Influence of Inhibitory Serotonergic Inputs to Orexin/Hypocretin Neurons on the Diurnal Rhythm of Sleep and Wakefulness

Abstract

Study Objective:

Serotonergic (5HT) neurons of the dorsal raphe nuclei receive excitatory input from hypothalamic orexin (hypocretin) neurons and reciprocally inhibit orexin neurons through the 5HT1A receptor. However, the physiological significance of this negative feedback circuit for sleep/wakefulness regulation is little understood.

Design:

5HT1A receptor expression level was specifically and reversibly controlled in the orexin neurons using the Tet-off system. The responsiveness of orexin neurons to 5HT in vitro and the sleep/wakefulness patterns were compared between 5HT1A-overexpressing and control mice.

Measurements and Results:

When the 5HT1A receptor was overexpressed in orexin neurons of Orexin-EGFP; orexin-tTA; TetO Htr1a mice, 5HT-induced inhibition of orexin neurons was prolonged. In the absence of doxycycline, Orexin-tTA; TetO Htr1a mice exhibited severe fragmentation of sleep/wakefulness during the first half of the dark period—the time of maximal activity in nocturnal rodents—without affecting sleep/wakefulness during the light period when sleep time is maximal. However, when the 5HT1A receptor in orexin neurons was reduced to basal expression levels in the presence of doxycycline, sleep/wakefulness patterns in Orexin-tTA; TetO Htr1a mice during the early active period were indistinguishable from those of littermate TetO Htr1a mice. These results strongly suggest that enhancement of inhibitory serotonergic input to orexin neurons caused fragmentation of wakefulness. In contrast, sleep/wakefulness architecture in the light period was unaffected by 5HT1A receptor overexpression in the orexin neurons.

Conclusion:

Inhibitory serotonergic input likely functions as negative feedback to orexin neurons in the early dark period and helps stabilize wakefulness bouts, thereby contributing to the diurnal rhythm of sleep and wakefulness.

Orexin, hypocretin, serotonin, 5HT1A receptor, transgenic, tTA, TetO, doxycycline

INTRODUCTION

The orexins, also called the hypocretins,^1,2 are a pair of neuro-peptides produced in a small number of neurons distributed in the lateral hypothalamic and perifornical areas that send efferent projections throughout the brain.^3,4 The orexin system has been implicated in sleep/wakefulness regulation because animals that lack the prepro-orexin gene,⁵ the orexin receptor 2 (OX2R) gene,⁶ or orexin neurons⁷ show phenotypes similar to the human narcolepsy. Narcoleptic patients exhibit fragmented sleep and wakefulness, sleep onset REM sleep periods and cataplexy, a sudden muscle weakness usually precipitated by positive emotional stimuli. The concentration of orexin-A (hypocretin-1) peptide in the cerebrospinal fluid (CSF) of narcoleptic patients is greatly reduced,⁸ and specific degeneration of orexin neurons has been reported.^9,10 These observations suggest that orexin neurons play a critical role in the regulation of sleep/wakefulness.

Serotonergic neurons in the raphe nuclei are also known to be involved in sleep/wakefulness regulation. Destruction of serotonergic neurons in the raphe nuclei or inhibition of serotonin (5-hydroxytryptophan [5HT]) synthesis with p-chlorophenylalanine (pCPA) produces severe but transient insomnia that can be reversed by restoring 5HT synthesis.^11,12 On the other hand, the firing rate of serotonergic neurons in the dorsal raphe (DR) nucleus is highest in waking, low in slow wave sleep (SWS), and almost quiescent in REM sleep.¹³ Consistent with its likely role in promoting wakefulness, blockade of 5HT neurotransmission through 5HT2A antagonists reduces wakefulness and enhances SWS.¹⁴ The DR sends inhibitory projections to sleep-active neurons in the ventrolateral preoptic area¹⁵ and also innervates the cerebral cortex, thereby influencing cortical arousal. However, the pharmacology underlying the involvement of the serotonergic system in sleep/wakefulness regulation is complicated by the fact that there are at least 14 different 5HT receptor subtypes distributed throughout the brain.

Recent studies have revealed a functional relationship between orexin and serotonergic neurons. Serotonergic neurons are innervated by orexin neurons and are directly and indirectly activated by orexin.¹⁶ Conversely, orexin neurons are densely innervated by serotonergic inputs from the DR.^17,18 Orexin neurons are inhibited by 5HT in vitro via activation of the 5HT1A receptor and subsequent opening of G protein-coupled inwardly rectifying potassium (GIRK) channels.¹⁸ These observations suggest negative feedback circuitry between orexin and serotonergic neurons. However, the functional significance of serotonergic inhibition of orexin neurons is little understood.

To address this, 5HT1A receptor expression level was specifically and reversibly controlled in the orexin neurons using the Tet-off system. Overexpression of 5HT1A receptor in the orexin neurons enhanced serotonergic inhibition and resulted in severe fragmentation of wakefulness, specifically early in the dark (active) period. Normalization of 5HT1A expression in the orexin neurons in the presence of doxycycline (DOX) eliminated this fragmentation. These results suggest that serotonergic inhibition of orexin neurons may function as a negative feedback circuit early in the active period and could thereby contribute to the diurnal rhythms of sleep and wakefulness.

MATERIALS AND METHODS

Animal Usage

All experimental procedures involving animals were approved by the Institutional Animal Care and Use Committee of the National Institutes of Natural Sciences and SRI International and were in accordance with NIH guidelines. All efforts were made to minimize animal suffering or discomfort and to reduce the number of animals used.

Generation of Multiple Orexin-tTA Transgenic Mouse Strains

The characteristics of the 7 transgenic mouse strains used in the present study are summarized in Table 1. The transgenic construct to generate Orexin-tTA transgenic mice was made by substituting the nLacZ gene (SalI-BamHI fragment) of the orexin/nLacZ transgenic construct¹⁹ with 0.7 kb of the mammalianized tetracycline-controlled transcriptional activator (tTA) fragment.²⁰ The transgene was excised and microinjected into pronuclei of fertilized mouse eggs (C57BL/6 mice) to generate transgenic founders. Founders were bred with C57BL/6J mice (Clea-Japan Inc., Tokyo, Japan) to produce stable Orexin-tTA transgenic lines. A total of 9 Orexin-tTA transgene-positive founders were obtained. In situ hybridization analysis of the N1 generation revealed that lines 29, C5, and G5 showed the highest tTA mRNA expression.

