Dopamine Receptor Type 2-Expressing Medium Spiny Neurons in the Ventral Lateral Striatum Have a Non-REM Sleep-Induce Function

Abstract Dopamine receptor type 2-expressing medium spiny neurons (D2-MSNs) in the medial part of the ventral striatum (VS) induce non-REM (NREM) sleep from the wake state in animals. However, it is unclear whether D2-MSNs in the lateral part of the VS (VLS), which is anatomically and functionally different from the medial part of the VS, contribute to sleep-wake regulation. This study aims to clarify whether and how D2-MSNs in the VLS are involved in sleep-wake regulation. Our study found that specifically removing D2-MSNs in the VLS led to an increase in wakefulness time in mice during the dark phase using a diphtheria toxin-mediated cell ablation/dysfunction technique. D2-MSN ablation throughout the VS further increased dark phase wakefulness time. These findings suggest that VLS D2-MSNs may induce sleep during the dark phase with the medial part of the VS. Next, our fiber photometric recordings revealed that the population intracellular calcium (Ca2+) signal in the VLS D2-MSNs increased during the transition from wake to NREM sleep. The mean Ca2+ signal level of VLS D2-MSNs was higher during NREM and REM sleep than during the wake state, supporting their sleep-inducing role. Finally, optogenetic activation of the VLS D2-MSNs during the wake state always induced NREM sleep, demonstrating the causality of VLS D2-MSNs activity with sleep induction. Additionally, activation of the VLS D1-MSNs, counterparts of D2-MSNs, always induced wake from NREM sleep, indicating a wake-promoting role. In conclusion, VLS D2-MSNs could have an NREM sleep-inducing function in coordination with those in the medial VS.

The sleep-inducing function of dopamine receptor type 2-expressing medium spiny neurons (D2-MSNs) in the medial part of the ventral striatum (VS) has been previously reported; however, their function in the lateral part of the VS (VLS) has not been elucidated.We demonstrated that the diphtheria toxin-induced ablation of D2-MSNs in the VLS, as well as in the entire VS, increased wakefulness time in mice during the dark phase.VLS D2-MSNs had higher average Ca 21 signals during non-REM (NREM) and REM sleep than wake state via fiber photometric recording.Furthermore, optogenetic activation of VLS D2-MSNs during wake state induced NREM sleep in mice.In conclusion, D2-MSNs in the VLS have an NREM sleep-inducing function in coordination with those in the medial VS.

Introduction
The regulation of sleep-wake behavior involves various neural populations, including the medium spiny neurons that express dopamine receptor type 2 (D2-MSNs) in the striatum (Oishi and Lazarus, 2017).D2-MSNs express receptors for adenosine, one of the sleep-promoting substances (Huang et al., 2014).These neurons are uniformly distributed throughout the striatum (Kemp and Powell, 1971); however, its functional manifestation is divided between the dorsal striatum and ventral striatum (DS and VS), and further varies depending on the subregion within the DS/VS, such as during reward processing (Cox and Witten, 2019).Thus, the sleep-wake regulatory function of these neurons could also be exhibited differently in each striatal subregion.
In the DS, a region primarily involved in motor control in animals, D2-MSNs contribute to non-REM (NREM) sleep induction in the rostral, centromedial, and centrolateral subregions, but not in the caudal subregion, and their sleep-inducing role is exhibited only in the dark phase (Yuan et al., 2017).In contrast, in the VS, it is not clear in all subregions whether the D2-MSNs are involved in sleep-wake control.The VS can be anatomically and functionally divided into at least three subregions: the nucleus accumbens (NAc) medial shell, referred to as the ventral medial striatum (VMS), NAc core, and ventral lateral striatum (VLS; also known as NAc lateral shell).Previous studies have reported that D2-MSNs in the NAc core subregion contribute to NREM sleep induction in mice (Oishi et al., 2017;Luo et al., 2018).Another study has suggested that these neurons in the NAc medial shell have the same function (Satoh et al., 1999), and mediates caffeine-induced arousal via the adenosine A 2A receptor which co-express with dopamine D2 receptor (Lazarus et al., 2011).However, it is unclear whether D2-MSNs in the VLS subregion play a role in regulating sleep-wake cycles in animals.The VLS D2-MSNs receive distinctive input from cortical glutamatergic neurons in the insular cortex and dopaminergic neurons in the ventral tegmental area (VTA; Hunnicutt et al., 2016;Mingote et al., 2019), and specifically constitute neuronal circuitry with parvalbumin-expressing interneurons (Yoshida et al., 2020).This could result in distinct firing patterns and functions in reward processing compared with those in the medial part of the VS (Tsutsui-Kimura et al., 2017a;Yang et al., 2018;R. Chen et al., 2021).This anatomic and functional uniqueness of the VLS predicts a distinct role for VLS D2-MSNs in sleep-wake control.
To elucidate the role of VLS D2-MSNs in sleep-wake regulation, we evaluated the sleep-wake architecture using a 24 h polysomnogram in mice, in which the D2-MSNs were spatiotemporally progressively ablated from the specific VLS to the entire VS based on DOX-dependent diphtheria toxin induction in these neurons (D2-DTA; Tsutsui- Kimura et al., 2017b).Next, we investigated the Ca 21 signal patterns of the D2-MSNs in the VLS during physiological sleep-wake states in mice.Furthermore, we conducted optogenetic activation of the VLS D2-MSNs to reveal a causal relationship between the activity of these neurons and sleep-wake regulation.Our findings provide evidence that D2-MSNs in the VLS are involved in sleep induction in animals, in cooperation with those in the medial part of the VS.

