tonically fired. Interestingly, the firing frequency clearly increased after dexmedetomidine perfusion (fig. 4A). There was a statistically significant difference in the firing frequencies before and after dexmedetomidine perfusion (before: 5.046 ± 5.057 Hz and after: 8.600 ± 7.387 Hz, P = 0.031; n = 11 neurons; fig. 4B). The current-injected firing also showed that after dexmedetomidine perfusion, the peak density visibly increased compared to that with artificial cerebrospinal fluid perfusion (fig. 4C). Statistical analyses revealed that as the injection current increased from –50 pA to 250 pA, the number of spikes was significantly higher after dexmedetomidine perfusion than after artificial cerebrospinal fluid perfusion, especially at 150 pA and 200 pA (F[1, 27], P < 0.0001; n = 28 neurons; fig. 4D).

In response to peer review, additional experiments providing a more in-depth electrophysiologic characterization for mechanistic insights were performed. Given that dexmedetomidine is a highly selective α2 adrenoceptor agonist, we used a selective α2 adrenoceptor antagonist (RS79948)^30,31  to explore whether it can antagonize the effect of dexmedetomidine activation of ventral tegmental area neurons. The results showed that the intrinsic neuronal firing after dexmedetomidine (2 μM) perfusion was similar to artificial cerebrospinal fluid in the presence of RS79948 (1 mM; fig. 5A). Statistical analysis indicated that the firing frequency showed no difference between artificial cerebrospinal fluid and dexmedetomidine in the presence of RS79948 (before: 3.970 ± 1.927 Hz and after: 3.810 ± 1.849 Hz, P = 0.100; n = 10 neurons; fig. 5B). Koo et al.³²  found that morphine increased the excitability of ventral tegmental area dopamine neurons by inhibiting potassium ion currents. We speculated that dexmedetomidine activation of ventral tegmental area dopamine neurons was also related to this mechanism. In voltage-clamp mode, the potassium ion current was significantly suppressed after perfusion with dexmedetomidine compared with artificial cerebrospinal fluid (fig. 5C). The peaks of the potassium ion currents in the ventral tegmental area dopamine neurons were reduced by dexmedetomidine compared with artificial cerebrospinal fluid (F[1, 14] = 16.09, P = 0.001; n = 15 neurons; fig. 5D). However, the potassium ion currents showed almost no difference between artificial cerebrospinal fluid and dexmedetomidine perfusion in the presence of RS79948 (fig. 5E). The peaks of the potassium ion currents in the ventral tegmental area dopamine neurons also showed no statistically significant difference between artificial cerebrospinal fluid and dexmedetomidine perfusion in the presence of RS79948 (F[1, 9] = 4.019, P = 0.076; n = 10 neurons; fig. 5F).

We used chemogenetics to inhibit ventral tegmental area dopamine neurons to test for exacerbation of dexmedetomidine sedation. In response to peer review, additional experiments with chemogenetic activation of ventral tegmental area neurons to test for resistance to dexmedetomidine sedation were also performed. The experimental design is shown in figure 6A. We bilaterally infused Cre-dependent AAV-DIO-hM4Di-mCherry, AAV-DIO-hM3Dq-mCherry, and AAV-DIO-mCherry vectors into the ventral tegmental area of DAT-Cre mice (fig. 6B). After administration of clozapine N-oxide, ventral tegmental area dopamine neurons in the brain slice preparations were significantly inhibited or activated (fig. 6, C and D). During the test, all mice underwent systemic delivery of clozapine N-oxide (1 mg/kg, intraperitoneally) 30 min before systemic delivery of dexmedetomidine (40 μg/kg, intraperitoneally) and EEG recording. The dose of dexmedetomidine is sufficient to induce slow-wave sedation in mice.^4,33  The EEG waveform pattern and spectrum showed that the power of low-frequency waves was increased in the hM4Di group and reduced in the hM3Dq group in comparison to the mCherry group (fig. 6E). The percentages of delta waves per minute among the mCherry, hM3Dq, and hM4Di groups were significantly different (F[2, 33] = 8.016, P = 0.002, n = 12 mice, fig. 6F; P < 0.0001, n = 12 mice, fig. 6G). The percentages of delta waves per minute between the mCherry and hM3Dq groups were significantly different (F[1, 22] = 6.974, P = 0.015, n = 12 mice). There were no significant differences between the mCherry and hM4Di groups (F[1, 22]=1.294, P = 0.268, n = 12 mice). However, there was a trend that the percentages of delta waves per minute in the hM4Di group were greater than those in the mCherry group (fig. 6, F and G). The percentages of theta waves per minute among the mCherry, hM3Dq, and hM4Di groups were also significantly different (F[2, 33] = 22.8, P < 0.0001, n = 12 mice, fig. 6H; P < 0.0001, n = 12 mice, fig. 6I). The percentages of theta waves per minute between the mCherry and hM3Dq groups were significantly different (F[1, 22]=14.82, P = 0.001; n = 12 mice). The percentages of theta waves per minute between the mCherry and hM4Di groups were also significantly different (F[1, 22]=7.348, P = 0.013; n = 12 mice).

The purpose of this study was to examine the mechanisms underlying arousability properties of dexmedetomidine using in vivo Ca²⁺ measurement, in vitro slice recording, chemogenetics, and EEG recording. The results showed that ventral tegmental area dopamine neurons were activated and dopamine neurotransmitters were increased in the targeted forebrain regions after systemic administration of dexmedetomidine. Patch-clamp recording showed that dexmedetomidine activated ventral tegmental area dopamine neurons through the α2 adrenoceptors and inhibited potassium ion conductance by activating the α2 adrenoceptors. EEG recordings confirmed that after the inhibition of ventral tegmental area dopamine neurons, the depth of sedation with dexmedetomidine was deepened. Conversely, after activation of ventral tegmental area dopamine neurons, the depth of sedation with dexmedetomidine was attenuated.