Table 1

Names and characteristics of transgenic mouse strains used in the present study

To confirm tTA function, each line was bred with the tetracycline operator (TetO) mCherry red fluorescent protein (RFP)-expressing mouse²¹ to generate Orexin-tTA; TetO RFP mice (Table 1). Among these, line G5 showed the highest penetrance of RFP expression: 85% of orexin neurons expressed RFP. Thus, line G5 was used for further experiments.

To manipulate 5HT1A receptor expression in orexin neurons, Orexin-tTA (line G5) mice were bred with TetO Htr1a mice²² to generate Orexin-tTA; TetO Htr1a mice (Table 1). The TetO Htr1a mice had been backcrossed 7 times with C57BL/6J mice (Clea-Japan Inc., N = 7). In this implementation of the “Tet-off” system, Orexin/tTA;Tet-O Htr1a mice have very high Htr1a mRNA expression in the orexin neurons. However, in the presence of doxycycline (DOX) in the chow, Htr1a mRNA levels in the orexin neurons decline to the basal levels observed in TetO Htr1a mice. Upon subsequent removal of DOX, Htr1a mRNA expression returns to the high level observed in orexin neurons before DOX (see Results), demonstrating reversible Htr1a gene expression in orexin neurons using the Tet-off system. Normal chow and DOX-containing chow (100 mg/kg) were obtained from Nosan Co. (Yokohama, Japan).

In Situ Hybridization

The in situ hybridization method using tissue sections has been described previously.²³ Briefly, digoxigenin-labeled prepro-orexin (584bp, NM_010410.1) or tTA (coding region) cRNA probes were hybridized to tissue sections of mouse hypothalamus, NBT/BCIP compounds (Roche) were used for color development, and Nuclear Fast Red (Vector Lab, Burlingame, CA) was used for counterstaining.

For double fluorescence in situ hybridization, FITC- labeled prepro-orexin cRNA probes and DIG- labeled tTA cRNA probes were hybridized. After stringent washing, a peroxidase -conjugated anti-DIG antibody (Roche, Indianapolis, IN)) was applied and probes were visualized with Cy3 (TSA Plus Cyanine 3 & Fluorescein System, Perkin Elmer, Foster City, CA). After quenching the conjugated peroxidase of the anti-DIG antibody with hydrogen peroxide, peroxidase -conjugated anti-FITC antibody (Perkin Elmer) was applied and probes were visualized with FITC (TSA Plus Cyanine 3 & Fluorescein System, Perkin Elmer).

Electrophysiological Recordings

Triple transgenic Orexin-tTA; orexin-enhanced green fluorescent protein (EGFP); TetO Htr1a mice were generated by breeding Orexin-EGFP mice²⁴ with Orexin-tTA; TetO Htr1a mice (Table 1). Orexin-tTA; orexin-EGFP; TetO Htr1a mice (3-4 week old male and female mice) were used for whole cell slice patch recordings. The mice were deeply anesthetized with isoflurane (Abbott Japan, Tokyo, Japan) and decapitated. Brains were quickly isolated in ice-cold cutting solution consisting of (in mM): 280 sucrose, 2 KCl, 10 HEPES, 0.5 CaCl₂, 10 MgCl₂, 10 glucose, pH 7.4 with NaOH, bubbled with 100% O₂. Brains were cut coronally into 350 μm slices with a microtome (VTA-1200S, Leica, Wetzlar, Germany). Slices containing the lateral hypothalamic area were transferred to an incubation chamber filled with physiological solution containing (in mM): 135 NaCl, 5 KCl, 1 CaCl₂, 1 MgCl₂, 10 HEPES, 10 glucose, pH 7.4 with NaOH, bubbled with 100% O₂ and incubated for at least 1 h at room temperature ([RT] 24-26°C). The slices were transferred to a recording chamber (RC-27L, Warner Instrument Corp.) on a fluorescence microscope stage (BX51WI, Olympus, Tokyo, Japan). Neurons with EGFP fluorescence were subjected to electrophysiological recording. The fluorescence microscope was equipped with an infrared camera (C2741-79, Hamamatsu Photonics, Hamamatsu, Japan) for infrared differential interference contrast (IR-DIC) imaging and a CCD camera (IK-TU51CU, Olympus) for fluorescent imaging. Each image was displayed separately on a monitor (Gawin, EIZO, Tokyo, Japan) and was saved on a computer through a graphic converter (PIX-MPTV, Pixcela, Osaka, Japan).

Recordings were carried out with an Axopatch 200B amplifier (Axon Instruments, Foster City, CA) using borosilicate pipettes (GC150-10, Harvard Apparatus, Holliston, MA) prepared by a micropipette puller (P-97, Sutter Instruments) filled with intracellular solution (4-10 MΩ) consisting of (in mM): 145 KCl, 1 MgCl₂, 10 HEPES, 1.1 EGTA-Na₃, 2 MgATP, 0.5 Na₂GTP, pH 7.3 with KOH. The osmolarity of the solutions were checked by a vapor pressure osmometer (model 5520, Wescor, Logan, UT) and found to be 280-290 and 320-330 mOsm/L for the internal and external solutions, respectively. The liquid junction potential of the patch pipette and perfused extracellular solution was estimated to be 4.3 mV and applied to the data. Recording pipettes were under positive pressure while advancing toward individual cells in the slice. Tight seals on the order of 1.0-1.5 GΩ were made by negative pressure. The membrane patch was then ruptured by suction. The series resistance during recording was 10-25 MΩ. The reference electrode was an Ag-AgCl pellet immersed in bath solution. During recordings, cells were super-fused with extracellular solution at a rate of 1.6 mL/min using a peristaltic pump (Dynamax, Rainin, Oakland, CA). In current clamp recordings, the membrane potential was set at -60 mV by current injection before the experiments. 5HT (Sigma, St. Louis, MO) was dissolved in the extracellular solution and applied by local application through a fine tube (100 μm diameter) positioned near the neuron being recorded. Tetrodotoxin (TTX) (Wako, Osaka, Japan) was dissolved in extracellular solution and applied by bath application.

