Effect of basal forebrain somatostatin and parvalbumin neurons in propofol and isoflurane anesthesia

Abstract Aims The basal forebrain (BF) plays an essential role in wakefulness and cognition. Two subtypes of BF gamma‐aminobutyric acid (GABA) neurons, including somatostatin‐expressing (GABASOM) and parvalbumin‐positive (GABAParv) neurons, function differently in mediating the natural sleep–wake cycle. Since the loss of consciousness induced by general anesthesia and the natural sleep–wake cycle probably share similar mechanisms, it is important to clarify the accurate roles of these neurons in general anesthesia procedure. Methods Based on two transgenic mouse lines expressing SOM‐IRES‐Cre and PV‐IRES‐Cre, we used a combination of genetic activation, inactivation, and chronic ablation approaches to further explore the behavioral and electroencephalography (EEG) roles of BFSOM and BFParv neurons in general anesthesia. After a single intravenous injection of propofol and the induction and recovery times of isoflurane anesthesia, the anesthesia time was compared. The changes in cortical EEG under different conditions were also compared. Results Activation of BF GABASOM neurons facilitates both the propofol and isoflurane anesthesia, manifesting as a longer anesthesia duration time with propofol anesthesia and a fast induction time and longer recovery time with isoflurane anesthesia. Moreover, BF GABASOM‐activated mice displayed a greater suppression of cortical electrical activity during anesthesia, showing an increase in δ power bands or a simultaneous decrease in high‐frequency power bands. However, only a limited and nuanced effect on propofol and isoflurane anesthesia was observed with the manipulated BF GABAParv neurons. Conclusions Our results suggested that BF GABASOM neurons play a critical role in propofol and isoflurane general anesthesia, while BF GABAParv neurons appeared to have little effect.


| INTRODUC TI ON
General anesthetics have been widely used since the introduction of it in the 1840s. 1 However, few explicit mechanisms have been elucidated that explain how general anesthetics cause a sudden reversible loss of consciousness. The sedation effects of anesthetics, like calmness, drowsiness, and muscle relaxation, are behaviorally similar to the features of endogenous sleep, especially in the nonrapid eye movement (NREM) period. 2,3 Moreover, some whole-brain imaging studies also showed that the state of "unconsciousness" during deep sleep and anesthesia are remarkably similar. 4 Recently, there has been growing appreciation that neural pathways that regulate the endogenous sleep-wake systems are involved in general anesthesia. [5][6][7] Thus, several studies on the mechanisms of general anesthesia have focused on the sleep-wake control systems.
The basal forebrain (BF), a large heterogeneous structure in the ventral forebrain, receives projections from the ascending reticular activating system (ARAS) that are then projected to the cerebral cortex. 8,9 The BF is key to sleep-wake control. BF has three main neuronal populations: cholinergic, glutamatergic, and GABAergic neurons. 10 Some studies have suggested that stupor or coma can be induced by destroying all these neurons in the BF, 5 whereas the specific lesion of BF cholinergic neurons produced limited changes in arousal, including in the sleep-wake cycles. 5,[11][12][13] This means the cholinergic neurons are more likely to be related to wakefulness, rapid eye movement (REM) sleep, arousal, and memory than NREM sleep. 11,12,14,15 Glutamatergic and GABAergic neurons potentially have a critical roles in sleep-wake control. 10, 13 We previously found that the activity of cholinergic neurons in BF could influence propofol and isoflurane anesthesia process. 16 Furthermore, several studies have indicated that propofol decreases the activity of BF cholinergic neurons via GABA A receptors. 17,18 GABAergic neurons in the BF are mainly separated into two subtypes with opposite functions: wakefulness-promoting parvalbumin-expressing (GABA Parv ) neurons and sleep-promoting somatostatin-expressing (GABA SOM ) neurons. 10,19 To clarify their accurate functions in the anesthesia process, we destroyed them respectively and used chemogenetics methods to activate and inactivate the two types of neurons in mice. Moreover, the mice were subjected to behavioral test and electroencephalograph (EEG) under propofol and isoflurane anesthesia. Our findings suggest that GABA SOM neurons in the BF region play a critical role in modulating general anesthesia.

