Abstract
In mammals, the central circadian clock is located in the suprachiasmatic nucleus (SCN) of the hypothalamus, which transmits circadian information to other brain regions and regulates the timing of sleep and wakefulness. Neurons in the lateral hypothalamus (LH), particularly those producing melanin-concentrating hormone (MCH) and orexin, are key regulators of sleep and wakefulness. Although the SCN receives nonphotic input from other brain regions, the mechanisms of functional input from the LH to the SCN remain poorly understood. Here, we show that orexin and MCH peptides influence the circadian period within the SCN of both sexes. When these neurons are ablated, the circadian behavioral rhythms are lengthened under constant darkness. Using anterograde and retrograde tracing, we found that orexin and MCH neurons project to the SCN. Furthermore, the application of these peptides to cultured SCN slices shortened circadian rhythms and reduced intracellular cAMP levels. Additionally, pharmacological reduction of intracellular cAMP levels similarly shortened the circadian period in SCN slices. These findings suggest that orexin and MCH peptides from the LH contribute to the modulation of the circadian period in the SCN.
Significance Statement
In mammals, the central circadian clock is located in the suprachiasmatic nucleus (SCN) of the hypothalamus, where it regulates circadian rhythms, including sleep and wakefulness. The SCN receives both neuronal and humoral input signals from external brain regions, which can modify circadian rhythms within the SCN. While several brain regions that project to the SCN have been anatomically identified, the specific regions, neuronal cell types, and neurotransmitters that influence SCN circadian rhythms remain largely uncharacterized. This study identifies two neuronal populations within the lateral hypothalamus that project to the SCN and modulate the circadian period.
Introduction
Circadian rhythms are endogenous physiological and behavioral rhythms with a cycle length of ∼24 h, thought to be driven by a transcription–translation negative feedback loop (TTFL) involving several clock genes. In mammals, transcription factors such as BMAL1 and CLOCK promote the transcription of clock genes, including period (Per) and cryptochrome (Cry) whose protein products subsequently inhibit their transcription (Reppert and Weaver, 2002). The TTFL mechanism operates in nearly all cell types, including fibroblasts, astrocytes, and neurons, to generate cellular circadian rhythms (Takahashi, 2017).
In mammals, daily physiological and behavioral rhythms are orchestrated by the central circadian clock located in the suprachiasmatic nucleus (SCN) of the hypothalamus (Moore and Eichler, 1972; Stephan and Zucker, 1972). Individual SCN cells exhibit circadian rhythms in both clock gene expression and spontaneous electrical activity (Welsh et al., 1995; Yamaguchi et al., 2003). However, in dispersed cell cultures lacking intercellular networks, the period and amplitude of these rhythms vary among cells, suggesting that the SCN comprises heterogeneous cell populations (Herzog et al., 2004; Honma et al., 2004). At the circuit level, nearly all SCN neurons are GABAergic and express various neuropeptides, including arginine vasopressin (AVP), vasoactive intestinal peptide (VIP), and gastrin-releasing peptide (GRP; Abrahamson and Moore, 2001). These neuropeptides serve as synchronizers, generating coherent cellular circadian rhythms within the SCN (Aton et al., 2005; Maywood et al., 2006, 2011; Ono et al., 2016).
The SCN receives direct retinal inputs via the retinohypothalamic tract and synchronizes circadian rhythms with environmental light/dark (LD) cycles (Hattar et al., 2002; Lucas et al., 2001). The SCN subsequently transmits circadian information to regulate various physiological functions, including sleep/wakefulness, corticosterone secretion, and osmolality (Gizowski et al., 2016; Jones et al., 2021; Ono et al., 2020; Paul et al., 2020). Anatomical studies indicate that several neuronal populations outside the SCN provide inputs to the nucleus, including the intergeniculate leaflet, median raphe nuclei, lateral septal nucleus, paraventricular thalamus, preoptic areas, ventromedial hypothalamus, dorsomedial hypothalamus, and LH (Moga and Moore, 1997; Pickard, 1982). Some of these populations are implicated in regulating circadian rhythms and modulating SCN light responses (Janik and Mrosovsky, 1994; Johnson et al., 1988; Rea et al., 1994; Selim et al., 1993).
In addition, midbrain dopamine input to the SCN accelerates circadian entrainment (Grippo et al., 2017), and diffusible signals from the choroid plexus modulate the circadian period in the SCN (Myung et al., 2018). Furthermore, hypothalamic orexin (hypocretin) with neuropeptide Y has been suggested to modulate circadian rhythms in the SCN (Belle et al., 2014). Collectively, these reports suggest that afferent projections to the SCN play a role in modulating circadian rhythms.
In the LH, several neuronal populations regulate sleep and wakefulness. Among these, melanin-concentrating hormone (MCH)-producing neurons and orexin-producing neurons are located in the LH and are critical for sleep and wake regulation, respectively (Chemelli et al., 1999; Sakurai et al., 1998; Shimada et al., 1998; Verret et al., 2003). Previous studies have shown that orexin-deficient mice exhibit fragmented sleep–wake cycles, indicating the importance of orexin in maintenance of wakefulness (Chemelli et al., 1999; Sakurai et al., 1998). Similarly, mice lacking MCH show alterations in sleep patterns, including changes in REM sleep duration and distribution (Shimada et al., 1998; Verret et al., 2003). Knockout mice for orexin receptors (OX1R and OX2R) or MCH receptors (MCHR1) display similar phenotypes (Adamantidis et al., 2008; Mieda et al., 2011). These findings suggest a reciprocal regulation between the circadian clock in the SCN and the sleep–wake centers in the brain. However, the specific mechanisms through which these neurons influence the SCN remain unclear. Therefore, we aimed to investigate the functional roles of orexin and MCH neurons in the LH on circadian rhythms in the SCN. Using genetic, anatomical, and imaging methods, we explored the input mechanisms of these neurons in regulating circadian rhythms in the SCN.