Output signals were low pass filtered at 5 kHz and digi-tized at 10 kHz. Data were recorded on a computer through a Digidata 1322A A/D converter using pClamp software (ver. 10, Axon Instruments, Union City, CA, USA). Traces were processed for presentation using Origin 8.1 (Origin Lab Corporation, Northampton, MA, USA) and Canvas X (ACD systems, British Columbia, Canada) software.

Immunohistochemical Studies

Male and female Orexin-tTA; TetO RFP double transgenic mice (10 weeks old) were deeply anesthetized with isoflurane and perfused sequentially with 20 mL of chilled saline and 20 mL of chilled 10% formalin solution (Wako). The brains were removed and immersed in the above fixative solution for 24 h at 4°C, and then immersed in a 30% sucrose solution for at least 2 d. The brains were quickly frozen in embedding solution (Sakura Fine-technical Co., Ltd., Tokyo, Japan). For orexin and RFP double-staining, coronal sections (40 μm) of Orexin-tTA; TetO RFP mouse brains were incubated with rabbit polyclonal anti-RFP antiserum (1/1000, Cat. #632496, Clontech, CA, USA) for 24 h at 4°C. These sections were incubated with Alexa 594-labeled goat anti-rabbit IgG (1/800, Cat. #A-11037, Invitrogen, CA) for 1 h at RT. The sections were then incubated with a guinea pig anti-orexin-A antiserum (1/500) for 24 h at 4°C. The antiserum was produced in our laboratory by immunizing guinea pigs with the same antigen reported in Nambu et al.³ The antiserum was affinity purified using an affinity column. Sections were then incubated with Alexa 488-labeled goat anti-guinea pig IgG (1/800, Cat.# A-11073, Invitrogen) for 1 h at RT. The sections were mounted and examined with a fluorescence microscope (BZ-9000, Keyence, Osaka, Japan). This orexin antiserum did not produce any labeling on brain sections from orexin knockout mice.

EEG/EMG Recording and Analyses

Male mice (12 weeks old at the time of surgery) were anes-thetized with pentobarbital (50 mg/kg, i.p.) and implanted with electroencephalogram (EEG) and electromyogram (EMG) electrodes for polysomnographic recording. EEG signals were recorded using gold-coated pins placed on the dura mater (4 pins located ± 1.5 mm lateral of the bregma, ± 1.25 mm anterior or posterior to bregma). The EMG signal was acquired through a pair of multi-stranded stainless-steel wires inserted into the neck extensor muscles. Mice were then housed separately for a recovery period ≥ 7 d before recording. Continuous EEG and EMG recordings were carried out through a slip ring (Air Precision, Le Pressis Robinson, France) designed so that the movement of the mouse was unrestricted. EEG and EMG signals were amplified (AB -610J, Nihon Koden, Tokyo, Japan), filtered (EEG 1.5-30 Hz; EMG 15-300 Hz), digitized at a sampling rate of 128 Hz, and recorded using SleepSign version 3 (Kissei Comtec, Nagano, Japan). The animal's behavior was monitored through a CCD video camera and recorded on a computer synchronized with EEG and EMG recordings using the SleepSign video option system (Kissei Comtec).

EEG and EMG records were automatically scored in 4-sec epochs and classified as wakefulness, SWS, or REM sleep by SleepSign software according to standard criteria.^25,26 All vigilance state classifications assigned by SleepSign were examined visually and corrected. The same individual, blinded to genotype and experimental condition, scored all EEG/EMG recordings. Spectral analysis of the EEG was performed by fast Fourier transform (FFT). This analysis yielded a power spectra profile over a 0-40 Hz window with a 1 Hz resolution divided into delta (1-5 Hz), theta (6-10 Hz), alpha (10-13 Hz), and beta (13-25 Hz) bandwidths.

Statistical Analyses

Data were analyzed by one-way ANOVA followed by Fisher protected least significant difference test using the Stat View 4.5 software package (Abacus Concepts, Berkeley, CA, USA). Probability (P)-values less than 0.05 were considered statistically significant.

RESULTS

tTA mRNA Is Specifically Expressed in the Orexin Neurons

To control gene expression specifically in the orexin neurons using the tetracycline gene expression system, transgenic mice were generated in which orexin neurons specifically express tTA. To confirm that tTA mRNA was expressed in the orexin neurons, in situ hybridization was performed on brain slices from Orexin-tTA transgenic mice brains. Figures 1A show that prepro-orexin mRNA was expressed in the lateral hypothalamic area. Figures 1B show that tTA mRNA was specifically expressed in the same area. Next, tTA mRNA expression in the orexin neurons was confirmed using double fluorescent in situ hybridization (Figure 1C, D, E). The merged picture in Figure 1E shows that tTA mRNA (Cy3, red, middle panel) was expressed in the same neurons that expressed prepro-orexin mRNA (FITC, green, left panel). These results confirmed that tTA was expressed in the orexin neurons in the Orexin-tTA mouse brain.

Figure 1

Generation of Orexin-tTA transgenic mice. In situ hybridization revealed that tTA mRNA was expressed in the orexin neurons of the Orexin-tTA mouse brain. Prepro-orexin mRNA (A) and tTA mRNA (B) are expressed in the same area within the lateral hypothalamus. (C, D and E) Double fluorescent in situ hybridization. Prepro-orexin mRNA (FITC, green) (C); tTA mRNA (Cy3, red) (D); merged picture (E). Scale bar B, 200 μm; E, 50 μm.

tTA-Induced Gene Activity in Orexin Neurons

Next, tTA-induced gene activity was confirmed by breeding Orexin-tTA mice with TetO RFP mice (which express RFP in the presence of tTA) to generate double transgenic Orexin-tTA; TetO RFP mice (Figure 2A; Table 1). In the double transgenic mouse brain, RFP was specifically expressed in the orexin neurons (Figure 2B) since RFP-ir (Alexa594, red) was observed only in orexin-ir neurons (Alexa488, green) and no RFP expression was found elsewhere in the brain. Furthermore, 85% of orexin-ir neurons expressed RFP (n = 4). These results indicate that tTA was expressed in the orexin neurons and correctly induced gene activity in the orexin neurons in the brains of these transgenic mice.