Ethics statement
All animal procedures were conducted following the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Keio University Animal Experiment Committee in compliance with the Keio University Institutional Animal Care and Use Committee (approval numbers A2022-315).

Surgical procedure
Surgeries were performed using a stereotaxic apparatus (SM-6 M-HT and SM-15R/L, Narishige).Mice were anesthetized with a mixture of ketamine and xylazine (100 and 10 mg/kg, respectively).Body temperature during surgery was maintained at 37 6 0.5°C using a heating pad (FHC-MO, Muromachi Kikai).Mice received permanent electroencephalography (EEG) and electromyography (EMG) electrodes for polysomnography.Three pits were drilled into the skull using a carbide cutter (drill diameter: 0.8 mm).We implanted two electrodes in each subject, each consisting of a 1.0 mm diameter stainless steel screw that served as an EEG electrode.One implant was placed over the right frontal cortical area (AP: 11.0 mm; ML: 11.5 mm) as a reference electrode, while the other was placed over the right parietal area (AP: 11.0 mm anterior to l ; ML: 11.5 mm) as a signal electrode.Another electrode was placed over the right cerebellar cortex (AP: À1.0 mm posterior to l ; ML: 11.5 mm) as the ground electrode.Two silver wires (AS633; Cooner Wire Company) were placed bilaterally into the trapezius muscles and served as the EMG electrodes.An optic fiber was inserted into the VLS at the following coordinates relative to bregma: AP, 11.1 mm; ML, 11.8 mm; DV, 14.4 mm from the skull unilaterally for photometry, AP, 11.1 mm; ML, 61.9 mm; DV, 13.1 mm from the brain surface bilaterally for optogenetic manipulation.Finally, the electrode and optical fiber cannula assembly were anchored and fixed to the skull using SuperBond (Sun Medical Co).

EEG/EMG recordings
The EEG/EMG signals were amplified (gain 1000Â) and filtered (EEG: 1-300 Hz, EMG: 10-300 Hz) using a DC/AC differential amplifier (AM-3000, AM Systems).The input was received via an input module (NI-9215, National Instruments), digitized at a sampling rate of 1000 Hz using a data acquisition module (cDAQ-9174, National Instruments), and recorded using a custom-made LabVIEW program (National Instruments).We habituated the 24 h EEG/EMG recordings more than three times, and REM sleep (see the vigilance state assessment) was often observed when we started the experiment.Three days before the control recordings on DOX-off day À1 (DOX-off day À4), the mice were placed in a soundproof box and connected to the EEG/EMG cable for the 24 h EEG/EMG recording of D2-DTA mice.They remained continuously tethered to the cable in the same soundproof box until day 11 of the DOX-off experiment.Cages and food were not changed during the stay in the soundproof box, except changing from DOX-containing chow to normal chow on Dox-off day 0.

Fiber photometry
The ratiometric fiber photometry method has been described previously (Natsubori et al., 2017;Kato et al., 2022).An excitation light (435 nm; silver light-emitting diode, Prizmatix) was reflected off a dichroic mirror (DM455CFP, Olympus), focused with a 2Â objective lens (numerical aperture 0.39, Olympus) and coupled into an optical fiber (400-mm diameter, 0.39 numerical aperture; catalog #M79L01, Thorlabs) through a pinhole (diameter 400 mm).The light-emitting diode power was ,200 mW at the fiber tip.The cyan and yellow fluorescence emitted by the YC-nano50 was collected via an optical fiber cannula, divided by a dichroic mirror (DM515YFP, Olympus) into cyan (483/ 32 nm band path filters, Semrock) and yellow (542/27 nm), and detected using a photomultiplier tube (H10722-210, Hamamatsu Photonics).The fluorescence signals were digitized using a data acquisition module (cDAQ-9174, National Instruments) and recorded simultaneously using a custom-made LabVIEW program (National Instruments).Signals were collected at a sampling frequency of 1000 Hz.

Optogenetic manipulation
An optical fiber (numerical aperture 0.39, Thorlabs) was inserted through the guide cannula.Blue (470 nm) and yellow (575 nm) light were generated using a SPECTRA 2-LCR-XA light engine (Lumencor).The blue and yellow light power intensities at the tip of the optical fiber were 1-2 and 3-4 mW, respectively.Using EEG and EMG monitoring, we illuminated 1 s of blue light to open the stepfunction-type opsin ChR2(C128S; Berndt et al., 2009) of D2-ChR2 or D1-ChR2(C128S) mice during the wake or NREM state lasting .10s, respectively.In control trials, yellow light was used instead of blue light.

Mouse vigilance state assessment
EEG/EMG signals were analyzed using MATLAB (MathWorks).Power spectral EEG data were obtained using multitaper spectral estimation (McCoy et al., 1998).A power spectral profile over a 1-to 50-Hz window was used for the analysis.We detected each sleep-wake state scored offline by visually characterizing 10-s epochs for 24-h polysomnography of D2-DTA mice and 1-s epochs for fiber photometric recording of D2-YC mice and optogenetic experiments with D2-ChR2 and D1-ChR2 mice (Funato et al., 2016;Kato et al., 2022).The state determination criteria were as follows: the wake state was characterized by a low-amplitude fast EEG and high-amplitude variable EMG.NREM sleep is characterized by a high-amplitude d (1-4 Hz) frequency EEG and low-amplitude tonus EMG.REM sleep was staged using u (6-9 Hz) dominant EEG and electromyography (EMG) atonia (Tobler et al., 1997).