Materials and Methods
Animals
We used Orexin-tTA and MCH-tTA mice, which express mammalianized tetracycline-controlled transcriptional activator (tTA) in orexin and MCH neurons, respectively (Tabuchi et al., 2014; Tsunematsu et al., 2014). We also used TetO DTA mice that express diphtheria toxinA (DTA) under the control of the Tet-off system [B6.Cg-Tg (TetO DTA) 1Gfi/j, The Jackson Laboratory]. Doxycycline (DOX)-containing chow was prepared by mixing 10% DOX powder (Kyoritsu Seiyaku) with normal chow (Labo MR Stock, Nosan) at a final concentration of 100 mg/kg, as previously described (Hung et al., 2020). We previously generated mice with ablated orexin, MCH, or both types of neurons by inducing tTA-driven DTA expression (Hung et al., 2020). This method ablates specific neurons in the absence of DOX treatment (Fig. 1A).
Ablation of orexin and MCH neurons shows a lengthening of circadian behavioral rhythms. A, Schematic drawing of the tTA-driven DTA expression system. B, C, Schematic drawing of the coronal brain slice and fluorescence images in the LH. Scale bar, 200 µm. D, Representative examples of locomotor activity rhythms for control and orexin neuron-ablated (Orexin Nx) mice (left), the circadian period under DD calculated by periodogram (middle; control, n = 10; male 4/female 6; orexin Nx, n = 7; male 3/female 4), and daily locomotor activity profiles of these mice under LD and DD (right). E, Representative examples of locomotor activity rhythms of control and MCH neurons ablated (MCH Nx) mice (left), the circadian period under DD calculated by periodogram (middle; control, n = 8; male 5/female 3; MCH Nx, n = 7; male 4/female 3), and daily locomotor activity profiles of these mice under LD and DD (right). F, Representative examples of locomotor activity rhythms of control and orexin and MCH neurons ablated (Orexin-MCH Nx) mice (left), the circadian period under DD calculated by periodogram (middle; control, n = 11; male 5/female 6; orexin and MCH Nx, n = 10; male 4/female 6; *p < 0.05; Student's t test), and daily locomotor activity profiles of these mice under LD and DD (right).
We used Orexin-tTA; TetO DTA, MCH-tTA; TetO DTA, and Orexin-tTA; MCH-tTA; TetO DTA mice, with DOX-containing chow administered prior to fertilization. From 10 to 12 weeks of age, DOX-containing chow was replaced with DOX-free chow for 4 weeks to induce neuronal ablation. Activity measurements were initiated when the mice were at 15–17 weeks old, using both male and female mice. For bioluminescence imaging of SCN slices, we used PER2::LUC mice carrying a PER2 fusion luciferase reporter (Yoo et al., 2004). For fluorescence imaging of SCN slices, we used Vgat-Cre mice. All mice had a C57BL/6J background and were bred and reared in the animal facility at Nagoya University under controlled environmental conditions (lights on from 8:00 to 20:00, light intensity ∼100–300 lux at the bottom of the cage, ambient temperature 23 ± 2°C, and humidity was 60 ± 10%). Male and female mice (aged 2–5 months) were used in all experiments as previously described (Ono et al., 2020). All experimental protocols were approved by the Institutional Animal Care and Use Committees of the Research Institute of Environmental Medicine, Nagoya University (Approval Numbers R230011 and R230013).
Locomotor activity recordings
Locomotor activity was measured using a passive infrared sensor that detects changes in thermal radiation from the animals due to movement, as previously described (Ono et al., 2016). Mice were individually housed in polycarbonate cages within a light-tight air–conditioned box. Locomotor activity data were recorded every minute using the ClockLab software (Actimetrics). The free-running period was calculated using a χ2 periodogram analysis using ClockLab.
Anterograde tracing
A tTA-dependent AAV carrying the synaptophysin-EGFP gene (AAV-TetO(3G)-Synaptophysin-EGFP-WPRE (Sr10); 600 nl, 4.9 × 1012 copies/ml) was unilaterally injected into the LH (AP, −1.4 mm; ML, 0.8 mm; DV, −5.0 mm from the bregma/skull surface) of Orexin-tTA or MCH-tTA mice. Three weeks after injection, animals were perfused with saline and fixed in 10% formalin (066-03847; Fujifilm Wako Pure Chemical Industries). The brains were removed, immersed in formalin overnight, and then placed in 30% sucrose in phosphate-buffered saline (PBS) for over 36 h. The brains were then frozen in embedding solution (4583, Sakura Finetek) at −80°C and sectioned serially at a 40 μm thickness using a cryostat (CM3050-S, Leica Microsystems K.K.). Brain sections were collected every 160 µm and stained with DAPI. Fluorescent images were captured using a microscope (BZ-X710, Keyence, or IX71, Olympus).
Retrograde tracing
C57BL/6 mice were stereotaxically injected with 50 nl of red fluorescent RetroBeads (Lumafluor) into the SCN (AP, −0.2 mm; ML, ±0.1 mm; DV, −5.8 mm from the bregma/skull surface). Four days after the RetroBeads injection, mice were anesthetized with isoflurane and perfused with saline, followed by a 10% formalin solution. The subsequent procedure was carried out as previously described.
Immunohistochemistry
After the brain was fixed and immersed in a 30% sucrose solution for at least 2 d, 40 µm coronal sections were prepared. Immunostaining was performed as previously described (Ono et al., 2020). Brain sections were incubated in blocking buffer (1% bovine serum albumin and 0.25% Triton X-100 in PBS: PBS-BX) and then incubated with primary antibodies (anti-orexin-A goat antibody at 1:1,000, Santa Cruz Biotechnology; anti-MCH rabbit antibody at 1:1,000, Sigma-Aldrich) overnight at 4°C. The sections were washed with blocking buffer and incubated with secondary antibodies (CF488- or CF647-conjugated anti-goat or anti-rabbit antibody at 1:1,000, Biotium) overnight at 4°C. The sections were then mounted on glass slides and examined under a fluorescence microscope (BZ-X710, Keyence). Cell counting was performed using the ImageJ software.