Figure 2

Functional confirmation of tTA-induced gene expression. (A) Drawing illustrating breeding scheme in which Orexin-tTA mice were bred with TetO RFP mice to generate double transgenic Orexin-tTA; TetO RFP mice. tTA induced RFP expression in the orexin neurons in Orexin-tTA; TetO RFP mice. (B) Immunohistochemical study revealed that RFP was specifically expressed in the orexin neurons. Left to right: RFP-ir (red); orexin-ir (green); merged picture. White square area in the upper panel is enlarged in the bottom panel. hOX promoter, human prepro-orexin promoter. Scale bars in B: upper panel, 100 μm; lower panel, 50 μm.

Reversible Control of Htr1a mRNA Expression in the Orexin Neurons

We previously reported that orexin neurons were densely innervated by serotonergic neurons and were significantly inhibited through the 5HT1A receptor expressed in orexin neurons.¹⁸ The 5HT1A receptor is coupled with the Gi/o subtype of Gα subunit and subsequently activates GIRK channels.²⁷ To evaluate the physiological significance of this inhibitory sero-tonergic input to orexin neurons, Orexin-tTA transgenic mice were bred with TetO Htr1a mice in which the TetO sequence was knocked into the promoter region of the 5HT1A receptor gene allele to generate Orexin-tTA; TetO Htr1a mice²² (Table 1; Figure 3A). Since tTA binds to the TetO sequence, Htr1a mRNA transcription should be elevated above the basal Htr1a promoter activity, resulting in overexpression of 5HT1A receptor in tTA-expressing cells.²² Although the TetO sequence was inserted into endogenous 5HT1A receptor promoter region, this insertion does not interfere with endogenous 5HT1A receptor promoter activity. Thus, the 5HT1A receptor expression pattern in these knock-in mice was comparable to wild-type mice.²⁸ However, in the presence of DOX, tTA loses its ability to bind to the TetO sequence (Figure 3B), which should result in the return of Htr1a mRNA expression level in orexin neurons to the basal level (Figure 3C).

Figure 3

Reversible control of Htr1a mRNA expression in the orexin neurons. (A) Drawing illustrating breeding scheme in which Orexin-tTA mice were bred with TetO Htr1a mice to generate double transgenic Orexin-tTA; TetO Htr1a mice. (B and C) 5HT1A receptor gene expression was reversibly regulated in the orexin neurons using the tTA gene expression system in the presence or absence of DOX. DOX(+) indicates the presence of DOX; DOX(−) indicates the absence of DOX. O, overexpression; B, basal expression. (D) Double fluorescent in situ hybridization revealed that Htr1a mRNA was expressed in prepro-orexin-expressing neurons in Orexin-tTA; TetO Htr1a mice. Left to right: Prepro-orexin (FITC, green); Htr1a (Cy3, red); merged image. (E)In situ hybridization demonstrated reversible control of Htr1a mRNA expression in the orexin neurons. The upper and middle panels show Htr1a mRNA (a, b, c, and d) and prepro-orexin mRNA (e, f, g, and h) in the hypothalamus. The lower panel shows Htr1a mRNA (i, j, k and l) in the hippocampus. Htr1a mRNA was overexpressed in the orexin neurons in the absence of DOX (b and d). However, in the presence of DOX for 5 d (c), Htr1a mRNA returned to expression levels comparable to those of monogenic TetO Htr1a mice (a). Neither prepro-orexin nor hippocampal Htr1a mRNA expression was altered in the presence or absence of DOX. Scale bars: D, 100 μm; E, 1.0 mm.

Htr1a mRNA overexpression in orexin neurons of Orexin-tTA; TetO Htr1a mice was confirmed by fluorescent in situ hybridization in the absence of DOX. Figure 3D demonstrates that Htr1a mRNA (red, Cy3) was indeed expressed in prepro-orexin mRNA-expressing neurons (green, FITC).

Next, we used the “ABA” experimental design illustrated in the upper panel of Figure 3E to determine whether Htr1a mRNA expression could be reversibly controlled in the orexin neurons of the Orexin-tTA; TetO Htr1a mice by applying or removing DOX in the chow. Htr1a mRNA expression level was visualized by in situ hybridization. Orexin/tTA; TetO Htr1a mice were fed normal chow until 14 weeks of age. Panel b of Figure 3E illustrates that, at 13 weeks of age, Htr1a mRNA was overexpressed in the orexin neurons in these mice (“preDOX” condition). The chow was then replaced with chow containing DOX for 7 d, from 14 to 15 weeks of age. After 5 d of DOX chow consumption, Htr1a mRNA expression in the orexin neurons returned to basal level (DOX(+), Figure 3E, Panel c). This expression level did not differ from that of TetO Htr1a control mice that lack the Orexin-tTA transgene (Figure 3E, Panel a). The Htr1a mRNA expression pattern and levels in the hypothalamus of TetO Htr1a mice did not differ from that of wild type mice (data not shown). When DOX chow was replaced with normal chow for two weeks from 15 to 17 weeks of age, Htr1a mRNA was overexpressed again (“postDOX,” Figure 3E, Panel d). Preproorexin mRNA in the hypothalamus and Htr1a mRNA in the hippocampus were unaffected by the presence or absence of DOX (prepro-orexin mRNA: Figure 3E, Panels e, f, g, and h; hippocampal Htr1a mRNA: Figure 3E, Panels i, j, k, and l). These results confirm that Htr1a mRNA expression level was reversibly controlled by the presence or absence of DOX in the chow.

Patch Clamp Electrophysiological Studies Reveal the Effects of 5HT1A Receptor Overexpression in Orexin Neurons

To determine the physiological effect of 5HT1A receptor overexpression in orexin neurons, slice patch clamp recording was performed. Since orexin neurons are sparsely distributed in the lateral hypothalamic area and there are no morphological features to distinguish them from other neurons, we generated triple transgenic mice, Orexin-EGFP; orexin-tTA; TetO Htr1a mice, by crossbreeding Orexin-EGFP mice with Orexin-tTA; TetO Htr1a mice (Table 1). In these mice, EGFP was expressed and the 5HT1A receptor was overexpressed in the orexin neurons in the absence of DOX in the chow. In contrast, in the presence of DOX in the chow, 5HT1A receptor expression level in the orexin neurons was normal (not overexpressed). Thus, orexin neurons of Orexin-EGFP; orexin-tTA; TetO Htr1a mice in the presence of DOX were used as control and compared to orexin neurons from Orexin-EGFP; orexin-tTA; TetO Htr1a mice in the absence of DOX.