Colorimetric in situ hybridization (ISH)
Mice were deeply anesthetized with ketamine (100 mg/ kg) and xylazine (10 mg/kg) and perfused with a 4% paraformaldehyde phosphate buffer solution.After removing the brains from the skull, they were fixed overnight in the same solution.The brains were then cryoprotected in 20% sucrose overnight, frozen, and sectioned to 25-mm thickness on a cryostat.The sections were mounted on silane-coated glass slides (Matsunami Glass), and then treated with proteinase K (40 mg ml À1 ; Merck).After washing and acetylation, sections were incubated with digoxigenin (DIG)-labeled complementary RNA (cRNA) probes.After the sections were washed in buffers with serial differences in stringency, they were incubated with an alkaline phosphatase-conjugated anti-DIG antibody (11093274910, 1:5000; Roche).The cRNA probes were visualized using freshly prepared colorimetric substrate (NBT/BCIP; Roche).Nuclear Fast Red (Vector Laboratories) was used for counterstaining (Tanaka et al., 2012).Probes for DTA, Drd1, and Drd2 were used in this study.Images were obtained using an all-in-one microscope (BZ-X710; Keyence).

Data processing
All animals and trials were randomly assigned to experimental conditions.The experimenters were not blinded to the experimental conditions during data collection and analysis.Mice were excluded when the optical fiber position was incorrectly targeted.We used custom-made programs in MATLAB for signal processing.

Fluorescence signal analysis
For fiber photometric signal analysis, the yellow and cyan fluorescence signals were detrended.Further, we used the YC ratio (R), which is the ratio of yellow to cyan fluorescence intensity, calculated as the population Ca 21 signals, and then Z-scored.For the power spectral analysis, fiber photometric signals (YC signals) were extracted during the uninterrupted wake, NREM sleep, and REM sleep bouts that lasted .100s, high pass filtered with a 0.01-Hz cutoff frequency, and calculated with a wavelet transform using a Morse wavelet from 0.01 to 1 Hz with time-bandwidth parameter (P 2 ¼ 10).The peak frequency is determined as the local maximum of the power spectrum.
For the state transition analysis, YC signals were extracted during uninterrupted wake, NREM sleep, and REM sleep bouts that lasted .20 s, and normalized using the mean and SD of YC signals from 20 s before the transition to the transition point.

Statistical analysis
All experimental data were analyzed using the following parametric statistics: Student's t test (independent samples t test), paired t test, paired t test with Bonferroni correction, and repeated-measures ANOVA followed by the Tukey-Kramer post hoc test.

Data availability
The datasets generated and/or analyzed in the current study are available from the corresponding author on reasonable request.