Fluorescence in situ hybridization (FISH)
Mice were anesthetized with isoflurane and perfused with saline followed by a 10% formalin solution between ZT5 and ZT8. The brain was removed and immersed in formalin overnight and then in 30% sucrose in PBS for over 36 h. After freezing the brain in an embedding solution at −80°C, they were sectioned at thickness of 20 µm. The sections were collected every 160 µm and mounted on glass slides for the subsequent steps.
The slides were treated using the RNAscope multiplex fluorescent v2 kit (323100, Advanced Cell Diagnostics) following the standard protocol. The slides were incubated in PBS at room temperature (RT) for 5 min, followed by incubation at 60°C for 1 h in a HybEZ hybridization oven (Advanced Cell Diagnostics). After drying, the slides were immersed in 10% formalin at RT for 1 h, followed by ethanol drying. Next, the slides were incubated with hydrogen peroxide at RT for 10 min, followed by boiling with target retrieval reagent at 98–102°C for 5 min, and treated with protease digestion (Protease III) at 40°C in the HybEZ oven for 30 min.
Subsequently, the slides were incubated at 40°C with target probes for 2 h and washed with wash buffer twice (2 min each) at RT. The slides were then incubated in AMP1 buffer for 30 min, AMP2 buffer for 30 min, and AMP3 buffer for 15 min, with washing in between each step. For fluorescence detection, RNA probes were conjugated to Cy3 or Cy5 using the TSA Plus Fluorescence System (Perkin Elmer). After fluorescent dye hybridization, the slides were immersed twice in PBS for 5 min, followed by incubation in a 500× dilution of NeuroTrace (N21480, Thermo Fisher Scientific) at RT for 20 min. After discarding the solution, the slides were washed for 3 min in PBS-BX and then for 3 min in PBS and stained with DAPI. Finally, the sections were mounted using ProLong Gold Antifade Mountant (Thermo Fisher Scientific). Fluorescent images were obtained using a microscope (BZ-X710, Keyence, or IX71, Olympus). The following RNA probes were used: OX1R (Mm-Hcrtr1-C2, 466631-C2), OX2R (Mm-Hcrtr2, 460881), and MCHR1 (Mm-Mchr1-C3, 317491-C3) from Advanced Cell Diagnostics.
Bioluminescence measurement of the SCN slice
We prepared 300-µm-thick coronal SCN slices from adult PER2::LUC mice using a Microslicer (DTK-1000, Dosaka EM). The animals were killed to harvest the SCN slices during the light period. The SCN slices were dissected from the midrostrocaudal region and cultured on Millicell-CM culture inserts (Merck Millipore). The culture conditions were the same as previously described (Ono et al., 2013). Briefly, the slices were cultured in the air at 36.5°C with 1.0 ml of Dulbecco's modified Eagle's medium (Invitrogen) containing 0.1 mM D-luciferin K and a 5% supplement solution. The bioluminescence of PER2::LUC in the SCN slices was measured using a luminometer equipped with a photomultiplier tube (PMT) at 10 min intervals, with an exposure time of 1 min. Four to five days after starting bioluminescence recordings, we applied water or DMSO (vehicle), orexin-A (4346-s, Peptide International; final concentration 100 nM), MCH (4369-v, Peptide International; final concentration 100 nM), orexin-A and MCH (final concentration 100 nM each), or MDL-12,330A (M182, Sigma-Aldrich; final concentration 5.0 µM) to the culture medium and measured bioluminescence again. The circadian period of the PER2::LUC rhythms in the SCN was calculated using a χ2 periodogram (ClockLab, Actimetrics). The normalized amplitude was calculated as the peak–trough difference after peptide administration divided by the peak–trough difference before administration.
Fluorescence measurement of the SCN slice
Cre-dependent AAV carrying GCaMP8s [AAV-CAG-flex-GCaMP8s (Sr9), 100 nl, 7.3 × 1012 copies/ml] or cAMPinG1 [AAV-CAG-flex-cAMPinG1-NE (Sr10), 100 nl, 5.9 × 1013 copies/ml] was injected into the SCN (AP, −0.2 mm; ML, ±0.1 mm; DV, −5.8 mm from the bregma/skull surface) of Vgat-Cre mice. Three to four weeks after the stereotaxic AAV injection, the mice were deeply anesthetized using isoflurane (Fujifilm Wako Pure Chemical Industries) and decapitated. The brains were rapidly isolated and chilled in ice-cold bubbled (95% O2 and 5% CO2) cutting solution [in mM: 110 K-gluconate, 15 KCl, 0.05 EGTA, 5 HEPES, 26.2 NaHCO3, 25 glucose, 3.3 MgCl2, 0.0015 3-((+/−)-2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid]. Coronal brain sections (300 µm thick) were made using a vibratome (VT-1200S, Leica). The slices were incubated in a bubbled (95% O2 and 5% CO2) artificial cerebrospinal fluid (aCSF; in mM: 124 NaCl, 3 KCl, 2 MgCl2, 2 CaCl2, 1.23 NaH2PO4.2H2O, 26 NaHCO3, 25 glucose) at 35°C for 30 min followed by another 30 min incubation at RT in the same solution.