First, we tested whether the basic membrane properties of orexin neurons were affected by overexpression of 5HT1A receptor in orexin neurons. Electrical properties of 5HT1A receptor-overexpressing orexin neurons were compared to those of control mice that were fed with DOX-containing chow until the day of experiment. There were no significant differences in the resting membrane potential, peak amplitude of action potentials, spontaneous firing, input resistance or membrane capacitance of orexin neurons between the overexpressing mice and control mice (Table 2). These results indicate that overexpression of the 5HT1A receptor did not alter the basic membrane properties of orexin neurons.

Table 2

Electrical membrane properties of 5HT1A receptor-overexpressing orexin neurons recorded from mice with and without DOX

Next, we measured membrane potential from orexin neurons in current clamp mode in the presence of tetrodotoxin (TTX, 1 μM) in the bath solution. Local application of 5HT (10 μM) significantly hyperpolarized orexin neurons from both control mice and overexpressing mice (Figure 4A). Local application of 5HT (10 μM) induced -13.0 ± 1.6 mV (n = 8) hyperpolarization of orexin neurons from control mice and -14.3 ± 1.3 mV (n = 14) hyperpolarization of orexin neurons from overexpressing mice. These changes in membrane potential induced by 5HT (10 μM) were not significantly different between the 2 mouse groups (P = 0.51, Student t-test). Figure 4B shows that the reversal potential (E_rev) induced by 5HT (10 μM) application was -93.6 ± 4.9 mV (n = 11), close to theoretical E_rev of potassium (-90.8 mV) calculated from the Nernst equation in these recording conditions. These results are consistent with those obtained previously,¹⁸ indicating that GIRK channels open downstream of 5HT1A receptor activation even in orexin neurons with overexpressed 5HT1A receptor levels. To confirm that GIRK-induced current is intact in 5HT1A-overexpressing orexin neurons, GIRK channelinduced hyperpolarization by a pathway other than the 5HT1A receptor was tested. Baclofen activates GABA_B receptors that are coupled to the Gi/o subtype of Gá subunit and which activate GIRK channels. GABA_B receptors are expressed in orexin neurons and baclofen-induced hyperpolarization of orexin neurons through activation of GIRK channels has previously been reported.²⁹ The magnitude of baclofen (30 μM)-induced hyperpolarization in 5HT1A-overexpressing orexin neurons from experimental and control mice was -9.2 ± 0.9 mV (n = 5) and -9.5 ± 1.1 mV (n = 7, P = 0.80, Student t-test), respectively. This result suggests that the intracellular signal transduction pathway to activate GIRK channels is not affected by 5HT1A receptor overexpression in orexin neurons.

Figure 4

Htr1a overexpression in the orexin neurons prolonged 5HT-induced inhibition of firing. (A) Under current-clamp mode, local application of 5HT (10 μM) induced a prolonged hyperpolarization of orexin neurons in 5HT1A receptor overexpressing Orexin-EGFP; orexin-tTA; TetO Htr1a mice in the absence of DOX (middle) compared to control Orexin-EGFP; orexin-tTA; TetO Htr1a mice in the presence of DOX (left), as illustrated best in the merged traces (right). In current clamp, the membrane potential was set at -60 mV by current injection before the experiments. (B) Graph showing current-voltage relationship in the presence (filled circles) or absence (open circles) of 5HT (10 μM) in orexin neurons recorded from overexpressing Orexin-EGFP; orexin-tTA; TetO Htr1a mice. Membrane potential in response to a series of 100 ms current steps (in 20 pA increments, -80 pA to 0 pA) from -60 mV was plotted (n = 11). Estimated reversal potential (E_rev) was -93.6 ± 4.9 mV (n = 11). (C and D) Bar graphs summarizing the data obtained from recordings such as those illustrated in A. C and D present the area and the half-width induced by 5HT-induced hyperpolarization, respectively, as illustrated by the insets (n = 3-16). (E) Representative traces of loose cell-attached recordings of orexin neurons in hypothalamic slices from Orexin-EGFP; orexin-tTA; TetO Htr1a mice in the presence of DOX (control, top) and Orexin-EGFP; orexin-tTA; TetO Htr1a mice in the absence of DOX (overexpression, bottom). 5HT was locally applied for 5 sec during the period represented by bars and inhibited the firing of action potentials in the orexin neurons. (F) Bar graphs comparing the duration of 5HT-induced inhibition illustrated in E (n = 10-17). (G) Representative traces showing loose cell-attached recordings of 5HT1A overexpressing orexin neurons (Orexin-EGFP; orexin-tTA; TetO Htr1a mice). 5HT (10 μM) was sequentially and repetitively applied 3 times. 5HT was locally applied for 5 sec during the period represented by bars and inhibited the firing of action potentials in the orexin neurons. (H) Bar graph comparing the duration of 5HT-induced inhibition illustrated in E (n = 9). Duration of inhibition is represented as percentage of 1^st application. *P < 0.05 vs. control. Values are mean ± SEM.

The 5HT-induced hyperpolarization was prolonged in orexin neurons in overexpressing mice compared with control mice (Figure 4A merged trace: control [black] and overexpressing mice [red]). To compare the magnitude of the hyperpolarization induced by 5HT application, the area and half-width were analyzed (Figure 4C and Figure 4D, respectively). The area of the hyperpolarizing deflection induced by 5HT application (see inset in Figure 4C) was calculated in units, in which 1 unit was 1 mV × 1 sec. The half-width was calculated as the time to reach 50% of the peak value induced by 5HT application (see inset in Figure 4D). The area and half-width of 5HT (10 μM)-induced hyperpolarization was increased approximately 3.0-fold and 2.3-fold, respectively, in orexin neurons from overexpressing mice compared with control mice. The areas induced by 5HT (10 μM) in orexin neurons in control and overexpressing mice were 155 ± 23 units (n = 8) and 455 ± 43 units (n = 10, P = 0.00002, ANOVA, vs. control). The half -widths of the hyperpolarization duration induced by 5HT (10 μM) in orexin neurons from control and overexpressing mice were 11.7 ± 1.3 sec (n = 8) and 26.3 ± 4.1 sec (n = 10, P = 0.005, ANOVA vs. control).

