Orexinergic innervations at GABAergic neurons of the lateral habenula mediates the anesthetic potency of sevoflurane

Abstract Aims The circuitry mechanism associated with anesthesia‐induced unconsciousness is still largely unknown. It has been reported that orexinergic neurons of the lateral hypothalamus (LHA) facilitate the emergence from anesthesia through their neuronal projections to the arousal‐promoting brain areas. However, the lateral habenula (LHb), as one of the orexin downstream targets, is known for its anesthesia‐promoting effect. Therefore, the current study aimed to explore whether and how the orexinergic projections from the LHA to the LHb have a regulatory effect on unconsciousness induced by general anesthesia. Methods We applied optogenetic, chemogenetic, or pharmacological approaches to regulate the orexinergicLHA‐LHb pathway. Fiber photometry was used to assess neuronal activity. Loss or recovery of the righting reflex was used to evaluate the induction or emergence time of general anesthesia. The burst‐suppression ratio and electroencephalography spectra were used to measure the anesthetic depth. Results We found that activation of the orexinergicLHA–LHb pathway promoted emergence and reduced anesthetic depth during sevoflurane anesthesia. Surprisingly, the arousal‐promoting effect of the orexinergicLHA‐LHb pathway was mediated by excitation of glutamate decarboxylase (GAD2)‐expressing neurons, but not glutamatergic neurons in the LHb. Conclusion The orexinergicLHA–LHb pathway facilitates emergence from sevoflurane anesthesia, and this effect was mediated by OxR2 in GAD2‐expressing GABA neurons.

role in promoting wakefulness. The orexinergic system conveys information to multiple output regions through the two peptides (orexin A [OA] and orexin B [OB]) and their widespread receptors (orexin receptor 1 [OxR1] and orexin receptor 2 [OxR2]). The extracellular levels of OAs have been reported to be high during the active wake state. [7][8][9] Our lab and others using electroencephalogram (EEG) and behavioral testing have demonstrated that orexinergic neurons facilitate emergence from general anesthesia through their projections to the excitatory down-stream neurons that promote emergence from anesthesia, such as the dopaminergic neurons of the ventral tegmental area and cholinergic neurons of the basal forebrain. 10,11 The lateral habenula (LHb) is one of the main downstream targets of orexinergic innervations. [12][13][14] Notably, although the LHb is predominantly composed of the excitatory glutamatergic neurons, the LHb glutamatergic output is active during propofol-and isofluraneinduced anesthesia, and activation of LHb glutamatergic neurons effectively shortens anesthesia induction and prolongs emergence from general anesthesia. 15,16 Therefore, whether the orexinergic neurons promote emergence via the anesthesia-facilitating center of the LHb is not only an interesting question, but the underlying mechanisms would also supplement understanding for the networks between the arousal center and the anesthesia center. In this study, we used chemogenetic, optogenetic, EEG recordings, and behavioral tests to investigate the role of orexinergic LHA-LHb projections in the regulation of consciousness and its neuronal mechanism in sevoflurane anesthesia, an extensively used volatile anesthetic.

| Animals
All experiments were in accordance with protocols approved by the Ethics Committee for Animal Experimentation and strictly abided by the Guidelines for Animal Experimentation of the Fourth Military Medical University. C57/BL6J male mice were obtained from Vital River Laboratory Animal Technology Co., Ltd. Glutamate decarboxylase 2 (GAD2)-Cre male mice were purchased from Jackson Laboratory. Hcrt-Cre male mice were the present from Luis de Lecea (Stanford University). Mice (aged 8-12 weeks) were housed separately in cages before experiments with the temperature of 23°C (22-24°C) and humidity of 40% (38%-42%) in a 12/12-h light-dark environment with food and water ad libitum.