Ablation of D2-MSNs in the VLS causes an increase in wakefulness time in the dark phase
To investigate the role of D2-MSNs in the sleep-wake regulation within the ventral lateral striatum (VLS), we conducted repeated 24-h polysomnography on Drd2-tTA:: tetO-DTA mice (D2-DTA; Tsutsui-Kimura et al., 2017b).In this mouse line, tTA-mediated diphtheria toxin expression is induced in D2-MSNs, which is initiated specifically in the VLS and follows spatial progress within the ventral striatum (VS) in doxycycline (DOX)-dependent manner (Fig. 1a): upon switching to normal chow from DOX-containing chow (DOX-off day 0), DTA induction is first observed only in the VLS (corresponding to the NAc lateral shell) on DOX-off day 3, and the DTA mRNA-positive cells were increased and the induction area gradually expands toward the medial part of the VS (corresponding to the   1).c, Time spent in each sleep state during the light and dark phases in control (DOX-off day À1) and DOX-off mice on days 5, 7, and 10 (n ¼ 5 mice).Repeated-measures ANOVA was followed by the Tukey-Kramer post hoc test (Results of statistical tests is shown in Table 2).d, Daily variations in wake time (left), NREM sleep (middle), REM sleep (right) every hour in control (DOX-off day À1), and DOX-off day 10 (n ¼ 5 mice).Twoway repeated-measures ANOVA was followed by the Tukey-Kramer post hoc test (Results of statistical tests is shown in Table 3).e, Episode numbers of wake, NREM sleep, and REM sleep for 24 h, light, and dark phases in control (DOX-off day À1) and DOX-off days 5, 7, and 10 (n ¼ 5 mice NAc medial shell and core) with each passing day (Tsutsui-Kimura et al., 2017b).Furthermore, based on the altered electrophysiological functions in viable DTA-expressing VLS D2-MSNs (Tsutsui-Kimura et al., 2017b), the VLS D2-MSNs could be regarded as a progressive hypofunction condition on DOX-off day 5 and day 7, respectively (Fig. 1b).On DOX-off day 10, DTA mRNA was expressed in the entire VS with loss of Drd2 mRNA expression in the VLS, indicating that the D2-MSNs in the entire VS could be in a hypofunctional state with VLS D2-MSNs dysfunction (Fig. 1c,d).
We performed 24-h polysomnography recordings to evaluate the effect of progressive ablation of the VS D2-MSNs in the D2-DTA mice on the sleep-wake architecture (Fig. 2a).The recordings were done on DOX-off day À1 (control), DOX-off days 5 and 7 (functional ablation of VLS D2-MSNs), and DOX-off day 10 (ablation of D2-MSNs in the entire VS) to assess the impact of these ablations (Fig. 1b).We observed that the daily amount of wakefulness time significantly increased on DOX-off day 10, alongside a decrease in NREM sleep time, compared with those on DOX-off day À1 (*p , 0.05, Repeated-measures ANOVA followed by the Tukey-Kramer post hoc test; Fig. 2b).
However, the daily sleep-wake time did not significantly change on DOX-off days 5 and 7. Next, we separately analyzed the effect of DOX-dependent D2-MSNs ablation on sleep-wake architecture in the light and dark phases.In the light phase, there were no significant differences in the percentage of time spent in the wake, in NREM sleep, or REM sleep (Fig. 2c).During the dark phase, we observed a significant increase in the percentage of time spent in the wake on DOX-off day 5, which continued to increase up to DOX-off day 10, whereas those in the NREM and REM sleep decreased on the DOX-off day 10 (*p , 0.05, repeated-measures ANOVA followed by the Tukey-Kramer post hoc test; Fig. 2c).This phenotype of increased wake time and decreased sleep time because of D2-MSNs ablation was most pronounced in the first half and end of the dark phase (*p , 0.05, two-way repeated-measures ANOVA was followed by the Tukey-Kramer post hoc test; Fig. 2d).
To determine the characteristics of sleep-wake alteration induced by D2-MSNs ablation, we calculated the number of episodes of wake, NREM sleep, and REM sleep.Although the mean duration of the sleep/wake bout showed a gradual increasing trend in the dark phase continued mean duration during each sleep-wake state in control (DOX-off day À1) and DOX-off day 10 (n ¼ 5 mice).Two-way repeated-measures ANOVA was followed by the Tukey-Kramer post hoc test).Results of statistical tests is shown in Table 5. Error bars indicate SEM; *p , 0.05.2f).The mean duration of the sleep/wake bout showed a gradual increasing trend in the dark phase under progressive D2-MSNs ablation, although this was not statistically significant (*p , 0.05, Repeated-measures ANOVA followed by the Tukey-Kramer post hoc test; Fig. 2g, twoway repeated-measures ANOVA was followed by the Tukey-Kramer post hoc test; Fig. 2h).
Next, we investigated the EEG changes in sleep or wakefulness that were altered by VS D2-MSN ablation.When we compared the EEG spectrum during each sleep-wake state throughout the day and in the light/ dark phases between the control (DOX-off day À1) and DOX-off day 10, there were no significant differences in any of the power bands of the EEG signal (d, u, s, b, and g ; *p , 0.05, paired t test; d and u ranges are shown in Fig. 3).VLS D2-MSNs display higher Ca 21 signal levels during NREM and REM sleep compared with wake Our ablation study revealed that D2-MSNs in the VLS may be involved in sleep-wake regulation.Thus, we monitored the compound intracellular calcium (Ca 21 ) signal patterns of the VLS D2-MSNs to determine their activity patterns across sleep-wake states and during sleep/wake stage transitions in mice.We used a fiber photometry   system to monitor intracellular Ca 21 signals from the VLS D2-MSNs in freely moving mice (Fig. 4a,b).We used transgenic mice expressing the FRET-based ratiometric Ca 21 indicator YC-nano50 (Horikawa et al., 2010), in D2-MSNs under the control of the Drd2 promoter (Drd2-tTA:: tetO-YCnano50: D2-YC mice; Natsubori et al., 2017).The ratio of yellow to cyan fluorescence intensity (YC ratio) represented the intracellular Ca 21 signal of D2-MSNs.We found that the VLS D2-MSNs displayed infra-slow periodic Ca 21 signal fluctuations during wake, NREM sleep, and REM sleep in both the light and the dark phases (Fig. 4c).The peak frequency of these events did not show a significant variation between the different sleep-wake states both in the light and the dark phases (Fig. 4d).Next, we investigated the temporal relationship between Ca 21 signal fluctuations in the VLS D2-MSNs and EEG/EMG activity during each sleep-wake state.When we calculated the correlation coefficient between the Ca 21 signal and the EEG or EMG power, the Ca 21 signal in the VLS D2-MSNs correlation with EEG g power (correlation coefficient: 0.16 6 0.02 and 0.17 6 0.03 in the light and dark phases, respectively; mean 6 SEM) and EMG power during the wake (correlation coefficient: 0.16 6 0.01 and 0.12 6 0.04, respectively; mean 6 SEM), EEG b (correlation coefficient: 0.17 6 0.02 and 0.15 6 0.02, respectively; mean 6 SEM) and g power (0.16 6 0.02 and 0.12 6 0.03, respectively; mean 6 SEM) during the NREM sleep, and EEG g power during the REM sleep (0.15 6 0.02 and 0.17 6 0.03, respectively; mean 6 SEM; Fig. 4e).Using cross-correlation analysis, we found that the Ca 21 signal in VLS D2-MSNs preceded the peak of EEG high-frequency or EMG power during both wake (lag time between peaks, EEG g power -Ca 21 signal: À0.83 6 0.10 and À0.51 6 0.04 s in the light and dark phases, respectively, EMG power -Ca 21 signal: À0.82 6 0.15 and À1.35 6 0.06 s, respectively; mean 6 SEM) and during the NREM sleep (lag time between peaks, EEG b power -Ca 21 signal: À0.31 6 0.46 and À1.77 6 0.39 s, respectively, EEG g power -Ca 21 signal: À0.77 6 0.18 and À1.81 6 0.71 s, respectively; mean 6 SEM; Fig. 4f).However, no such relationship was observed between the Ca 21 signal and EEG g power during REM sleep, likely because g power fluctuations are minimal during this state (Fig. 4f).
Furthermore, we focused on the averaged Ca 21 signal levels during each sleep-wake state, which was significantly higher during NREM and REM sleep than during wakefulness in the light phase (Fig. 4g).In contrast, in the dark phase, averaged Ca 21 signal level during REM sleep were higher than those during NREM sleep and wake states (Fig. 4g).Thus, we evaluated changes in the average Ca 21 signal level during the state transition (Fig. 4h).During the transition from NREM sleep to wake, the Ca 21 signal began to decrease in the light phase and showed a lesser tendency to decrease in the dark phase.During the transition from wake to NREM sleep, the Ca 21 signal began to increase in both the light and the dark phases.The Ca 21 signal did not change during the transition from NREM to REM sleep, and it began to decrease during the transition from REM sleep to wake in both the light and the dark phases (Fig. 4h).These Ca 21 signal pattern variations in the VLS D2-MSNs, particularly during NREM sleep induction, support the sleep-inducing function of these neurons, as indicated by our ablation experiment.