This prolongation of 5HT inhibition was also confirmed using loose cell-attached recording. Orexin neurons from control and overexpressing mice showed spontaneous firing at rates of 2.9 ± 0.2 (n = 21) and 3.3 ± 0.2 Hz (n = 16), respec-tively. Although 5HT application inhibited the generation of action potentials in orexin neurons from both control and overexpressing mice, the inhibitory effect was prolonged in orexin neurons from overexpressing mice compared with control mice (Figure 4E and F). Inhibitory duration was measured as the “silent period” induced by 5HT (see inset in Figure 4F). Bath application of 3, 10, and 30 μM 5HT prolonged inhibition 2.0-fold (n = 10, P = 0.002, ANOVA vs. control), 1.7-fold (n = 13, P = 0.005, ANOVA vs. control) and 1.7-fold (n = 16, P = 0.0003, ANOVA vs. control), respec-tively, compared with control (Figure 4F). To confirm that 5HT-induced inhibition of 5HT1A-expressing orexin neurons was a reproducible response, sequential repetitive application of 5HT was tested (Figure 4G). Three sequential 5HT (10 μM) applications induced the same duration of firing inhi-bition in 5HT1A-overexpressing orexin neurons. When the inhibitory duration of the first application was set as 100%, the second and third applications induced 109.4% ± 8.5% (n = 9, P = 0.22, ANOVA vs. 1st) and 94.4% ± 4.8% (n = 9, P = 0.46, ANOVA vs. 1st) inhibition, respectively (Figure 4H). These results indicate that 5HT1A receptor overexpression in orexin neurons significantly enhanced the inhibition produced by 5HT application, that is, the inhibitory serotonergic input to orexin neurons was intensified in orexin neurons of 5HT1A receptor-overexpressing mice.

Effects of Enhanced Serotonergic Inhibition of Orexin Neurons on Sleep/Wakefulness Regulation

To determine the in vivo significance of the inhibitory sero-tonergic input to orexin neurons, Orexin-tTA; TetO Htr1a mice were subjected to sleep/wakefulness analyses, and littermate TetO Htr1a monogenic mice were used as controls. Sleep/ wakefulness states were identified as wakefulness, SWS, and REM sleep from EEG and EMG recordings. For these experiments, DOX chow was provided and withdrawn according to the protocol presented in Figure 3E. In TetO Htr1a monogenic control mice, wakefulness was consolidated in the first half of the dark period (20:00-02:00), with mice maintaining wakefulness bouts as long as 2 h, since mice are most active at that time of day (Figure 5A, top). In contrast, during the first half of the light period (08:00-14:00), TetO Htr1a mice frequently transitioned between wakefulness and sleep since mice are less active at that time of day (Figure 5A, bottom). Sleep/wakefulness patterns were analyzed in 3-h bins across the 24-h cycle. The total time spent in wakefulness, SWS, and REM sleep in the dark (Figure 5B, top) and light (Figure 5B, bottom) periods was analyzed and the frequency of transitions between states was calculated (Figure 5C). In TetO Htr1a mice, neither the total time in each state (Figure 5B) nor the transition frequency (Figure 5C) was affected by the presence or absence of DOX in the chow. These results indicated that DOX application had no direct effect on the physiological regulation of sleep/wakefulness.

Figure 5

DOX application does not affect sleep/wakefulness pattern in TetO Htr1a mice. (A) hypnograms from TetO Htr1a mice recorded in the presence or absence of DOX (100 mg/kg) in chow. The upper panel presents representative hypnograms from the dark period (from 20:00 to 08:00). The lower panel presents representative hypnograms from the light period (from 08:00 to 20:00). PreDOX (left panels) is a recording of 13-week-old mouse fed normal chow; DOX (middle panels) is 5 d after DOX chow; postDOX is 14 d after removal of DOX chow. (B) summarizes the time spent in wakefulness, SWS, and REM sleep in each 3-h period throughout the light (bottom) and dark (top) periods. (C) Bar graphs showing transition frequency in 3-h bins across the 24-h period. Values are mean ± SEM (n = 8). W or Wake, wakefulness; S or SWS, slow wave sleep; R or REM, REM sleep.

In contrast to TetO Htr1a mice, Orexin-tTA; TetO Htr1a mice showed severe fragmentation of wakefulness during the first half of the night in the absence of DOX in the chow (“preDOX,” Figure 6A top left). Wakefulness was consolidated at this time of day in the DOX(+) condition (“DOX(+),” Figure 6A top middle). The hypnogram of DOX(+) condition in Orexin-tTA; TetO Htr1a mice (Figure 6A) was similar to that of TetO Htr1a mice (Figure 5A). The mice maintained wakefulness bouts as long as 2 h (Figure 6A top middle) but sleep/wakefulness fragmentation returned when DOX was again removed from the chow for 14 d (“postDOX,” Figure 6A top right).

Figure 6

Htr1a mRNA overexpression in orexin neurons increases the sleep time during the early dark period but does not affect sleep/wakefulness pattern in the light period. (A) hypnograms from Orexin-tTA; TetO Htr1a mice recorded in the presence or absence of DOX in the chow. The upper panel presents representative hypnograms from the dark period (from 20:00 to 08:00). The lower panel presents representative hypnograms from the light period (from 08:00 to 20:00). PreDOX (left panels) is during 5HT1A receptor overexpression; DOX (middle panels) is 5 d after DOX application which reduces 5HT1A receptor overexpression in orexin neurons; postDOX is 14 d after removal of DOX chow. (B) summarizes the time spent in wakefulness, SWS and REM sleep in each 3-h period throughout the light (bottom) and dark (top) periods. (C) Bar graphs showing transition frequency in 3-h bins across the 24-h period. *P < 0.05 vs. DOX. Values are mean ± SEM (n = 12). W or Wake, wakefulness; S or SWS, slow wave sleep; R or REM, REM sleep.

Sleep/wakefulness patterns of Orexin-tTA; TetO Htr1a mice were analyzed in 3-h bins across the 24-h cycle (Figure 6B and 6C). The total time in wakefulness was increased and total amounts of SWS and REM sleep were decreased during the first 6 h (20:00-23:00 and 23:00-02:00) of the dark period in the presence of DOX (DOX(+)) compared to the absence of DOX (preDOX). The transition frequency significantly decreased in the first 3 h of the dark period (20:00-23:00) (Figure 6C), consistent with sustained wakefulness bouts (Figure 6B). During the second 3 h (23:00-02:00), there was a tendency toward decreased transition frequency in the presence of DOX.