| Animals
All experimental procedures were approved by the Animal Care and Use Committees of Zunyi Medical University, Guizhou, China, and followed the ARRIVE guidelines and the Guide of Care and Use of Laboratory Animals. 20 Adult (8-12 weeks, 20-25 g) SOM-IRES-Cre (stock N° 018973, Sst tm2.1(cre)Zjh /J), and PV-IRES-Cre (stock N° 008069, 129P2-Pvalb tm1(cre)Arbr /J) male mice were used in all experiments (provided by Prof. Minmin Luo, National Institute of Biological Science, Beijing, China). Under the control of the SOM/ PV gene promoter, all experimental mice were genotyped by PCR and identified as Cre recombinase. All animals were maintained in an ambient temperature of 23 ± 0.5°C with a relative humidity of 55 ± 2% and 12-h light/12-h dark cycle (light on at 8:00 am). Food and water were provided ad libitum.

| Stereotactic surgery
The mice were anesthetized with pentobarbital (40 mg/kg, intraperitoneal [i.p.]) and then placed on a stereotaxic apparatus (RWD Life Science, Shenzhen, China). Lidocaine (1%) was subcutaneously injected for local anesthesia before exposing the surface of the skull. and PV-IRES-Cre (n = 8) mice, respectively, through a glass micropipette (1-mm glass stock, tapering to a 10-20 micron tip) using a micro-syringe pump. The pipette was kept in the region for 10 min to allow the virus to diffuse and was then slowly withdrawn. The electroencephalographic electrodes were placed on the skull (AP: +1.0 mm, ML: ±1.5 mm; AP: −3.5 mm, ML: ±1 mm) simultaneously. 22,23 The animals underwent further behavioral testing and EEG recording after 3 weeks.

| Experimental procedure
All dosage of behavioral testing and EEG recording were unified. The loss of righting reflex (LORR) and recovery of righting reflex (RORR) time in mice are considered standardized indexes of the induction and emergence time of general anesthesia, respectively. For propofol anesthesia, a single dosage of 20 mg/kg was intravenously injected through the caudal vein with mouse injection fixation devices (Chuangbo Global Biotechnology Co., Ltd., Beijing). The duration of anesthesia, which means the time from LORR to RORR, was recorded. For isoflurane anesthesia, the mice were placed in a recording chamber (RWD Company, Shenzhen, China) filled with 1.4% isoflurane with oxygen at a rate of 1 L/min. The time interval from the start of the isoflurane application to when the mice demonstrated LORR for 30 s was determined as the latency to LORR. The mice were kept anesthetized with 1.4% isoflurane for 30 min and then immediately and gently removed from the chamber. Then, the RORR was quantified in a supine position in room air. The latency to RORR refers to the duration of time before isoflurane stops acting on the mice, and they revert to a prostrate position and landing on all fours.
In the lesioned group, we recorded the duration and EEG under propofol anesthesia, as well as LORR, RORR, and EEG under isoflurane anesthesia ( Figure 1A). In chemogenetics groups, Clozapine N-oxide (CNO) (1 mg/ml, 1 mg/kg, i.p.) or saline (0.9%, equal volume, i.p.) were injected randomly 1 h before behavioral testing and EEG recording. The duration and EEG were recorded under propofol anesthesia, as well as LORR, RORR, and EEG were recorded under isoflurane anesthesia ( Figure 1B,C). During all tests, a heating pad with a rectal temperature probe was used to keep the mouse body temperature at 37•C. All mice were sacrificed and subjected to immunofluorescence to verify viral expression and specific transfection after all the experiments were performed. All experiments were performed between 7:00 AM and 6:00 PM .