Optogenetic activation of the VLS D2-MSNs during wake causes NREM sleep
Our ablation study and Ca 21 signal recordings suggest that VLS D2-MSNs have an NREM sleep-inducing function.and artificially activated their VLS D2-MSNs (Fig. 5a).In this variant, short-term light induction cause to initiate photocurrent, and it terminated ;100 s after light induction because the "tau off" of C128S mutant ion channel is ;106 6 9 s (Berndt et al., 2009).The D2-ChR2(C128S) mice received 1 s of blue light illumination to open the ChR2(C128S) in the bilateral VLS D2-MSNs or 1 s of yellow light as a control (Fig. 5a).
Optogenetic activation of VLS D2-MSNs during the wake state in mice always prompted changes in EEG/ EMG activity and induced NREM sleep within 100 s after light illumination during both the light and dark phases; however, control light (yellow light) did not always induce these changes (Fig. 5b,c).In contrast, optogenetic activation of the VLS D2-MSNs during NREM sleep resulted in no changes in sleep-wake state of animals (Fig. 5d).Notably, REM sleep did not occur during the first 100 s after the onset of photostimulation, whether in NREM sleep or wake state.
The NREM sleep-induction latency by the VLS D2-MSNs photoactivation during the wake state was 12.3 6 2.0 and 10.2 6 1.9 s in the light and the dark phases, respectively, which was significantly shorter than those of the control light illumination [mean 6 SEM, p ¼ 6.9 Â 10 À5 and p ¼ 1.5 Â 10 À3 (light and dark phases, respectively), independent t test; Fig. 5e].The percentages of NREM sleep time during 100 s after the start of VLS D2-MSNs photoactivation were 87.6 6 1.9% and 88.1 6 2.4% in the light and the dark phases, respectively, which were significantly higher than that of control light illumination [mean 6 SEM, p ¼ 1.2 Â 10 À4 and p ¼ 2.4 -Â 10 À4 (light and dark phases, respectively), independent t test; Fig. 5f].Furthermore, EEG d and u powers were at the same level between natural NREM sleep and the optogenetically induced NREM sleep during both the light and dark phases (Fig. 5g).These results suggest that the VLS D2-MSNs activity induced NREM sleep in mice.

Optogenetic activation of the VLS D1-MSNs induces wakefulness from NREM sleep
In the striatum, another population of MSNs, expressing dopamine receptor type 1 (D1-MSNs), is also distributed.While D2-MSNs and D1-MSNs in the dorsal striatum have opposing roles in motor control and reinforcement learning, in the VLS they have been reported to have cooperative roles in goal-directed behavior in mice (Kravitz et al., 2010;Natsubori et al., 2017).Therefore, we tried optogenetic activation of the D1-MSNs in the VLS to compare the sleep-wake regulation functions of VLS D2-MSNs and D1-MSNs.We used D1-ChR2 mice (Pde10a2-tTA::tetO-ChR2(C128S)-EYFP; Adora2a-Cre triple-transgenic mice), which harboring ChR2(C128s) only in the D1-MSNs (Fig. 6a).As a result, optogenetic activation of the VLS D1-MSNs during wake state did not cause any state changes (Fig. 6b).However, activation of the VLS D1-MSNs during NREM sleep always promptly induced wake in mice during the light and dark phases, whereas control light (yellow light) illumination did not always induce it (Fig. 6c-f).The wake-induction latency by the VLS D1-MSNs photoactivation was 0.15 6 0.02 and 0 s in the light and the dark phases, respectively [mean 6 SEM, p ¼ 0.01 and p ¼ 0.03 (light and dark phases, respectively), independent t test; Fig. 6e].The percentages of wake time during 100 s after the start of VLS D1-MSNs photoactivation were 96.7 6 1.3% and 100% in the light and the dark phases, respectively [mean 6 SEM, p ¼ 6.3 Â 10 À6 and p ¼ 2.4 Â 10 À5 (light and dark phases, respectively), independent t test; Fig. 6f].These results indicates that VLS D1-MSNs have a wake-inducing function and that VLS D2-MSNs and D1-MSNs encode opposing functions in sleep-wake regulation.