| Stereotaxic surgery
After being anesthetized with 1.2% isoflurane and fixed on the stereotaxic apparatus, mice were exposed of skull and cleaned with 10% hydrogen peroxide. During the surgery, the mice were maintained warm by a heating plate. After viruses were microinjected (described later in this section) into the LHA (AP: −1.6 mm, ML: 0.4 mm, DV: −4.5 mm) using a Nanoject III injector (Drummond Scientific) at a rate of 23 nL/min, an optical fiber (diameter, 300 μm, Inper, Hangzhou, China) was implanted into the ipsilateral LHb (AP: −1.6 mm, ML: −0.4 mm, DV: −2.5 mm) for optogenetic experiments, or a guide cannula (RWD, Inc.) for chemogenetic or pharmacological manipulations. To obtain EEG recording, three stainless-steel screws were anchored to the skull as electrodes: the positive and negative electrodes on the left and right sides of the head, respectively, or vice versa (AP: −1.5 mm, ML: −1.5 mm) and a reference electrode at the back of the head (AP: −5.5 mm, ML: 0 mm). After surgery, mice were housed for recovery for 7 days. A total of 3 weeks was required for the expression of virus in the brain.

| Optogenetic stimulation
During deep anesthesia, mice remained unconscious and lost the righting reflex. The deep anesthesia was achieved by controlling the inhalational concentration of sevoflurane at around 2.4% under 100% O 2 at a flow rate of 1.5 L/min, during which mice presented the steady burst-suppression ratio (BSR) on the EEG up to 60-70%.
Burst-suppression pattern of EEG is a fundamental characteristic of a deeply inactivated brain during anesthesia. Under deep anesthesia, optogenetic manipulation was applied. The 2-min blue laser (473 nm, 20 Hz, 30 ms, 15 mW from tips, Thinker Tech) was applied for activation, whereas 2-min yellow laser (594 nm, 1 Hz, 1 s, 10 mW from tips, Thinker Tech) was applied for inhibition. Light anesthesia was achieved by administering the inhalational concentration of sevoflurane at 1.4-1.7% to avoid the presentation of burst suppression waves. During the state of light sevoflurane anesthesia, blue light (473 nm, 20 Hz, 30 ms, 15 mW from tips, for 120 s) was used to activate neurons. The position of each optical fiber tip was verified experimentally, and the EEG spectrogram was simultaneously recorded.

| Fiber photometry for recording calcium signals
We attached the optical fiber to the fluorescence photometer (Thinker Tech) and confirmed the laser intensity at the fiber tip was between 30 and 40 μW persistently. The average ΔF/F values were calculated by a custom-written MATLAB code, as (F duration-F baseline)/F baseline, in which F baseline was the mean calcium signal before time zero.

| Behavioral assessment
Mice were placed in a horizontal cylindrical observation cage  (2). With regard to the righting and walking status of each animal during opto-stimulation, retention of LORR was scored as 0, and the appearance of RORR was scored as 2.
Subsequently, the RORR scoring was performed as follows: walking with no further movements, 0; crawling without the abdomen off the chamber bottom, 1; or walking with the abdomen off the chamber bottom, 2. The total score was the sum of the scores for body movements, righting, and walking. 17

| EEG recording and analysis
Electroencephalogram signals were continuously recorded by the PowerLab 16/35 amplifier system and LabChart Pro version 8.0.10 software (AD Instruments). Raw EEG data were digitized at a rate of 1000 Hz and were bandpass filtered at 0.3-50 Hz. The burstsuppression ratio (BSR), total power percentage, and spectrum were analyzed using MATLAB (R2014a; MathWorks).
To calculate the BSR, the EEG voltage threshold was set based on the amplitude of suppression of each animal. If the amplitude of EEG was lower than the threshold for 0.5 s, it was defined as a suppression event and assigned a value of 1. Otherwise, signals above the threshold were defined as burst events and were assigned a value of 0. Finally, the BSR was calculated as the percentage of suppression events for 2 min before and during optical stimulation. 11,18

| Immunohistochemistry
Immunofluorescence assay was performed as described previously. 19,20 In specific, brain slices were incubated with anti-OxR2

| Statistical analysis
using the Shapiro-Wilk test. Statistical significance was assessed using the student's t-test for comparison between two groups. Twoway ANOVA followed by Bonferroni correction was used to evaluate statistical differences in the chemogenetic experiments. Behavioral scores for arousal are presented as median (25-75th percentile) and were analyzed using the Mann-Whitney U-test.