| EEG recording and spectral analyses
Electroencephalography signals were captured during all the experimental procedures using a neuronal recording system (Appolo, Bio-Signal Technologies, USA). These were then digitized and analyzed using the Spike2 software package (Cambridge Electronic Design, Cambridge, United Kingdom). Delta (δ), theta (θ), alpha (α), beta (β), gamma (γ), and total spectral powers were calculated using the frequency bands 1-4, 5-8, 9-12, 13-25, 26-60, and 1-60 Hz, respectively. Relative powers were calculated by dividing the averaged signal power across the frequency range of each band by the total power in 1-60 Hz. Furthermore, GraphPad Software was used for the statistical analysis.

| Perfusion and immunofluorescence
All mice were deeply anesthetized with pentobarbital for the perfusion of the phosphate-buffered saline (PBS), followed by 4% paraformaldehyde (PFA). The brains were removed and post-fixed in PFA overnight at 4°C and put in 30% sucrose in PBS at 4°C until they sank. The brains were coronally sectioned into 30μm slices using a cryostat (Leica CM1950).
The hM3Dq and hM4Di expressing mice were injected with CNO (1 mg/ml, 1 mg/kg, i.p.) or saline (0.9%, equal volume, i.p.), and then kept in their home cage for 2 h before perfusion.
For immunofluorescence, the brain sections were first incubated in blocking solution (PBS containing 2.5% normal goat serum, 1.5% bovine serum albumin, and 0.1% Triton™ X-100) for 2 h at room temperature. The sections were then incubated with the primary antibody (c-Fos staining, rabbit anti-c-Fos, 1:500, Synaptic Systems; SOM staining, sc-74556, 1:100, SantaCruz; PV staining, ab104224, 1:1000, Abcam) in a blocking solution overnight at 4°C and washed with PBS. The sections were then incubated with the secondary antibody (goat anti-rabbit Alexa 594 and Alexa 488, 1:1000, Invitrogen; goat anti-mouse, Alexa 488, Invitrogen, 1:1000) at room temperature for 2 h. After another wash with PBS, the sections were mounted on glass slides and cover-slipped with a mounting media (Gold antifade reagent with DAPI, Life Technologies, USA). All images were captured on the virtual microscopy system (Olympus BX63).

| Statistical analysis
The statistical analysis was performed using commercial software (GraphPad Prism; GraphPad Software). All data were subjected to Kolmogorov-Smirnov tests for normality. Unpaired student's t-tests were to detect all behavioral differences between the lesioned and control groups and cell counts. The differences in the chemogenetics behavioral recording experiments (hM3Di-Saline and hM3qi-CNO; hM4Di-Saline and hM4Di-CNO) were detected by paired t-test. A two-way analysis of variance (ANOVA) was used to analyze the EEG recordings. For all results, significant threshold was placed at *p < 0.05, **p < 0.01, ***p < 0.001, and p > 0.05 was considered non-significant (n.s.). All data were shown as mean ± SEM. F I G U R E 1 A, The behavioral and EEG recording experiment procedure diagram of neuron lesioned group. B, The behavioral and EEG recording experiment procedure diagram of hM3Dq and hM4Di group under propofol anesthesia. C, The behavioral and EEG recording experiment procedure diagram of hM3Dq and hM4Di group under isoflurane anesthesia. D, The diagram of injection sites (showing by the red boxes) of AAV-CAG-DIO-DTA virus or saline in BF. And immunoflurescence of BF GABASOM neurons in control (left) and lesioned mice (right), scale bar: 100 μm. E, Mean numbers of BF GABASOM neurons in control and lesioned mice (54.00 ± 2.6 to 8.83 ± 2.88). F, The duration time of propofol anesthesia in BF GABASOM neuron lesioned group, n = 8, p < 0.05, unpaired t-test. G, H, The LORR (G) and RORR (H) time of isoflurane in BF GABASOM neuron lesioned group, n = 8, *p < 0.05, **p < 0.01, unpaired t-test. I, Representative EEG traces under propofol anesthesia in BF GABASOM neuron lesioned and control mice. J, The power spectrum analysis of EEG recording in single dose propofol anesthesia of BF GABASOM neuron lesioned group, n = 8, *p < 0.05, **p < 0.01, two-way ANOVA. K, Representative EEG traces under isoflurane anesthesia in BF GABASOM neuron lesioned and control mice. L, The power spectrum analysis of EEG recording under 30 min isoflurane anesthesia in BF GABASOM neuron lesioned group, n = 8, **p < 0.01, ***p < 0.001, two-way ANOVA. All graphs show mean ± SEM [Colour figure can be viewed at wileyonlinelibrary.com]