After incubation, the brain slices were placed in a recording chamber (RC-26G, Warner Instruments) and perfused with bubbled (95% O2 and 5% CO2) aCSF at a flow rate of 1.5 ml/min using a peristaltic pump (Miniplus3, Gilson). An infrared camera (C3077-78, Hamamatsu Photonics) was integrated with in an upright fluorescence microscope (BX51WI, Olympus), equipped with an electron-multiplying charge–coupled device camera (Evolve 512 Delta, PhotoMetrics) and a monitor to visualize images. A light source (Spectra Light Engine, Lumencor, Beaverton) was used for the excitation of GCaMP8s or cAMPinG1. SCN-containing slices were illuminated with blue light (475 ± 17.5 nm wavelength, 2.18 mW) through the objective lens of a fluorescence microscope (40×, LUMPlanFL N, Olympus). The fluorescence intensity of GCaMP8s or cAMPinG1 was recorded using the MetaMorph software (Molecular Devices) at a rate of 1 Hz with an exposure time of 100 ms. Each substance was dissolved in aCSF according to the target concentration and applied for 2 min. After each application, the signal was allowed to return to basal levels. As a control, aCSF was applied to each slice.
Analysis of calcium and cAMP imaging data
Fluorescence imaging data were analyzed by defining region of interest on GCaMP8s or cAMPinG1-expressing SCN neurons, and the z-score was calculated from baseline data before the application of substances. Statistical analyses were performed using the paired Wilcoxon signed-rank test to compare the average intensities measured 2 min before and during substance application.
Analysis of single-cell RNA sequencing data
The scRNA-seq data (filtered HDF5 binary files, GSE167927) from SCN were used as previously reported (Morris et al., 2021). The data were imported into R version 4.3.1 and primarily analyzed using Seurat V4.2.0. Briefly, we retained the threshold parameters set by the original authors including at least 200 unique transcription profiles and <10% of mitochondrial transcripts. After log normalization, the count data were scaled, and the replicates were merged before integrating samples from different time points. Clustering and UMAP visualization were performed using 25 principal components and a resolution of 0.5 for the shared nearest neighbor clustering algorithm. Fourteen clusters were identified and used for downstream differential expression analysis.
Statistical analysis
Statistical analyses were conducted using the Prism 8.0 (GraphPad) or OriginPro 2020 software. The following tests were employed: Student's t test, two-way repeated–measure ANOVA, mixed-effect model with post hoc Sidak's multiple-comparison test, and paired Wilcoxon signed-rank test.
Results
Ablation of orexin and MCH neurons shows a lengthening of circadian behavioral rhythms
We used Orexin-tTA; TetO DTA, MCH-tTA; TetO DTA, or Orexin-tTA; MCH-tTA; TetO DTA mice to selectively ablate orexin, MCH, or both neuron populations by removing DOX (Fig. 1A). We confirmed the ablation of orexin, MCH, or both neurons in the LH (Fig. 1B,C; Hung et al., 2020). After reintroducing DOX-containing chow, locomotor activity was recorded under LD and constant darkness (DD) conditions using an infrared sensor, as previously described (Ono et al., 2020). Control mice were continuously fed DOX-containing chow throughout the experiment.
Orexin neuron-ablated (Orexin Nx) mice showed no significant changes in the free-running period or the daily patterns of circadian behavioral rhythms compared with littermate controls (Fig. 1D). Similarly, MCH neuron-ablated (MCH Nx) mice did not show significant changes in the free-running period but exhibited higher activity levels than littermate controls (Fig. 1E). In contrast, mice with ablation of both orexin and MCH neurons (orexin-MCH Nx) exhibited longer free-running periods and reduced activity levels compared with littermate controls (Fig. 1F). These results suggest that orexin and MCH neurons may contribute to the shortening of the period of circadian behavioral rhythms.
Presynapse of orexin and MCH neurons in the LH are located in the SCN and peri-SCN
We observed that orexin-MCH Nx mice exhibited lengthened circadian behavioral rhythms (Fig. 1). Therefore, we hypothesized that both orexin and MCH neurons serve as inputs to the SCN. To identify neuronal inputs from these neurons to the SCN, we performed anterograde tracing. We injected the LH of Orexin-tTA or MCH-tTA mice with an adeno-associated virus (AAV) carrying a tTA-dependent expression of synaptophysin (a presynaptic marker) fused to an enhanced green fluorescent protein [AAV-TetO (3G)-synaptophysin-EGFP; Fig. 2A,B). Histochemical analysis revealed that presynapses of both orexin and MCH neurons in the LH were located in the SCN and peri-SCN (Fig. 2C,D). Moreover, the innervation of these neurons was predominantly observed on the ipsilateral side of the brain. To confirm that synaptophysin-EGFP was specifically expressed in orexin or MCH neurons, we injected the AAV into the LH of wild-type (WT) mice (Fig. 2E). No expression was observed in the LH or SCN of WT mice (Fig. 2F,G). These results suggest that orexin and MCH neurons in the LH provide inputs to the SCN and peri-SCN.
Anterograde and retrograde identification of the projections of orexin and MCH neurons in the LH to the SCN and peri-SCN. A, Schematic drawing of anterograde tracing of orexin neurons in the LH. EGFP fused synaptophysin were expressed in orexin or MCH neurons in the LH (injection site). B, Representative example of fluorescent image of synaptophysin-EGFP in the LH. Scale bar, 1,000 µm. C, Representative example of fluorescent images of synaptophysin-EGFP expressed in orexin neurons in the anterior, middle, and posterior SCN (left). Magnified images in the dorsal and ventral areas of the SCN are described on the right. D, Representative example of fluorescent images of synaptophysin-EGFP expressed in MCH neurons in the anterior, middle, and posterior SCN (left). Magnified images in the dorsal and ventral areas of the SCN are described on the right. Scale bar, 100 µm (left) and 10 µm (right). E, Schematic drawing of anterograde tracing in WT mice. F, Representative example of fluorescent image of synaptophysin-EGFP in the LH of WT mice. There is no expression in the LH (injection site). Scale bar, 1,000 µm. G, Representative example of fluorescent images of synaptophysin-EGFP expressed in WT SCN. Magnified images in the dorsal and ventral areas of the SCN are described on the right. Scale bar, 100 µm (left) and 10 µm (middle and right). H, Schematic drawing of retrograde tracing from the SCN. Retrobeads were injected in the SCN. I, Fluorescent image of retrobeads injected in the SCN. Scale bar, 100 µm. J, Representative fluorescent image of retrobeads in the LH (magenta, retrobeads; yellow, orexin; light blue, MCH). Scale bar, 100 µm (left) and 10 µm (right). K, The percentage of retrobead-positive neurons in the orexin or MCH-positive neurons in the LH (n = 3).