These results indicate that the amount of wakefulness was increased during the first 6 h of the dark period. These effects on sleep/wakefulness were eliminated by removing DOX from chow for 14 d (postDOX condition). No significant differences were observed between preDOX and postDOX during any 3-h epoch during the dark phase. Interestingly, however, the sleep/ wakefulness pattern in the Orexin-tTA; TetO Htr1a mice during the light period was unaffected by the presence or absence of DOX. No significant differences were observed between preDOX, DOX(+), and postDOX during any time period in the light phase (Figure 6B bottom).

Table 3 summarizes the total time in wakefulness, SWS, and REM sleep in the dark and light periods for both mouse strains. In TetO Htr1a mice, the time spent in wakefulness, SWS, and REM sleep in both the dark and light periods was unaffected by the presence or absence of DOX. However, in Orexin-tTA; TetO Htr1a mice, ANOVA revealed that the total time in wakefulness was significantly increased and the total time in SWS and REM sleep were significantly decreased during the dark period in DOX(+) condition compared with the preDOX condition. These effects on sleep/wakefulness were eliminated by removing DOX from chow for 14 d (postDOX condition); the total time in wakefulness, SWS and REM sleep were comparable to those in the preDOX condition. The fragmentation of wakefulness in the absence of DOX (preDOX and postDOX) during the first half of the dark phase in conjunction with Htr1a mRNA overexpression in the orexin neurons suggests that enhancement of inhibitory serotonergic input to orexin neurons is causally related to the fragmentation of wakefulness during the early dark period.

Table 3

Sleep/wake amounts in TetO Htr1a and Orexin-tTA; TetO Htr1a mice in the presence and absence of DOX

The mean wakefulness bout duration and mean SWS bout duration were also calculated for the first half of the dark period (6 h, 20:00- 03:00). The mean wakefulness bout duration in Orexin-tTA; TetO Htr1a mice was significantly prolonged in the presence of DOX. Mean wakefulness bout durations of Orexin-tTA; TetO Htr1a mice in the preDOX, DOX(+), and postDOX conditions were 12.3 ± 2.7 min, 25.6 ± 5.2 min (P = 0.03 vs. preDOX) and 13.7 ± 5.2 min (P = 0.83 vs. preDOX), respectively. In contrast, mean wakefulness bout durations of TetO Htr1a mice were not affected by the presence or absence of DOX. Mean wakefulness bout durations of TetO Htr1a mice in preDOX, DOX(+), and postDOX conditions were 27.6 ± 8.2 min, 30.7 ± 10.0 min (P = 0.80 vs. preDOX), and 26.2 ± 8.4 (P = 0.90 vs. preDOX), respectively. On the other hand, mean SWS bout durations were unaffected by the presence of DOX in either Orexin-tTA; TetO Htr1a mice or TetO Htr1a mice. Mean SWS bout durations of Orexin-tTA; TetO Htr1a mice in the preDOX, DOX(+), and postDOX conditions were 4.4 ± 0.3 min, 4.2 ± 0.4 min (P = 0.55 vs. preDOX), and 4.0 ± 0.3 min (P = 0.42 vs. preDOX), respectively. Mean SWS bout durations of TetO Htr1a mice from the preDOX, DOX(+), and postDOX conditions were 5.6 ± 0.7 min, 5.4 ± 0.6 min (P = 0.86 vs. preDOX), and 4.8 ± 0.5 min (P = 0.38 vs. preDOX), respectively. These results indicate that 5HT1A receptor overexpression in orexin neurons induced fragmentation of wakefulness.

DISCUSSION

A combination of electrophysiological and histochemical studies had previously identified a neural circuit between the serotonergic neurons in the raphe nuclei and the orexin neurons in the hypothalamus. However, the function of this circuit in the regulation of sleep/wakefulness is little understood. Here, we generated transgenic mice that enabled reversible control of 5HT1A receptor expression specifically in the orexin neurons and found that this negative feedback circuit likely contributes to stable, sustained periods of wakefulness early in the active (dark) period.

Inhibitory Serotonergic Input to Orexin Neurons

Serotonergic neurons in the raphe nuclei are well known to be involved in sleep/wakefulness regulation. However, the effects of 5HT on sleep/wakefulness are complex, and how 5HT regulates sleep/wakefulness has been controversial: 5HT promotes wakefulness and inhibits REM sleep in some cases, whereas it increases sleep propensity in other circumstances.³⁰ These conflicting reports of the effect of 5HT on sleep/wakefulness might be explained by the relationship between serotonergic neurons and other neurons involved in sleep/wakefulness regulation, such as orexin neurons in the hypothalamus. Orexin neurons densely innervate serotonergic neurons in the DR nucleus and activate these neurons directly and indirectly through activation of both orexin 1 and orexin 2 receptors.^31,32 On the other hand, we previously reported that orexin neurons received dense serotonergic innervation and were strongly inhibited by 5HT through the 5HT1A receptor and subsequent activation of GIRK channels.^17,18 It is currently unclear how this circuit functions to regulate sleep/wakeful-ness, since both orexin neurons and serotonergic neurons in the raphe nuclei are reported to be active during wakefulness and discharge less or are silent during SWS and REM sleep.^33–36 However, the firing rate of orexin cells is not simply a function of arousal state, as they are relatively inactive during quiet wakefulness but discharge in active waking with maximal activity during exploratory behavior.³⁴ Although the effects of 5HT1A overexpression on the firing of orexin neurons is yet to be directly determined, from the results presented here, enhanced 5HT1A-mediated inhibition of these cells apparently reduces wakefulness drive and allows sleep to occur early in the dark period, a time of day when there is normally a low probability of sleep.