| Chemogenetic inactivation of BF GABA SOM and GABA Parv neurons in propofol and isoflurane anesthesia
To reversibly inactivate the BF GABA SOM or GABA Parv neurons, we injected AAV-Ef1α-DIO-hM4Di-mcherry and AAV-Ef1α-DIO-mCherry virus in the transgene mice separately. The immunofluorescence images show the virus transfection of the BF GABA SOM (Figure 5A,B, Figure S1C) and GABA Parv neurons ( Figure 5F,G, Figure S2C). The paired t-test). Furthermore, they were also harder to anesthetize with isoflurane ( Figure 6B,C Moreover, the slow-delta power bands of the EEG during anesthesia showed the similar effect by the inhibition (Figure 6D-G, Propofol, hM4Dq group, 0.32 ± 0.008 to 0.27 ± 0.02, p = 0.038, n = 8; Isoflurane, hM4Di group, 0.31 ± 0.02 to 0.25 ± 0.02, p = 0.029, n = 8, F I G U R E 4 A, The duration time of propofol anesthesia in BF GABASOM neuron with hM3Dq group, n = 8, *p < 0.05, ***p < 0.001, paired t-test. B, C, The LORR (B) and RORR (C) time of isoflurane anesthesia in BF GABASOM neuron with hM3Dq group, n = 8, **p < 0.01, paired t-test. D, E, The power spectrum analysis (D) and representative EEG traces (E) of EEG recording of mice with activated BF GABASOM neurons in propofol anesthesia, n = 8, ***p < 0.001, two-way ANOVA. F, G The power spectrum analysis (F) and representative EEG traces (G) of EEG recording of mice with activated BF GABASOM neurons in isoflurane anesthesia, n = 8, **p < 0.01, ***p < 0.001, two-way ANOVA. H, The duration time of propofol anesthesia in BF GABAParv neuron with hM3Dq group, n = 8, n.s., no significant, paired t-test. I, J, The LORR (I) and RORR (J) time of isoflurane in BF GABAParv neuron with hM3Dq group, n = 8, **p < 0.01, n.s., no significant, paired t-test. K-N, The power spectrum analysis of EEG recording of mice with activated BF GABASOM neurons in propofol (K, L) and isoflurane (M, N) anesthesia, n = 8, two-way ANOVA [Colour figure can be viewed at wileyonlinelibrary.com] F I G U R E 5 A, The injection sites of AAV-Ef1α-DIO-hM4Di-mcherry or AAV-Ef1α-DIO-mCherry vector in BF of SOM-IRESCre mice. B, Immunofluorescent of BF GABASOM neurons (green) and hM4Di virus expression (red), and the merged picture. Scale bar: 100 μm. C, D, C-Fos expression of BF in GABASOM-hM4Di mice after saline or CNO i.p. injection. Scale bar: 100 μm. E, The percent of C-Fos positive cells in BF GABASOM nucleus. Data are presented as mean ± SEM. The percent of C-Fos expression in GABASOM neurons with CNO pretreatment was significantly lower than that in the saline pretreatment group (p < 0.005, unpaired t-test). F, The injection sites of AAV-Ef1α-DIO-hM4Di-mcherry or AAV-Ef1α-DIO-mCherry vector in BF of PV-IRESCre mice. G, Imnunoflurescent of GABAParv neurons (green) and hM4Di virus expression (red), and the merged picture. H, I C-Fos expression of BF in GABAParv-hM4Di mice after saline or CNO i.p. injection 2 h before perfusion. Scale bar: 100 μm. J, The percent of C-Fos positive cells in BF GABAParv nucleus. Data are presented as mean ± SEM. The percent of C-Fos expression in GABAParv neurons with CNO pretreatment was significantly lower than that in the saline pretreatment group (p < 0.005, unpaired t-test) [Colour figure can be viewed at wileyonlinelibrary.com] two-way ANOVA). There were no significant behavioral differences in the GABA Parv neurons inactivated mice with propofol and isoflurane anesthesia ( Figure 6H-J). However, a few differences were found in the delta power bands of the cortical EEG ( Figure 6K-N).
These results illustrate that GABA SOM neurons in the BF promote propofol and isoflurane anesthesia, similar to the sleep-promoting function in natural sleep-wake cycle. However, the GABA Parv neurons did not appear to have obvious effect in anesthesia.