Retrograde identification of the projections of orexin and MCH neurons in the SCN
To gain further insight into the neuronal inputs from orexin and MCH neurons to the SCN, we performed retrograde tracing. We injected retrogradely transported red fluorescent beads (RetroBeads) into the SCN of WT mice to determine whether these neurons project to the SCN (Fig. 2H,I). Four days after the injection, we prepared brain slices and used immunohistochemistry to assess whether orexin and MCH neurons exhibited red fluorescence. As expected, fluorescent signals were observed in the LH, with ∼20% of both orexin and MCH neurons showing overlap (Fig. 2J,K). These results collectively indicate that orexin and MCH neurons in the LH innervate the SCN.
Orexin and MCH receptors expressed in the SCN and peri-SCN
We identified neuronal inputs from orexin and MCH neurons in the SCN (Fig. 2). It is likely that these neurons release orexin or MCH, which are then received by SCN neurons. Mice have two types of orexin receptors (orexin receptor-1 and orexin receptor-2: OX1R and OX2R, respectively) and one type of MCH receptor (MCH receptor 1, MCHR1; Chambers et al., 1999; Saito et al., 1999; Sakurai et al., 1998). To determine whether OX1R1, OX2R, and MCHR1 are expressed in the SCN, we performed FISH with fluorescent Nissl staining. We observed that all three receptors were expressed in both the SCN and peri-SCN (Fig. 3A,B). MCHR1 appeared more predominantly expressed in the dorsal SCN than in the ventral SCN. OX1R and OX2R were observed throughout the SCN without any distinct distribution patterns. These results indicate that orexin and MCH receptors are expressed in the SCN.
Orexin and MCH receptors expressed in the SCN and peri-SCN. A, Representative FISH with fluorescent Nissl staining images in the SCN (gray, Nissl; light blue, orexin receptor 1; magenta, MCH receptor 1). B, Representative FISH with fluorescent Nissl staining images in the SCN (gray, Nissl; light blue, orexin receptor 2; magenta, MCH receptor 1). Scale bar, 100 µm (left) and 10 µm (right). C–E, Results of single-cell RNA sequencing using data from Morris et al. Avp, Vip, Grp, or Nms expression patterns in OX1R-, OX2R-, or MCHR1-positive (>0.2, green) or OX1R-, OX2R-, or MCHR1-negative (<0.2, red) cells in the SCN are shown in each panel.
Previously, two groups have performed single-cell RNA sequencing on SCN tissues (Morris et al., 2021; Wen et al., 2020). To further identify which SCN neurons express OX1R, OX2R, and MCHR1, we re-analyzed these data. We selected the Morris et al. dataset for our analysis due to its higher RNA read count as compared with the Wen et al. dataset. Almost all of the OX1R-, OX2R-, or MCHR1-positive cells in the SCN expressed the Avp gene (OX1R, 91.9%; OX2R, 95.9%; and MCHR1, 86.7%), and ∼40% of these receptor-positive cells expressed Vip (OX1R, 40.3%; OX2R, 46.3%; and MCHR1, 39.6%; Fig. 3C–E). In contrast, only a small fraction of these receptor-positive cells expressed Grp or Nms. These results indicate that orexin and MCH receptors are predominantly expressed in Avp-positive cells in the SCN.
Orexin and MCH peptides shorten circadian PER2::LUC rhythms in the SCN slice
Given that orexin-MCH Nx mice exhibited lengthened circadian behavioral rhythms and that neuronal inputs from orexin and MCH neurons were detected in the SCN, we hypothesized that orexin and MCH peptides shorten circadian rhythms in the SCN. Indeed, several studies have shown that circadian rhythms in SCN slices tend to be longer than behavioral rhythms under DD conditions (Hamnett et al., 2021; Myung et al., 2012; Nakamura et al., 2015; Tso et al., 2017; van der Vinne et al., 2018), suggesting that certain factors in vivo contribute to the shortening of the period of circadian rhythms in the SCN. To test whether orexin and MCH peptides shorten circadian rhythms in the SCN, we prepared SCN slices from PER2::LUC reporter mice (a fusion protein of PER2 and firefly luciferase) and monitored their bioluminescence using a PMT, as previously reported (Ono et al., 2013). Four to five days into recording PER2::LUC rhythms, we applied the vehicle, orexin-A (100 nM), MCH (100 nM), or a combination of both peptides to the culture medium. We found that the circadian period of PER2::LUC rhythms in SCN slices was longer than 24 h before peptide application. However, following the application of orexin-A and MCH, the circadian period in the SCN slices was shortened (Fig. 4A,B). Furthermore, a single application of either orexin-A or MCH alone also resulted in the shortening of circadian rhythms in SCN slices. The amplitude of circadian rhythms remained unchanged across all conditions following peptide application (Fig. 4C). These results suggest that orexin and MCH peptides are capable of shortening circadian rhythms in SCN slices.