Reversible Control of 5HT1A Receptor Expression in Orexin Neurons

To reveal the physiological role of the inhibitory serotonergic input to orexin neurons on sleep/wakefulness regulation, 5HT1A receptor expression level in the orexin neurons was reversibly controlled using the Tet-off system. In situ hybridization studies confirmed that Htr1a mRNA expression level in the orexin neurons of Orexin-tTA; TetO Htr1a mice was reversibly controlled by adding or removing DOX from the chow. In the absence of DOX in the chow, Htr1a mRNA was overexpressed because tTA enhanced transcription of Htr1a mRNA in the orexin neurons. 5HT1A receptor overexpression in the orexin neurons returned to basal levels by adding DOX to the chow for 5 d. Although the TetO sequence was inserted into endogenous 5HT1A receptor promoter region, this insertion does not interfere with the normal 5HT1A receptor expression pattern and DOX does not interfere endogenous 5HT1A receptor promoter activity.²⁸ Thus, Htr1a mRNA expression in non-orexin neurons, such as hippocampal neurons, was unaffected by the presence or absence of DOX and histological studies confirmed that Htr1a mRNA levels were specifically controlled in the orexin neurons of Orexin-tTA; TetO Htr1a mice.

Slice patch clamp recording from orexin neurons revealed that Htr1a mRNA overexpression prolonged the inhibitory effect of 5HT in orexin neurons approximately 2-fold. This result suggests that the number of 5HT1A receptors in the plasma membrane of orexin neurons was increased by Htr1a mRNA overexpression. However, the effect of 5HT1A receptor overexpression on 5HT-induced response was weaker than expected since in situ hybridization showed that Htr1a mRNA was robustly overexpressed in the absence of DOX. It is possible that 5HT1A receptor protein in the plasma membrane did not increase in parallel with the mRNA increase. Trafficking of GPCRs from the endoplasmic reticulum/Golgi apparatus to the plasma membrane is tightly regulated with multiple steps to ensure correct GPCR targeting.³⁷ These multiple steps might restrict robust expression of GPCRs in the membrane. Alternatively, a limitation on the number of G proteins and GIRK channels expressed in the orexin neurons might be involved in this mismatch between mRNA expression level and physiological response.

Physiological Significance of Inhibitory Serotonergic Input to the Orexin Neurons for Sleep/Wakefulness Regulation

Reversible gene expression control using the Tet-off system is a powerful approach for the analysis of the effects of cell type-specific gene expression on physiology or behavior because it enables comparisons to be made before and after gene expression control within the same animal. In the present study, we reversibly controlled 5HT1A receptor expression level in the orexin neurons. The pattern and amount of wakefulness in the early dark period was significantly altered by 5HT1A receptor expression level in the orexin neurons as mice showed severe fragmentation of wakefulness when 5HT1A receptor was over-expressed in these cells. These results strongly suggest that enhancement of inhibitory serotonergic input to orexin neurons caused fragmentation of wakefulness. However, sleep/wakeful-ness architecture in the light period was unaffected by 5HT1A receptor overexpression in the orexin neurons. These results suggest that inhibitory serotonergic input to orexin neurons may function as negative feedback in the early dark period. Since mice are nocturnal, they are highly active during the dark period and many mouse strains are most active shortly after lights off. During this time, inhibitory serotonergic input might keep the activity of orexin neurons in the moderate range as a negative feedback circuit. However, enhancement of inhibitory seroto-nergic input to orexin neurons by 5HT1A receptor overexpression likely reduces the firing of the orexin neurons that normally provide excitatory input to monoaminergic and cholinergic systems as well as to the serotonergic raphe neurons. Fragmentation of wakefulness could thus result from reduced excitatory drive onto these arousal systems, resulting in the inability to sustain prolonged bouts of wakefulness during this period.

We recently reported that the activity of serotonergic neurons in vivo is highly dependent on the activity of orexin neurons.³⁸ Acute inhibition of orexin neurons using optogenetics induced SWS in conjunction with a decrease in serotonergic neuron activity. Decreased firing of serotonergic raphe neurons presumably results in decreased transmitter release at 5HT nerve terminals, including those that synapse on orexin neurons. Most orexin neurons express c-fos in the early dark period,³⁹ suggesting elevated discharge of orexin neurons at this time. Orexin neurons are hypothesized to receive high frequency excitatory inputs that depolarize orexin neuron membrane potential to maintain wake-fulness early in the dark phase. Decreased serotonergic inhibition might result in orexin neuron overexcitation, resulting in increased locomotor activity and sustained wakefulness. Imbalance between excitatory and inhibitory inputs to orexin neurons by 5HT1A receptor overexpression in Orexin-tTA; TetO Htr1a mice might be the cause of wakefulness fragmentation in the dark period. During the light period, however, serotonergic inhibitory input may be too weak to affect the firing of orexin neurons. A similar fragmentation of wakefulness was observed in mice with selective loss of GABA_B receptors in orexin neurons.²⁹ This result supports our hypothesis that interruption of the balance between excitatory and inhibitory input to orexin neurons is important in the physiological regulation of sleep/wakefulness. However, these GABA_B1^flox/flox; orexin-Cre mice showed fragmentation in the light period as well,²⁹ suggesting inhibitory GABAergic input to orexin neurons functions in both the light and dark periods.

Utility of Reversible Gene Expression to Understand Physiology and Behavior Mediated Through the Orexin System

The present study illustrates the power of reversible gene expression to elucidate the functional significance of a particular receptor in a neural circuit thought to be involved in sleep/ wakefulness regulation. Over the past decade, anatomical and in vitro cellular physiological studies have identified a number of afferent inputs to the orexin neurons and the receptors that mediate the complex balance of excitatory and inhibitory inputs onto these cells. In the coming decade, Orexin-tTA mice, when bred with various lines of TetO mice, will provide a wide reper-toire of tools to control gene expression in the orexin neurons and thereby elucidate the physiological significance of specific inputs to this neuronal system.⁴⁰ Although the present study has focused on sleep/wakefulness regulation, given the involvement of the orexin system in energy metabolism, addiction, reward, stress, and other aspects of physiology and behavior, these combinatorial transgenic approaches should have wide-spread utility beyond sleep/wakefulness control.

ACKNOWLEDGMENTS

This study was supported by the JST PRESTO program, Grant-in-Aid for Scientific Research on Innovative Areas “Mesoscopic Neurocircuitry” (23115103), Grant-in-Aid for Scientific Research (B) (23300142) (to Dr. Yamanaka), a Japan Society for Promotion of Science postdoctoral fellowship (to Dr. Tsun-ematsu), and NIH R01 NS057464 (to Dr. Kilduff). We thank Dr. René Hen for kindly providing TetO Htr1a knock-in mice. We thank Claire Saito and Keiko Nishimura for technical assistance.