Orexin neurons inhibit sleep to promote arousal

Orexin (hypocretin) neurons in the lateral hypothalamus are essential for stabilizing and maintaining wakefulness; loss of these neurons in narcolepsy produces excessive daytime sleepiness, sleep attacks, and wake-state instability. Prevailing models suggest orexin neurons stabilize wake by exciting other arousal-promoting nodes (e.g., locus coeruleus noradrenergic neurons). The authors hypothesized that orexin neurons may also produce wake-promoting, wake-stabilizing effects via inhibition of sleep-promoting neurons in the ventrolateral preoptic nucleus (VLPO), a key sleep-initiating and -maintaining region containing GABA/galanin neurons. Because orexin is excitatory and co-released with glutamate, the team predicted orexin would not directly inhibit VLPO GABA/galanin neurons but instead inhibit them indirectly through activation of local GABAergic interneurons or an inhibitory relay. They propose a polysynaptic pathway from lateral hypothalamus to an intra-VLPO circuit where orexin-responsive VLPO GABAergic interneurons inhibit sleep-promoting VLPO GABA/galanin neurons, thereby promoting arousal. The study aims to test whether activating orexin terminals in VLPO evokes arousal, identify orexin’s postsynaptic targets in VLPO, and define the intra-VLPO synaptic mechanisms that support arousal.

Background work demonstrates: (1) Orexin neurons excite arousal systems (e.g., locus coeruleus) to promote wake. (2) VLPO contains sleep-active GABA/galanin neurons; lesions cause insomnia; stimulating VLPO promotes sleep. (3) Orexin administration into VLPO promotes wake, but postsynaptic targets were unclear. (4) Some arousal inputs inhibit VLPO via GABAergic mechanisms; VLPO GABA/galanin neurons are inhibited by noradrenaline and other wake-promoting signals. The authors build on these findings to test an alternative mechanism where orexin activates local VLPO GABAergic neurons (lacking galanin) that feedforward inhibit VLPO sleep-promoting neurons.

• In vivo optogenetics: Ox-IRES-Cre mice received bilateral AAV-DIO-ChR2 (mCherry/eYFP) injections into the orexin neuron field in lateral hypothalamus; bilateral optical fibers were implanted above VLPO. Real-time sleep-state detection triggered 10 s trains of 5 ms blue light pulses at 1, 5, 10, or 20 Hz (12–15 mW, 473 nm) during NREM or REM sleep between ZT3–ZT9. EEG/EMG were recorded; arousal probability and spectral/EMG responses were analyzed (two-way ANOVA, post-hoc Holm-Šidák, bootstrap CIs). WT mice served as controls.
• In vitro electrophysiology in slices: Whole-cell and cell-attached recordings from VLPO neurons in C57BL/6J (WT), Vgat-IRES-Cre (VLPO GABAergic), Gal-IRES-Cre (VLPO GABA/galanin), and Vgat-Flp::Gal-IRES-Cre mice. Pharmacology tested orexin-A (0.3–1 µM), noradrenaline, carbachol; TTX and bicuculline assessed indirect GABAergic mechanisms. sIPSCs measured at Vh=0 mV; effects quantified via MiniAnalysis, K–S tests, RM-ANOVA.
• CRACM (circuit mapping): Photostimulation of ChR2-expressing terminals or cells using 10 ms blue pulses. For VLPO local circuits, expressed ChR2-eYFP in Vgat neurons (Flp-dependent) and TdTomato in galanin neurons (Cre-dependent) in Vgat-Flp::Gal-IRES-Cre mice, recording from VLPO GABA/galanin neurons. Tested VLPOVgat→VLP OGABA/Gal and VLPOGABA/Gal→VLPOGABA/Gal connectivity, including TTX+4-AP to assess monosynaptic input. Measured opto-evoked IPSC amplitudes/latencies and modulation by orexin-A (1 µM). Peak-scaled non-stationary fluctuation analysis estimated unitary current (i) and number of activated GABA_A channels (N).
• Ox→VLPO connectivity: ChR2 in orexin neurons (Ox-IRES-Cre) with photostimulation of orexin axon terminals in VLPO while recording from VLPO neurons. AMPA antagonist DNQX tested glutamatergic mediation. NA responsiveness and single-cell RT-qPCR for Gal determined cell identity of connected postsynaptic targets.
• Single-cell transcriptomics: Reanalysis of Moffitt et al. 2018 POA dataset including MERFISH spatial data to isolate VLPO neuronal clusters (4174 neurons, 20 clusters). Identified VLPO GABA vs VLPO GABA/Gal clusters, assessed Hcrtr1/Hcrtr2 expression (Wilcoxon rank-sum, Bonferroni correction).
• RNAscope in situ hybridization: Triplex detection of Gal, Slc32a1 (Vgat), and Hcrtr2 mRNAs in VLPO to validate receptor expression patterns.
• Animal details: Adult mice (6–12 weeks), both sexes; multiple transgenic lines noted. Viral constructs: AAV-DIO-ChR2-mCherry/eYFP, AAV-DIO-TdTomato/GFP, AAV-fDIO-ChR2-eYFP. Standard slice preparation solutions and recording parameters provided. Statistical analyses included ANOVAs, paired t-tests, bootstrap resampling; significance at p<0.05.