Orexin and MCH peptides shorten circadian PER2::LUC rhythms in the SCN slice. A, Representative double-plotted PER2::LUC rhythms in the SCN slice. Stars in each panel indicate the timing of peptide or vehicle application. B, The circadian period of PER2::LUC rhythms in the SCN before and after application of peptides (vehicle, n = 11; orexin-A, n = 11; MCH, n = 13; orexin-A and MCH, n = 12; **p < 0.01; mixed-effect model with post hoc Sidak's multiple-comparison test). C, The normalized amplitude of PER2::LUC rhythms in the SCN slice after peptide application. The amplitudes were normalized with the difference in peak–trough before the administration as 1.0.
Orexin and MCH peptides modulate Ca2+ and cAMP dynamics in the SCN slice
Orexin and MCH receptors are G-protein–coupled receptors capable of modulating intracellular Ca2+ and cAMP dynamics. It has been suggested that intracellular Ca2+ and cAMP play a role as inputs to the TTFL (Hastings et al., 2008; Ono et al., 2023). Given that ablation of orexin and MCH neurons lengthened circadian behavioral rhythms (Fig. 1) and that the application of orexin and/or MCH peptides to SCN slices shortened circadian PER2::LUC rhythms (Fig. 4), these peptides may influence Ca2+ or cAMP dynamics in the SCN.
To investigate the effects of orexin and MCH on Ca2+ signaling in the SCN, we expressed a fluorescent Ca2+ indicator (GCaMP8s) in the SCN using AAV and performed time-lapse fluorescence imaging in SCN slices from mice killed during the day or night. We applied MCH, orexin-A, glutamate, and GABA to aCSF and measured the fluorescence intensity before and after the application of each substance. During the daytime, Ca2+ levels increased in 63.4 and 43.9% of SCN neurons and decreased in 19.5 and 29.3% of SCN neurons following orexin (100 nM) or MCH (100 nM) application, respectively (Fig. 5A–D). At night, Ca2+ levels increased in 10.9 and 0% of SCN neurons and decreased in 76.4 and 84.4% of SCN neurons after orexin and MCH application, respectively (Fig. 5F–I). All SCN neurons exhibited increased Ca2+ levels after glutamate (100 µM) application and decreased Ca2+ levels after GABA (100 µM) application (Fig. 5E,J). These results indicate that orexin and MCH peptides modulate Ca2+ dynamics in SCN neurons.
Orexin and MCH peptides modulate Ca2+ dynamics in the SCN neurons. A, F, Fluorescence images of GCaMP8s expressing Vgat neurons in the SCN. Scale bar, 20 µm. B, G, The intensity of fluorescence changes obtained from SCN neurons before, during, and after treatment with orexin-A or MCH peptides. Green, magenta, and gray lines indicate neurons with increased, decreased, and unchanged fluorescence intensity upon peptide administration, respectively. Fluorescence intensity is expressed as z-score. C, H, Pie chart showing the percentage of cells that exhibited increase, decrease, or no change in fluorescence intensity during peptide administration. The paired Wilcoxon signed-rank test (p < 0.05) was used to compare the average intensity 2 min before and during peptide application. D, I, Difference in fluorescence intensity between orexin-A and MCH administrations. Each symbol represents one neuron (n = 41 from 3 mice at ZT 4–10; n = 55 from 4 mice at ZT 14–20). E, J, Fluorescence intensity changes in SCN neurons before, during, and after treatment with aCSF, glutamate, and GABA during the subjective day and night. A–E for day and F–J for night.
Next, to examine the effects of orexin and MCH on cAMP signaling in the SCN, we expressed a fluorescent cAMP indicator (cAMPinG1) in the SCN using AAV and performed time-lapse fluorescence imaging of SCN slices from mice killed during the day or night (Yokoyama et al., 2024). Interestingly, almost all SCN neurons exhibited decreased cAMP levels following the application of orexin-A as well as MCH during both day and night (Fig. 6A–J). Furthermore, the application of GABA decreased cAMP levels, and glutamate also decreased cAMP levels (Fig. 6E,J). These results indicate that orexin and MCH peptides decreased cAMP levels in SCN neurons.
Orexin and MCH peptides modulate cAMP dynamics in the SCN neurons. A, F, Fluorescence images of cAMPinG1-expressing Vgat neurons in the SCN. Scale bar, 20 µm. B, G, Fluorescence intensity changes from SCN neurons before, during, and after treatment with orexin-A or MCH peptides. Green, magenta, and gray lines indicate neurons with decreased or unchanged fluorescence intensity upon peptide administration, respectively. Fluorescence intensity is expressed as z-score. C, H, Pie chart showing the percentage of cells that exhibited decrease or unchanged fluorescence intensity during peptide administration. Paired Wilcoxon signed-rank test (p < 0.05) was used to compare average intensity 2 min before and during peptide application. D, I, Difference in fluorescence intensity between orexin-A and MCH administrations. Each symbol represents one neuron (n = 51 from 3 mice at ZT 4–10; n = 53 from 4 mice at ZT 14–20). E, J, Fluorescence intensity changes in SCN neurons before, during, and after treatment with aCSF, glutamate, and GABA during the subjective day and night. A–E for day and F–J for night.
Pharmacological reduction of cAMP levels shortens circadian PER2::LUC rhythms in the SCN slice
Since the application of orexin-A or MCH peptides shortened the period of the circadian PER2::LUC rhythms and reduced cAMP levels in the SCN slices (Figs. 4, 6), it is possible that the reduction in cAMP levels is associated with the shortening of circadian rhythms in the SCN. To test this, we measured PER2::LUC rhythms in the SCN slices and applied MDL-12,330A (an adenylate cyclase inhibitor) or vehicle to the culture medium. As expected, the period of circadian PER2::LUC rhythms in SCN slices was shortened during the application of MDL-12,330A (Fig. 7A,B). These results indicate that a reduction in cAMP levels leads to the shortening of circadian rhythms in SCN slices.