| DISCUSS ION
Our study aimed to clarify whether propofol and isoflurane induce unconsciousness mediated by BF GABA SOM and GABA Parv neu- laboratory to some extent. [26][27][28] Accordingly, we suspected that the increased sensitivity to general anesthetics in BF GABA SOM neurons lesioned mice appears to be due to the accumulation of sleep debt, just like the increased efficacy of general anesthetics after 24 days of lesions of the ventrolateral preoptic nucleus, which is a classic sleep-promoting brain area. 29  which is in accordance with our results that the manipulation of GABA Parv neurons did not alter the process of general anesthesia, suggesting that these neurons may be not critical in arousal and/or wakefulness maintaining. Nevertheless, a more significant influence might exist if we manipulated the circuits that BF Parv neurons involved in, as they have some connections to other important areas, like thalamic reticular nucleus (TRN), cingulate cortex, neocortex, etc. [36][37][38] Furthermore, the technical limitations might cause some influences, such as the incompletely ablation of the target neurons might induce the compensatory effect and caused the results seen in the lesioned groups. In addition, we recorded EEG without EMG of mice synchronically and did not analyze the NREM and REM sleep separately under anesthesia, which might neglect some changes when manipulate these neurons.
Numerous evidences indicated that GABA SOM neurons contain diverse regional populations and functions in the BF. GABA SOM neurons in VP of BF regulate local gamma oscillations to drive prefrontal cortical activity. 39,40 In sleep-wake cycle, BF GABA SOM neurons exert a sleep promoting effect, 10 and the synthetic SOM analog, octreotide, can either suppress NREM sleep or increase REM sleep. 41,42 Furthermore, they can also promote anxiety emotion and the intake of high-calorie food. 43 The projections from the amygdala to different BF (VP/SI) subregions share certain organizational features with prefrontal cortical, and there are hippocampal projections to the medial septum and SI areas in the BF as well. 34,44,45 In our study, we did not separate these subregions in detail. Rather, according to the slice immunofluorescence, the region we injected the virus in were mostly the horizontal limb of the diagonal band (HDB), magnocellular preoptic nucleus (MCPO), and VP. We found that the SOM neurons in these subregions were involved in anesthesia induced by isoflurane and propofol, and activation of these neurons increased the sensitivity to anesthetics, prolonged anesthesia, and synchronized cortical EEG.
Future study should divide these neurons into different subgroups according to different subregions, such as the wake-active Kv2.2expressing GABAergic neurons. 46,47 Additionally, some experiments indicated that the GABAergeic system could be affected by anesthetics, and this effect is related to age. 48,49 Therefore, in our study, we only included young mice (8-12 weeks) to avoid any possible effects caused by aging.
In addition, the BF not only connects with other brain regions but also contains local neuronal connectivity. The sleep-promoting GABA SOM neurons inhibit other three types of neurons in the sleepwake cycle. 10 The increased effect of anesthetics may partly be caused by inhibition of other active neurons in the BF and relevant regions when BF GABA SOM neurons are activated. Although BF Parv neurons have some connections to other important areas, they appeared to have little effect on the narcotism induced by anesthetics.
Thus, our results suggest that the unconsciousness induced by general anesthetics is principally achieved by acting on a specific neural network that is involved in consciousness maintenance.