Discussion
This study demonstrated that ablation of D2-MSN in the VLS causes an increase in the amount of wake time during the dark phase, accompanied by a decrease in sleep time.This effect becomes more pronounced as the ablation area expands to include the entire VS.Next, our fiber photometric recording revealed that the average intracellular Ca 21 Our ablation and optogenetic activation studies strongly suggest that D2-MSNs in the VLS have an NREM sleep-inducing function (Figs. 2 and 5), and the fiber photometric recording supports these results (Fig. 4).The VLS is distinguished from the medial parts of the VS, such as the NAc medial shell and core, based on cortical input patterns (Berendse et al., 1992;Hunnicutt et al., 2016).Several studies have demonstrated the functional specificities of D2-MSNs in the VLS, such as reward processing (Tsutsui-Kimura et al., 2017a;Yang et al., 2018;R. Chen et al., 2021).However, our findings regarding the sleep-inducing function of D2-MSNs in the VLS are consistent with those found in the medial portion of the VS (NAc medial shell/ core; Satoh et al., 1999;Lazarus et al., 2011;Oishi et al., 2017;Luo et al., 2018).Furthermore, in the DTA-induced ablation experiment, we observed that the alteration of sleep/wake architecture was even more pronounced under the ablation of the D2-MSNs in the entire VS, compared with that of in the VLS (Fig. 2b,c).Therefore, the alterations in the sleep/wake pattern might be strengthened by the ablation of D2-MSNs in both the VLS and the medial part of the VS, indicating a coordinated and consistent regulation of the animal's sleep-wake states by the D2-MSNs in these subregions of the VS.The degree of the D2-MSNs effect on the sleep/wake architecture may depend on the extent of its distribution area.D2-MSNs in the VS project to the VP (Basar et al., 2010), of which the projection from the NAc core to the VP has been reported to be involved in sleepwake regulation (Oishi et al., 2017;Luo et al., 2018).These consistent efferent patterns could be associated with the common sleep-wake regulatory role of D2-MSNs in the medial and lateral VS.
We demonstrated that the optogenetic activation of the VLS D2-MSNs induced NREM sleep from wakefulness whereas that of the VLS D1-MSNs induced wake from NREM sleep in mice (Figs. 5 and 6).These findings strongly suggest that the VLS D1-MSNs and D2-MSNs activities themselves change the sleep-wake state of animals in the opposite direction.The reversal sleep-wake regulation function of the VLS D1-MSNs and D2-MSNs could be related to the differential dopaminergic control and/or output patterns of these two neuronal subpopulations.First, striatal D1-MSNs and D2-MSNs activities are inversely affected by the extracellular dopamine: as the D1 and D2 receptors couple with G s and G i proteins, respectively (Kebabian and Calne, 1979), the dopamine signal has excitatory and inhibitory effects on the activity of D1-MSNs and D2-MSNs, respectively.Thus, dopaminergic control often causes the opposing function of striatal D1-MSNs and D2-MSNs, and their balance is expressed in functional outputs such as motor controls (Cox and Witten, 2019).It could be the same for the sleep-wake regulation function of the VLS D1-MSNs and D2-MSNs.Next, the partially differential output patterns of the VS D1-MSNs and D2-MSNs might cause their reversal of sleep-wake regulation functions.D2-MSNs in the VS consistently project to the VP, while the D1-MSNs in the NAc project to not only the VP but also the ventral mesencephalon (VM; Kupchik et al., 2015;Liu et al., 2022).D1-MSNs in the NAc core have a wake-promoting role via the VM and lateral hypothalamus (Luo et al., 2018) and artificial inhibition of GABAergic neurons in the VM increases wakefulness time (Takata et al., 2018), whereas the VS D2-MSNs projection to the VP could involve NREM sleep induction (Oishi et al., 2017;Luo et al., 2018).These findings suggest that the VLS D2-MSNs activation could induce NREM sleep only through the VP whereas the VLS D1-MSNs activation-induced wake could be mediated by the inhibition of VM neurons.
NREM-sleep inducing function of the D2-MSNs and opposing wake inducing function of the D1-MSNs that we demonstrated in the VLS are consistent with D1-/D2-MSNs functions in the NAc core region (Oishi et al., 2017;Luo et al., 2018).The common reciprocal sleep-wake regulation function of D1-MSNs and D2-MSNs across the VS subregion may aid in the identification of sleep-wake regulation neuronal circuits centered on the VS D1-MSNs/ As previously mentioned, since the dopaminergic input from the SNr to the entire VS has a reverse effect on D1-MSNs and D2-MSNs (Scofield et al., 2016), it may be involved in the sleep-wake regulation function.Meanwhile, the glutamatergic inputs from the distinct cortical area to each VS subregion (Berendse et al., 1992;Hunnicutt et al., 2016) may have consistent effects on the MSNs if they participate in sleep-wake regulatory mechanisms.In turn, the common output of VS D1-MSNs and D2-MSNs to the VM and VP, respectively, could be engaged for sleep-wake regulation (Basar et al., 2010;Kupchik et al., 2015;Liu et al., 2022).Additionally, the latency to sleep/wake transition differs between D1 or D2-MSN stimulation (Figs. 5 and 6).
Even in the natural sleep-wake cycles of animals, the transition from sleep to wakefulness occurs rapidly, whereas the transition from wakefulness to sleep is more gradual (Saper et al., 2010).Differences in the effects of each sleepwake inducing neuronal activity on the entire cortex may be responsible for the differences in sleep/wake transition latency.
Our fiber photometric recordings revealed that the compound intracellular Ca 21 signal in the VLS D2-MSNs showed the sleep-wake state-dependent variation and the mean signal level was elevated during the wake-to-NREM sleep transition both in the light and dark phases (Fig. 4).These findings support our conclusion of the sleep-inducing function of the VLS D2-MSNs derived from our ablation and optogenetic experiments (Figs. 2 and 5).The VLS D2-MSNs activity can be affected by dopaminergic input from the ventral tegmental area (VTA; Mingote et al., 2019) and glutamatergic input from the insular cortex (Berendse et al., 1992;Hunnicutt et al., 2016).The activity of VTA dopaminergic neurons varies depending on the sleep-wake state, which is higher during wakefulness and REM sleep and lower during NREM sleep (Eban-Rothschild et al., 2016).Therefore, VTA dopaminergic inputs, which could suppress D2-MSNs activity via the Gi-coupled D2 receptor, could cause the VLS D2-MSNs activity patterns to be lower during wakefulness and higher during NREM sleep.However, the higher activity of VLS D2-MSNs during REM sleep, particularly in the dark phase (Fig. 4g), cannot be explained by variations in the dopaminergic input.The increase in the Ca 21 signal level of the VLS D2-MSNs during REM sleep may be influenced by glutamatergic neuronal input from the insular cortex.A previous study showed that lesions in the insular cortex led to decreased wake time and increased NREM and REM sleep times (M.Chen et al., 2016).The fluctuation of neuronal activity in the insular cortex across sleep-wake states is currently unclear; however, given its potential involvement in sleep/wake regulation, it is not unexpected that glutamatergic neuronal projections from the insular cortex may impact the activity of D2-MSNs in the VLS across different sleep-wake states.In addition, these diverse state-dependent neuronal inputs might cause detailed differences in the VLS D2-MSNs Ca 21 signal fluctuations between each sleep-wake state in terms of their relationship with EEG/EMG activity in mice (Fig. 4e,f).Further studies are required to elucidate the sleep-inducing functions of the VLS D2-MSNs at the neuronal circuit level.
We observed that the change in sleep/wake architecture caused by VLS/VS D2-MSNs ablation occurred only during the dark phase (Fig. 2).Previous studies also reported that the inhibition D2-MSNs in the NAc core induced wake state in mice during the dark phase, although they did not mention during the light phase (Oishi et al., 2017;Luo et al., 2018).Meanwhile, Ca 21 signals in the VLS D2-MSNs increased during the transition from wake to NREM sleep in both light and dark phases (Fig. 4).These suggest that the VLS D2-MSNs increase their activity at the onset of NREM sleep in both the light and dark phases; however, this activity might exert its sleep-inducing function during the dark phase.Besides, we observed that optogenetic activation of VLS D2-MSNs and D1-MSNs promptly induced sleep-wake state change regardless of the light or dark phases (Figs. 5  and 6).These findings suggest that the firing activity of D1/D2-MSNs has a potential role in sleep-wake state change regardless of the light/dark phases, however, there might be a time-dependent gating mechanism for VLS D1/D2-MSNs output for sleep-wake regulation.
Our study has two limitations.First, the D2-MSN ablation method in D2-DTA mice (Tsutsui-Kimura et al., 2017b) cannot evaluate the sleep-wake regulatory function of D2-MSNs in each subregion of the VS separately, except for the VLS.Several prior studies demonstrate a causal relationship between the activity of the D2-MSNs in the medial part of the VS, particularly the NAc core, and NREM-sleep-inducing function, and these findings complement our result (Oishi et al., 2017;Luo et al., 2018).Next, it has been pointed out that fluorescent probes expressed in the living brain and detected by the fiber photometry system could be affected by cerebral blood volume (Ikoma et al., 2023a, b).The evaluation of YC signal fluctuations particularly during REM sleep should be cautious because cerebral blood flow is increased (Braun et al., 1997;Tsai et al., 2021;Ikoma et al., 2023b).The proposed new fiber photometry setup and analysis method to exclude the effects of cerebral blood flow (Ikoma et al., 2023a, b) could improve the accuracy of our neuronal activity assessment based on the YC signal measurement in the future.
In conclusion, our results demonstrate that D2-MSNs in the VLS are essential for the maintenance of sleep/wake architecture and have an NREM sleep-inducing function.