Pharmacological reduction of cAMP levels shortens circadian PER2::LUC rhythms in the SCN slice. A, Representative double-plotted PER2::LUC rhythms in the SCN slice. Stars in each panel indicate the timing of vehicle or MDL-12,330A (5.0 µM) application. B, Circadian period of PER2::LUC rhythms in SCN slices before and after MDL application (vehicle, n = 5; MDL, n = 5; **p < 0.01; mixed-effect model with post hoc Sidak's multiple-comparison test).
Discussion
We identified that orexin and MCH neurons innervate the SCN through both anterograde and retrograde tracing methods. Ablation of these neurons, using genetically encoded DTA, resulted in a lengthening of circadian behavioral rhythms. Moreover, the application of orexin or MCH peptides to cultured SCN slices shortened the period of PER2::LUC rhythms. These peptides also modulated Ca2+ and cAMP dynamics in SCN slices. Based on these results, we conclude that orexin and MCH peptides shorten the circadian period in the SCN (Fig. 8).
A hypothetical model of the regulation of the circadian period in the SCN by orexin and MCH neurons. Orexin and MCH neurons in the LH regulate wakefulness and sleep, respectively. Neuronal activity of orexin neurons is typically high during wakefulness and intermediate during NREM sleep, whereas the neuronal activity of MCH neurons is high during REM sleep and intermediate during wakefulness. Orexin and MCH peptides are released from neuronal terminals located in the SCN and peri-SCN, where they reduce cAMP levels and shorten circadian rhythms in the SCN.
Our experiments demonstrated that the application of orexin-A, as well as MCH peptides, reduced intracellular cAMP levels across almost all SCN neurons (Fig. 6). Since not all SCN neurons express receptors for orexin (OX1R, OX2R) or MCH (MCHR1), the reduction of cAMP levels by orexin-A and MCH may be mediated by indirect mechanisms, possibly through GABAB receptors. Additionally, the high density of orexin and MCH presynapses and their receptors in the peri-SCN suggest that circadian rhythms in the SCN could be modulated indirectly by peri-SCN neurons.
Intracellular Ca2+ and cAMP are second messengers that have been suggested to be involved in the input and output of the molecular circadian clock in the SCN (Hastings et al., 2008; Lundkvist et al., 2005; O'Neill et al., 2008). Indeed, pharmacological induction of cAMP in U2OS cells has been shown to lengthen the circadian period, and optical induction of cAMP in the SCN induces phase delay shifts (Jagannath et al., 2021; Ono et al., 2023). Thus, the reduction in cAMP levels following orexin-A and MCH application likely contributes to the observed shortening of the circadian rhythm (Fig. 4). This hypothesis is further supported by our findings that the pharmacological reduction of cAMP levels also shortened the circadian PER2::LUC rhythms in SCN slices (Fig. 7). Previous studies have reported that caffeine lengthens circadian rhythms both in vivo and ex vivo (Burke et al., 2015; Oike et al., 2011). Therefore, our results suggest that internal factors, such as orexin and MCH, and external factors, such as caffeine, influence the circadian period in the SCN through modulation of cAMP signaling.
Recently, we reported that cAMP is regulated by VIP-associated networks in the SCN (Ono et al., 2023). Furthermore, using AAV encoding the CRISPR-Cas9 system, Hamnet et al. demonstrated that knocking out the DUSP4 gene in the SCN shortened circadian PER2::LUC rhythms in SCN slices (Hamnett et al., 2019). The transcription factor cAMP response element modulator α regulates the expression of DUSP4 (Hofmann et al., 2019). Therefore, DUSP4, a downstream target of the cAMP pathway, may influence the shortening of the circadian period in the SCN following the application of orexin-A and MCH.
Belle et al. showed that orexin-A-containing fibers surround the SCN (Belle et al., 2014). They also demonstrated that orexin-A application suppressed intracellular Ca2+ levels in 47 and 39% of SCN neurons during the day and night, respectively. However, we observed a smaller population (19.5%) of SCN neurons showing a decrease in Ca2+ upon orexin-A administration during the day. This discrepancy may be attributed to differences in calcium recording methods. In their research, they used the Ca2+ fluorescent dye, fura-2, to monitor intracellular Ca2+ levels in SCN cells, whereas we used AAV with Cre-dependent expression of the genetically encoded Ca2+ probe, GCaMP8s, in SCN neurons of Vgat-Cre mice. The fluorescent dye can be incorporated into both neurons and glial cells, but Cre-dependent expression of GCaMP8s occurs exclusively in GABAergic neurons in the SCN. Therefore, the orexin-A-induced decrease in Ca2+ levels observed in their study may involve glial cells. It has also been shown that orexin-A application reduces the firing rate of SCN neurons (Belle et al., 2014; Brown et al., 2008). However, firing and Ca2+ rhythms in the SCN can exhibit different patterns (Ikeda et al., 2003; Ono et al., 2017). It is possible that firing and Ca2+ patterns can be dissociated in response to orexin-A application in the SCN.
Since the presynapses of orexin and MCH neurons in the LH were observed in both the SCN and peri-SCN (Fig. 2), these neuropeptides are likely released into the SCN via volume transmission. Alternatively, orexin or MCH neurons may release these peptides into the cerebrospinal fluid, allowing them to circulate throughout the brain. This diffusible pathway could influence SCN circadian rhythms in vivo. Additionally, orexin neurons coexpress glutamate, dynorphin, galanin, and neurotensin, while MCH neurons coexpress GABA, glutamate, cocaine- and amphetamine-regulated transcript, secretoneurin, and EM66 (Beekly et al., 2023). These neurotransmitters could also affect circadian rhythms in the SCN. Further experiments are required to elucidate the roles of these neurotransmitters in regulating SCN circadian rhythms.