• In vivo optogenetic activation of orexin terminals in VLPO rapidly triggered arousals from both NREM and REM sleep. NREM: significant main effect of genotype (F(1,24)=22.92, p=0.003); increased arousal probability at 10 and 20 Hz (t=2.98, p=0.019; t=3.73, p=0.004). REM: significant interaction (F(2,12)=7.02, p=0.010); increased arousal probability at 1 and 10 Hz (t=3.82, p=0.002; t=5.16, p<0.001). Arousals were typically brief; NREM stimulation produced longer arousals than REM.
• Orexin-A exerted opposite effects on VLPO neurons in slices: excitation in ~35% (firing +173.29±41.29%; n=21; depolarization +2.80±0.44 mV; n=26) and inhibition in ~60% (firing −84.37±4.66%; n=37; hyperpolarization −4.53±0.47 mV; n=44). Inhibition was blocked by TTX or bicuculline; excitation persisted, indicating indirect GABAergic inhibition vs direct excitation.
• Within VLPO GABAergic neurons: 42.8% were excited by orexin-A (firing +225.61±95.01%; depolarization +2.95±1.06 mV), 50% inhibited (firing −72.92±14.26%; hyperpolarization −3.40±1.13 mV). Cells inhibited by orexin-A were also inhibited by noradrenaline and expressed Gad1/2 and galanin (VLPO GABA/Gal). Cells excited by orexin-A were also excited by noradrenaline and lacked galanin (VLPO GABA).
• In identified VLPO GABA/Gal neurons, orexin-A increased sIPSC frequency in 75% (+58.76±13.29%; n=9), indicating enhanced GABAergic afferent input mediating inhibition.
• Local circuit mapping: Photostimulating VLPO Vgat neurons evoked short-latency GABA_A-mediated oIPSCs in VLPO GABA/Gal neurons (92.7% connected; n=55) that persisted in TTX+4-AP, indicating monosynaptic input; photostimulating VLPO GABA/Gal neurons did not evoke oIPSCs in VLPO GABA/Gal neurons (n=6), indicating lack of reciprocal inhibitory connectivity among them.
• Orexin-A enhanced VLPOVgat→VLPO GABA/Gal oIPSC amplitude in TTX+4-AP (+59.71±26.3%; 56% of cells; n=16). Peak-scaled non-stationary fluctuation analysis showed increased number of activated GABA_A channels (+89.33±33.23%; n=8; p=0.0091) without change in unitary current, consistent with increased GABA release.
• Transcriptomics: Hcrtr2 (Ox2R) expression was absent in all VLPO GABA/Gal clusters and restricted largely to VLPO GABA cluster #1 (489 neurons). Hcrtr1 was not detected in VLPO clusters.
• RNAscope confirmed Hcrtr2 expression in 28.5% of VLPO neurons that were Vgat+ and Gal− (VLPO GABA) and absence in Vgat+ Gal+ (VLPO GABA/Gal).
• Ox→VLPO CRACM: Photostimulation of orexin terminals evoked DNQX-sensitive oEPSCs (glutamatergic, AMPA-mediated) in 24% of recorded VLPO neurons (n=63), with no detectable orexin peptide postsynaptic current during prolonged trains. Connected postsynaptic VLPO neurons were NA-excited and Gal−, identifying them as VLPO GABA neurons.
• Overall model: Orexin neurons (with co-released glutamate) excite Hcrtr2-expressing VLPO GABA interneurons, which in turn monosynaptically inhibit sleep-promoting VLPO GABA/Gal neurons via feedforward inhibition, promoting arousal.

The findings delineate a polysynaptic orexin→VLPO circuit wherein orexin neurons indirectly inhibit sleep-promoting VLPO GABA/galanin neurons through excitation of local VLPO GABA interneurons, thereby promoting arousal. This mechanism complements classical models where orexin directly excites ascending arousal systems and integrates with the flip-flop switch concept governing state stability. By identifying Hcrtr2 expression in a subset of VLPO GABA neurons and showing functional glutamatergic orexin inputs selectively target these neurons, the study explains how orexin can rapidly trigger and stabilize wake by suppressing VLPO sleep-promoting output. In narcolepsy, loss of orexin would reduce excitation of VLPO GABA interneurons, weakening feedforward inhibition of VLPO GABA/Gal neurons and biasing the system toward sleep and instability. The results also suggest VLPO GABA neurons act as a critical interface for multiple wake-promoting inputs to influence VLPO output, potentially generalizable to other afferents (noradrenergic, histaminergic).

This study identifies and functionally validates a feedforward inhibitory circuit by which orexin neurons promote wakefulness: orexin/glutamate inputs excite Hcrtr2-expressing VLPO GABA neurons that inhibit sleep-promoting VLPO GABA/galanin neurons. In vivo activation of orexin terminals in VLPO rapidly elicits arousals from NREM and REM sleep, and in vitro recordings, circuit mapping, and single-cell transcriptomics establish cell-type specificity and synaptic mechanisms. These insights provide an alternative framework for orexin’s role in maintaining consolidated wakefulness and suggest new therapeutic targets within the VLPO network for treating disorders of arousal stability such as narcolepsy and perhaps insomnia. Future research should test causal necessity of the identified VLPO GABA subset in vivo, dissect upstream/downstream partners, explore pharmacological modulation of Hcrtr2-expressing VLPO neurons, and determine how other arousal/sleep signals converge on this intra-VLPO circuit.

• The causal necessity of the VLPO GABA-mediated feedforward pathway for orexin-evoked arousal in vivo was not directly tested via selective inhibition or ablation; redundancy in arousal circuits may compensate. - Optogenetic stimulation of orexin terminals and in vitro slice experiments may not fully capture endogenous firing patterns, neuromodulator dynamics, or long-range network context. - Evidence for orexin peptide action at the VLPO synapse was indirect; glutamatergic oEPSCs were observed, but peptide-mediated postsynaptic currents were not detected under the recording conditions. - Findings are in mice; translational generalizability to humans remains to be established. - Single-cell transcriptomic inferences rely on reanalysis of publicly available datasets and RNAscope validation; low-abundance transcripts or receptor protein levels were not quantified across all subpopulations.