Patch-clamp recording revealed that dexmedetomidine inhibited potassium ion conductance and activated ventral tegmental area dopamine neurons. However, after perfusion with an α2 adrenoreceptor antagonist (RS79948), there were no statistically significant differences in the excitability and potassium ion conductance of ventral tegmental area dopamine neurons. Koo et al.³²  found that morphine increased the excitability of ventral tegmental area dopamine neurons by inhibiting potassium ion conductance. Therefore, dexmedetomidine inhibited potassium ion conductance in the ventral tegmental area dopamine neurons by activating α2 adrenoceptors on the ventral tegmental area dopamine neurons, resulting in increased excitability of ventral tegmental area dopamine neurons. This led to an increase in dopamine neurotransmitter in the targeted forebrain regions, i.e., the nucleus accumbens and medial prefrontal cortex.

Dexmedetomidine is commonly used as a sedative drug. It has primarily been thought to target the norepinephrine system. Many studies have shown that dexmedetomidine acts on α2 adrenoceptors on locus coeruleus norepinephrine neurons and then hyperpolarizes cell membranes through a Gi-coupling mechanism. The release of the excitatory neurotransmitter norepinephrine is subsequently decreased throughout the brain, which eventually leads to sedation.³  Previous studies have also shown that ventral tegmental area dopamine neurons are associated with sleep and arousal.^16–21,34  However, the relationship between dexmedetomidine and activity in the ventral tegmental area dopamine system has remained unknown. Our results showed that dexmedetomidine activated ventral tegmental area dopamine neurons and increased dopamine concentrations in the targeted forebrain regions. After the inhibition of ventral tegmental area dopamine neurons, sedation with dexmedetomidine deepened. This is consistent with previous results showing that inhibiting ventral tegmental area dopamine neurons promoted sleep in mice.¹⁸  Conversely, after activation of ventral tegmental area dopamine neurons, sedation with dexmedetomidine was attenuated. This is consistent with previous results showing that activating ventral tegmental area dopamine neurons promoted awakening in mice.^17,18  We conclude that dexmedetomidine activates ventral tegmental area dopamine neurons (arousal-promoting neurons), which can lower the threshold of arousal during dexmedetomidine sedation. This may be involved in the unique sedative feature of dexmedetomidine that enables patients to be easily arousable (fig. 7).

Some studies have focused on changes in dopamine concentrations in the nucleus accumbens or medial prefrontal cortex after systemic administration of dexmedetomidine. Whittington et al.²³  and Whittington and Virág²⁴  found that dexmedetomidine decreased the concentrations of accumbal dopamine in rats, and this effect was partly mediated via the locus coeruleus. Ihalainen and Tanila^22,35  also showed the same results in the medial prefrontal cortex and nucleus accumbens of mice. However, our findings are not consistent with these studies. One reason for this may be that we used different animal strains. Another reason may be that we used different experimental methods. Previous studies have used microdialysis to measure changes in dopamine neurotransmitter concentrations in the medial prefrontal cortex and nucleus accumbens. However, there are several evident disadvantages of the microdialysis method. The temporal resolution of microdialysis is considerably slower than many dynamic processes in the central nervous system.³⁶  In addition, the probe (>250 µm in diameter) used in microdialysis can induce extensive tissue damage with inflammation, gliosis, and swollen axons.³⁷  Borland et al.³⁸  found that after acute implantation, the release of dopamine was strongly inhibited in the vicinity of the probe, as the dialysis probe was surrounded by scar tissue soon after implantation, and therefore, the microdialysis method might represent a neuropathological rather than a neurophysiological condition. However, we used fiber photometry recording technology to confirm that dexmedetomidine activated ventral tegmental area dopamine neurons and increased forebrain dopamine neurotransmitters. The optical fiber was implanted above the brain area and would not cause damage to the brain area of interest. We also confirmed that dexmedetomidine activated ventral tegmental area dopamine neurons in vivo and in vitro and how dexmedetomidine activated ventral tegmental area dopamine neurons using patch-clamp technology in vitro.

Our experiment also has some limitations. For example, our experiments used all male mice. However, many clinical drugs show sex differences,³⁹  and our experiments did not examine the effect of dexmedetomidine-induced sedation in mice of different sexes. Furthermore, in addition to the ventral tegmental area dopamine system, there might be other important neural systems involved in the sedation induced by dexmedetomidine, which should also be studied.

Based on our findings, dexmedetomidine may have many potential applications. It is well known that aberrant dopamine concentrations are also associated with several neurologic diseases, such as Parkinson’s disease,⁴⁰  depression,⁴¹  Alzheimer’s disease,⁴²  and more. Whether dexmedetomidine can treat these illnesses should be the subject of future studies.

In summary, dexmedetomidine activates ventral tegmental area dopamine neurons and increases dopamine concentrations in the forebrain, which attenuates the depth of sedation in mice. This mechanism may explain rapid arousability upon dexmedetomidine sedation.

The authors thank Yulong Li, Ph.D. (State Key Laboratory of Membrane Biology, Peking University School of Life Sciences, Beijing, China), for providing DA1m virus; and Suwen Zhao, Ph.D. (School of Life Science and Technology, ShanghaiTech University, Shanghai, China) and Fang Bai, Ph.D. (School of Life Science and Technology, ShanghaiTech University, Shanghai, China), for valuable discussion and comments during revision of the manuscript.

This work was supported by the National Natural Science Foundation of China (Beijing, China; grant Nos. 31922029, 31671086, 61890951, and 61890950 to Dr. Hu and grant No. 81701049 to Dr. Zhu.)

The authors declare no competing interests.