Figure 1 .
Figure 1.Spatiotemporally specific the VS D2-MSNs ablation.a, DOX-controllable DTA expression.Drd2-tTA::tetO-DTA (D2-DTA) mice were fed with doxycycline (DOX)-containing chow until the start of the experiment (DOX-on).DTA mRNA expression started when DOX-chow was replaced with normal chow (DOX-off).b, Time course of the VS D2-MSNs ablation, and timing of the polysomnographic recordings.c, Schematic illustration of the VS, including the VMS and the VLS.d, Drd1, Drd2, and DTA mRNA levels in the striatum on DOX-off day 10.Purple denotes the mRNA signal.Left, Drd1 mRNA expression throughout the striatum.Middle, Drd2 mRNA expression in the VLS disappears (yellow arrow).Right, DTA mRNA expression was observed throughout the VS (yellow arrow).

Figure 3 .
Figure 3.Comparison of EEG power spectrum before and after the VS D2-MSNs ablation.EEG spectra before (DOX-off day À1) and after (DOX-off day 10) the VS D2-MSNs ablation during each sleep-wake state (n ¼ 5 mice).Two-way repeated-measures ANOVA was followed by the Tukey-Kramer post hoc test).Results of statistical tests is shown in Table 6.Colored shades indicate SEM; *p , 0.05.