| Activation of orexinergic terminals in the LHb reduces the burst-suppression pattern of EEG during deep sevoflurane anesthesia
To further investigate the effect of the LHb orexinergic terminals on deep anesthesia maintenance, we modulated its excitability using an optogenetic approach when animals anesthetized with 2. When the burst suppression pattern was regularly induced for at least 10 min, the optical activation of the orexinergic terminals in the LHb significantly decreased BSR from 67.48% to 31.69% (p = 0.0006).
As the optical stimulation ceased, the BSR returned to the prestimulation state ( Figure 2E,I). In comparison, the BSR was increased from 64.78% to 77.54% (p = 0.0081) by the photo-inhibition of the LHb orexinergic projections, and also returned to the pre-stimulation state once the light was turned off ( Figure 2G,K). As shown in Figure 2D,F,H,J, the burst suppression pattern of the mCherry group did not change across the optical stimulation (p = 0.9242). Therefore, the orexinergic terminals in the LHb might be involved in the regulation of the sevoflurane anesthetic depth.

| Activation of the orexinergic terminals in the LHb induces cortical activation and wakeful manifestation of mice from light anesthesia
We further investigated the regulatory effect of the LHb orexinergic terminals during light anesthesia (steady sedation with 1.4-1.7% sevoflurane) using acute optical stimulation (Figure 2). The optical activation of the orexinergic terminals in the LHb induced an immediate behavioral wakeful manifestation in ChR2-expressing mice, including the appearance of movements of legs, heads, and whiskers, as well as righting and walking (Table S1). In contrast, mice in the controlled mCherry group exhibited none of the above behavioral manifestations during photostimulation. The overall arousal score of the behaviors was much higher in the ChR2 group than the mCherry group (p < 0.001).
In addition, photostimulation of the orexinergic terminal in the LHb also induced a brain-state transition in EEG from a slowwave activity pattern to a fast-wave activity pattern in ChR2 mice ( Figure 2M), but not in mCherry mice ( Figure 2L). Spectral analysis of the EEG revealed that acute photostimulation induced a significant decrease in delta power ( Figure 2P, p = 0.0061, n = 6), theta power ( Figure 2P, p = 0.0002, n = 6), and alpha power ( Figure 2P, p < 0.0001, n = 6). After the cessation of laser stimulation, the EEG spectral signatures gradually returned. Collectively, these findings indicated that activation of the LHb orexinergic terminals was sufficient to induce cortical activation and behavioral emergence in mice from the light anesthesia state. As shown in Figure 3F Figure 3G) and longer emergence time (TCS vs. Con:

| Administration either orexin A (OA) or orexin B (OB) at LHb decreases the depth of sevoflurane anesthesia as detected by EEG
We further microinjected OA, OB, SB334867, and TCS-OX2-29, respectively, into the LHb to observe their individual regulatory effects during maintenance of deep anesthesia ( Figure 4A-F).

| Knockdown of OxR2 in GAD2 neurons but not in glutamate neurons of the LHb prolongs emergence from sevoflurane anesthesia
Previous studies confirmed that OxR2 in the LHb is highly expressed in GAD2-positive GABAergic neurons. Therefore, we selectively