Orexin neurons are critical for the maintenance of wakefulness and exhibit higher activity during wakefulness and intermediate activity during REM sleep (Feng et al., 2020; Hassani et al., 2009; Li et al., 2022). The activity of orexin neurons is also elevated by exposure to novel objects and mechanical stimulation (Gonzalez et al., 2016). In contrast, MCH neurons are primarily involved in sleep regulation. These neurons are activated during REM sleep and wakefulness (Gonzalez et al., 2016; Hassani et al., 2009; Izawa et al., 2019). The activity of MCH neurons also increases during interactions with novel objects or exploratory behaviors (Blanco-Centurion et al., 2019; Gonzalez et al., 2016). These findings suggest that both internal and external factors affect orexin and MCH neuronal activity, thereby modulating circadian rhythms in the SCN.
An example of an external factor influencing circadian rhythms is running wheel activity. Circadian behavioral rhythms exhibit shorter periods when a running wheel is present (Yamada et al., 1988). This observation suggests that factors associated with wheel running can influence the periods of behavioral rhythms. Increased neuronal activity of orexin and MCH neurons due to wheel running may lead to enhanced releases of these neuropeptides, which could act as input signals to the SCN. Reciprocal interactions between the central circadian clock and sleep–wake centers are important to maintain the timing of sleep and wake rhythms and to adapt physiological functions to environmental changes.
Orexin-MCH neuron-ablated mice exhibited lengthened circadian behavioral rhythms under DD, whereas mice with ablation of either orexin or MCH neurons did not show any differences in the period of circadian behavioral rhythms (Fig. 1). These results suggest that both orexin and MCH neurons are necessary to shorten the period of circadian behavior. In SCN slice cultures, input signals from outside the SCN are eliminated, which may explain the lengthening of the circadian period of the PER2::LUC rhythm in the SCN slices as compared with behavioral rhythms. When we applied both orexin-A and MCH peptides to the SCN slice medium, the circadian period of the PER2::LUC rhythm was shortened. Similarly, the application of either orexin-A or MCH alone also shortened the circadian PER2::LUC rhythms (Fig. 4). The discrepancy between in vivo and ex vivo results could be attributed to differences in peptide concentration dynamics between the two environments. Another possible explanation involves the nature of the convergence to 24 h effect observed both inside and outside the SCN. For example, Ode et al. generated various CRY1mutants and expressed them in Cry-deficient mouse embryonic fibroblasts (MEFs; Ode et al., 2017). They showed that mutant CRY1 expressed in Cry-deficient MEFs exhibited a variety of circadian period lengths. However, mutant CRY1 expressed in Cry-deficient mice did not exhibit the same period length as compared with mutant CRY1 expressed in Cry-deficient MEFs. The circadian periods of behavior in mutant CRY1-expressing mice are consolidated at ∼24 h (Ode et al., 2017). This suggests that the mechanism regulating the 24 h period is located either outside the SCN or within the SCN network.
We found that orexin and MCH peptides play a role in modulating the circadian period in the SCN. Several other factors have also been reported to influence circadian rhythms in the SCN. For example, melatonin is secreted at night by the pineal gland in both nocturnal and diurnal animals, and its application can produce a phase advance of behavior in mice (Benloucif and Dubocovich, 1996). In contrast, the activity of serotoninergic neurons is well correlated with arousal levels, and the application of serotonin agonists can induce a phase advance in nocturnal animals during the subjective day but in diurnal animals during the night (Challet, 2007). These two types of factors can be defined as arousal-independent and arousal-dependent, respectively. Orexin and MCH peptides can be categorized as arousal-dependent factors, as their neuronal activities are modulated by sleep and wakefulness (Hassani et al., 2009; Izawa et al., 2019; Li et al., 2022). It would be valuable to investigate the functional role of these peptides in the SCN of diurnal animals.
The amount of locomotor activity in orexin and MCH neuron-ablated mice was reduced compared with that in control mice. These mice exhibited frequent sleep attacks with elevated spectral power in the delta and theta ranges (Hung et al., 2020). This behavioral phenotype likely contributed to the reduction in locomotor activity observed in these animals.
Our findings shed light on the input mechanisms that modulate the circadian period in the SCN. The SCN regulates the 24 h temporal timing of neuronal activity outside the SCN, including orexin and MCH neurons. Neuropeptides released by these neurons in turn influence the circadian period within the SCN. These nonphotic stimuli may play a crucial role in the daily phase adjustment of the circadian clock and the temporal organization of behavior.
Data Availability
All data required to evaluate the conclusions of this study are presented in the paper. Vgat-IRES-Cre and PER2::LUC mice are available from B. Lowell and J. Takahashi under a material transfer agreement with the Beth Israel Deaconess Medical Center and Northwestern University, respectively.
Footnotes
We thank J.S. Takahashi for supplying the PER2::LUC mice and S. Tsukamoto, S. Nasu, H. Noma, and the Center for Animal Research and Education at Nagoya University for their dedication to animal care. This work was supported by the SECOM Science and Technology Foundation, Konica Minolta Science and Technology Foundation, Inamori Foundation, Casio Science Promotion Foundation, LOTTE Foundation, Japan Science and Technology Agency Fusion Oriented Research for Disruptive Science and Technology Program (Grant Number JPMJFR211A, Japan, to D.O.), AMED-PRIME (JP 23gm6510022 to M.S.), and Japan Society for the Promotion of Science KAKENHI (Grant Numbers 21K19255, 21H02526, 21H00307, 21H00422, 20KK0177, and 18H02477 to D.O.).
↵*C.J.H., C.-T.T., and S.M.R. contributed equally to this work.
The authors declare no competing financial interests.
- Correspondence should be addressed to Daisuke Ono at dai-ono{at}riem.nagoya-u.ac.jp.