Figure 4 .
Figure 4. Population Ca 21 signals of the VLS D2-MSNs across the sleep-wake states.a, Schematic illustration of the fiber photometry system for monitoring intracellular Ca 21 signals of VLS D2-MSNs in D2-YC mice.The EEG and EMG were recorded simultaneously.PMT, photomultiplier tube.b, YCnano50 fluorescent expression in D2-YC mice (left) and a schematic of the fiber photometric recording site (right) in the VLS.The asterisks indicate the tip of the optical fiber.Scale bar, 1 mm.c, Representative examples of EEG signal trace and power spectrogram, EMG signal, and population Ca 21 signal trace and spectrogram of VLS D2-MSNs across

Figure 5 .
Figure 5. Sleep-wake state change by optogenetic activation of the VLS D2-MSNs.a, Schematic illustration of optogenetic activation of VLS D2-MSNs in Drd2-ChR2(C128S) mice (hereafter referred to as D2-ChR2 mice).Scale bar, 1 mm.The blue and yellow shades indicate the illumination times.b, Representative examples of EEG signal traces, power spectrograms, and EMG signals under the photostimulation of VLS D2-MSNs during the wake state in D2-ChR2 mice.Vertical blue and yellow shades indicate illumination times.The upper panel shows the optogenetic manipulation and the lower panel shows the control experiment.c, NREM sleep induction rate which indicates whether NREM sleep occurred within 100 s after light illumination.Upper column shows a light phase and lower column shows a dark phase.In each column, left side shows the activation group (blue light illumination) and right side shows control group [yellow light illumination; n ¼ 4 mice, 5-10 manipulations/recordings, a total of 49 trials for activation (light

Figure 6 .
Figure 6.Sleep-wake state change by optogenetic activation of the VLS D1-MSNs.a, Schematic illustration of optogenetic activation of VLS D1-MSNs in Pde10a2-tTA::tetO-ChR2(C128S)-EYFP; Adora2a-Cre triple-transgenic mice.Scale bar, 1 mm.The blue and yellow shades indicate the illumination times.b, Representative examples of EEG signal trace, EEG power spectrogram, and EMG signal before and after optogenetic manipulation during the wake.Vertical blue shade indicates illumination times.c, Representative examples of EEG signal traces, EEG power spectrograms, and EMG signals before and after optogenetic continued manipulation during NREM sleep.Vertical blue and yellow shades indicate illumination times.The upper panel shows the optogenetic manipulation and the lower panel shows the control experiment.d, Wake state induction rate which indicates whether the wake state occurred within 100 s after light illumination.Upper column shows light phase and lower column shows dark phase.In each column, left side shows the activation group (blue light illumination) and right side shows the control group [yellow light illumination; n ¼ 3 mice, 5-10 manipulations/recordings, a total of 23 trials for activation (light phase), 18 trials for control (light phase), a total of 18 trials for activation (dark phase), 20 trials for control (dark phase)].e, Wake latency after the initiation of optogenetic activation of VLS D1-MSNs during NREM sleep.The columns on the left and right show the light and dark phases, respectively.In each column, left bar shows the activation group, right bar shows the control group [n ¼ 3 mice, 5-10 manipulations/recordings, a total of 23 trials for activation (light phase), 18 trials for control (light phase), a total of 23 trials for activation (dark phase), 30 trials for control (dark phase); p light ¼ 0.01, p dark ¼ 0.03.independent t test].f, Percentage of wake time for 100 s after the initiation of VLS D1-MSNs photoactivation.The columns on the left and right show the light and dark phases, respectively.In each column, left bar shows the activation group, light bar shows the control group [n ¼ 3 mice, 5-10 manipulations/recordings, a total of 23 trials for activation (light phase), 18 trials for control (light phase), a total of 23 trials for activation (dark phase), 30 trials for control (dark phase); p light ¼ 6.3 Â 10 À6 , p dark ¼ 2.4 Â 10 À5 , independent t test].Error bars indicate SEM; *p , 0.05.

Table 2 :
Results of Tukey-Kramer post hoc test for daily percentage of sleep/wake times in light/dark phases among DOXoff days for the data in Figure2c * p , 0.05.

Table 1 :
Results of Tukey-Kramer post hoc test for daily amount of sleep/wake times among DOX-off days for the data in Figure2b * p , 0.05.under progressive D2-MSNs ablation, they were almost unaffected (*p , 0.05, repeated-measures ANOVA followed by the Tukey-Kramer post hoc test; Fig. 2e, paired t test at each time point with Bonferroni correction; Fig.

Table 3 :
Results of Tukey-Kramer post hoc test for daily variations of sleep/wake times between before and after the DOX-off for the data in Figure2dDOX-off day À1 vs day 10 * p , 0.05.ZT: zeitgeber time.

Table 4 :
Results of Tukey-Kramer post hoc test for daily variations of sleep/wake episode numbers between before and after the DOX-off for the data in Figure2f * p , 0.05.ZT: zeitgeber time.

Table 5 :
Results of Tukey-Kramer post hoc test for daily variations of mean durations of sleep-wake states between before and after the DOX-off for the data in Figure2h

Table 6 :
Results of Tukey-Kramer post hoc test for each EEG band power between before and after the DOX-off for the data in Figure3 * p , 0.05.