| DISCUSS ION
In this study, we found that chemogenetic activation of orexinergic LHA-LHb projections could significantly facilitate emergence from sevoflurane anesthesia. Optogenetic activation of orexinergic terminals in the LHb not only induced cortical arousal in terms of the change in EEG signatures during deep anesthesia but also led to behavioral emergence during light anesthesia. We found that administration of two peptides of orexins, OA, and OB in LHb decreased the depth of anesthesia and promoted emergence from sevoflurane anesthesia mainly through OxR2 in GAD2-expressing neurons but not glutamatergic neurons in the LHb. Therefore, we proposed that LHb GAD2-expressing neurons mediated the emergence-facilitative effect possibly through local inhibitory circuit on glutamatergic neurons.
LHb is one of the major projected targets of the orexin system.
The orexin LHA-LHb pathway has been widely reported to be involved in motivated behavioral regulations. 7,9 Particularly, orexin terminals in the LHb induced aggressive behavior in male mice by activating LHb GAD2-expressing GABA neurons, and this was mediated by the feedforward inhibition of LHb glutamatergic neurons. 14 Previously we also reported that orexin neurons and their terminal in the LHb significantly alleviated depression-like behaviors caused by chronic social defeat stress via activating the LHb glutamate neurons. 20 In the current study, we firstly reported a substantial arousal-facilitating effect of orexinergic terminals in the LHb during sevoflurane anesthesia by using chemogenetic, opotogenetic, and pharmacological approaches. Relevantly, Gelegen et al. found that blocking the LHb output caused natural sleep to be more fragmented, and this was largely reversed by systemic administration of dual orexinergic antagonist. 15 Moreover, we previously found the stimulation of glutamatergic terminals in the LHb derived from the LHA, where orexin neurons are exclusively located, also induced similar facilitation of emergence during isoflurane anesthesia. 16 It is worth noting that more than 50% of orexinergic neurons co-release glutamate from their terminals, [21][22][23] providing further support for the arousal effect of the orexin LHA-LHb pathway during general anesthesia.
As previously reported, OxR1 is specifically activated by the OA, but OxR2 combines evenly with both OA and OB. Both OA and OB, or OxR1 and OxR2, have been reported to involve in anesthesia emergence. 11,24 In this study, administration of both OA and OB to the LHb prolonged the induction time of sevoflurane anesthesia, isoflurane. 15 The anesthesia-facilitating effect of LHb glutamate neurons was speculated through their downstream of the rostromedial tegmental nucleus. 16 As a matter of fact, Flanigan reported a feedforward local inhibitory regulation between the GAD2-expressing GABA neurons and the vGlut2-expressing glutamate neurons within the LHb, suggesting that GABA neurons activation-induced glutamate neurons inhibition during anesthesia emergence may be involved in the pro-arousal effect of orexinergic terminals in the LHb. 14 To test this hypothesis, we recorded calcium signals of LHb GABA neurons over the whole sevoflurane anesthesia procedure and found the decrease of GABAergic activity during anesthesia and the recovery during emergence, which was consistent with the changes of excitability of orexin neurons and contrary to that of glutamate neurons in the LHb as reported previously. 16 We also blocked the local GABAergic effect by administering the inhibitor of GABA A receptor and found the emergence from sevoflurane was prolonged. In consideration of the great number of glutamate neurons in the LHb, the arousal-promoting effect of the LHb GABA neurons is rather likely to be mediated by the inhibition of LHb Glutamate neurons. Further investigations are needed to obtain more solid evidence.
The present study had some limitations. By administering exogenous OA and OB in the LHb, we confirmed the involvement of orexin peptides in regulating sevoflurane anesthesia, but the actual release of orexins in the LHb terminals still needs experimental confirmation. Orexin sensors may help to quantify the orexin release.
Moreover, in addition to the local inhibitory projections, LHB GABA neurons also have long-range projections in the brain; we did not fully clarify the downstream circuits of LHb GABA neurons in promoting arousal, which requires more studies in the future.
In conclusion, the present study provides evidence for the role of orexin LH-LHb in facilitating emergence from sevoflurane anesthesia, and this effect was mediated by OxR2 in GAD2-expressing GABA neurons. This study further expands on the effects of the LHb on general anesthesia. And, since LHb is highly associated with negative emotional behaviors, our findings could have implication regarding perioperative psychiatric disorders.

AUTH O R CO NTR I B UTI O N S
Qianzi Yang and Hailong Dong initiated and supervised the study;

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors state that there are no conflicts of interest to disclose.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.