Orexin neurons receive glycinergic innervations

Abstract

Glycine, a nonessential amino-acid that acts as an inhibitory neurotransmitter in the central nervous system, is currently used as a dietary supplement to improve the quality of sleep, but its mechanism of action is poorly understood. We confirmed the effects of glycine on sleep/wakefulness behavior in mice when administered peripherally. Glycine administration increased non-rapid eye movement (NREM) sleep time and decreased the amount and mean episode duration of wakefulness when administered in the dark period. Since peripheral administration of glycine induced fragmentation of sleep/wakefulness states, which is a characteristic of orexin deficiency, we examined the effects of glycine on orexin neurons. The number of Fos-positive orexin neurons markedly decreased after intraperitoneal administration of glycine to mice. To examine whether glycine acts directly on orexin neurons, we examined the effects of glycine on orexin neurons by patch-clamp electrophysiology. Glycine directly induced hyperpolarization and cessation of firing of orexin neurons. These responses were inhibited by a specific glycine receptor antagonist, strychnine. Triple-labeling immunofluorescent analysis showed close apposition of glycine transporter 2 (GlyT2)-immunoreactive glycinergic fibers onto orexin-immunoreactive neurons. Immunoelectron microscopic analysis revealed that GlyT2-immunoreactive terminals made symmetrical synaptic contacts with somata and dendrites of orexin neurons. Double-labeling immunoelectron microscopy demonstrated that glycine receptor alpha subunits were localized in the postsynaptic membrane of symmetrical inhibitory synapses on orexin neurons. Considering the importance of glycinergic regulation during REM sleep, our observations suggest that glycine injection might affect the activity of orexin neurons, and that glycinergic inhibition of orexin neurons might play a role in physiological sleep regulation.

Figure 1. Glycine-induced sleep phenotype during dark period.

A Protocols for glycine administration in light (left panel) or dark period (right panel). Glycine or saline was injected intraperitoneally 10 min before the start of recording (ZT0 or ZT12). EEG/EMG recordings were performed for the next 5 hours (until ZT5 or ZT17). B, C Total time (minutes, mean ± SEM) spent in each state in saline- (n = 4, white bar) and glycine-administered mice (n = 4, gray bar), itemized separately for light (B) and dark periods (C). D, E Episode duration (seconds, mean ± SEM) spent in each state in saline- and glycine-administered mice, in light (D) or dark period (E). F, G Stage count (count, mean ± SEM) is number of each episode during each period (light; F, dark; G). The glycine-administered group showed a significantly shorter total time and duration of episodes of wakefulness, suggesting fragmentation of sleep/wakefulness states during the dark period. *p<0.015. Graphs summarize the data recorded during the 5 h light/dark period.

Figure 2. Effects of glycine administration to mice on activity of orexin neurons.

A. Protocols for glycine administration in light (left panel) or dark period (right panel). Glycine or saline was injected intraperitoneally at ZT0 or ZT12, and then mice were sacrificed at ZT3 or ZT15. B–E, Representative immunohistochemical micrographs of double-staining with anti-c-Fos (black) and anti-orexin A (brown) antibody at ZT3 (B, control (saline); D, glycine (2 g/kg)) and at ZT15 (C, control; E, glycine). White arrowheads show orexin-immunoreactive cells. Black arrowheads show orexin neurons with Fos-immunoreactivity in their nuclei. Scale bar, 20 µm. F. Percentage of c-Fos-expressing orexin neurons at ZT 3 (control; n = 3, glycine; n = 4) and ZT 15 (control; n = 4, glycine; n = 6). Values are mean ± SEM. *p<0.02, **p<0.0001.

Figure 3. Glycine directly inhibits orexin neurons.

A. Under whole-cell current-clamp mode, glycine (1 mM) was applied to orexin neurons in normal extracellular solution, in the absence of TTX (upper panel), in the presence of TTX (1 µM) (middle panel) and in the presence of both TTX (1 µM) and strychnine (1 µM) (lower panel). Glycine was applied during the period indicated by the bars. B. Concentration dependency of glycine-induced outward currents in orexin neurons clamped at −60 mV. EC₅₀ and E_max were 10^{−3.00±0.17} M and 205±5 pA, respectively (n = 2–3). C. Strychnine (1 µM) significantly inhibited the effect of glycine in the presence of TTX. D. Records of membrane potential in response to a series of current steps (from −150 to +120 pA in 30 pA increments) from resting potential (−60 mV) in the absence (left) or presence (right) of glycine (10 mM). E. Current-voltage relationship derived from the data in F. The potential at the end of current injection was plotted; control (open circles) and 10 mM glycine (filled circles). Estimated reversal potential was −80.5 mV (n = 3). Values are mean ± SEM.

Figure 4. Specificity of GlyRα and GlyT2 antibodies.

A, E. Immunoblotting using GlyRα and GlyT2 (E) antibodies. Each lane was loaded with HEK293 cell lysates transfected with expression plasmid only (mock), transfected with expression plasmid encoding GlyRα1, GlyRα3, GlyRα4, GlyRβ, or GlyT2, and adult mouse brain homogenates. Because of low viability, HEK293 cell lysate of GlyRα2-transfected cells was omitted. Identical patterns of immunoblot labeling were repeatedly confirmed in three independent experiments. The position of protein size markers (kDa) is shown on the left. B, C, F, G. Immunofluorescence with use of GlyRα (B, C) or GlyT2 (F, G) in parasagittal brain sections. Note that preincubation of antibodies with antigen protein (C, G) almost completely abolished immunostaining. D. Double-labeling postembedding immunogold for GlyRα (10 nm) and VIAAT (15 nm) in the facial nucleus. Note that GlyRα localizes to a symmetrical synapse (black arrowheads) made with a VIAAT-positive inhibitory terminal (In), but not to an asymmetrical synapse (white arrowheads) made with a VIAAT-negative excitatory terminal (Ex). Arrows indicate immunogold particles for GlyRα. H. Preembedding silver-enhanced immunoelectron microscopy of facial nucleus. Metal particles for GlyT2 are preferentially distributed on the cell membrane of presynaptic inhibitory terminals (In), which contain flat synaptic vesicles and make symmetrical synapses with dendritic shafts (Dn). OB, olfactory bulb; Cx, cerebral cortex; St, striatum; Th, thalamus; Ht, hypothalamus; Mb, midbrain; Po, pons; Cb, cerebellum; MO, medulla oblongata; Ex, excitatory terminal; In, inhibitory terminal; Dn, dendrite. Scale bars, B, C, F, G; 1 mm; D, H, 400 nm.

Figure 5. Immunofluorescence showing glycinergic innervations to orexin neurons.

A. Double immunofluorescence showing distribution of GlyT2-positive glycinergic varicose fibers (red) and GFP-positive orexin neurons (green) in the LHA. B, C. Triple immunofluorescence showing that orexin neurons (GFP, green) are associated with GlyT2-positive (red)/VIAAT-positive (blue) varicosities (arrows). D. Double immunofluorescence showing the distribution of GlyRα- (red) and GFP-positive orexin neurons (green) in the LHA. E, F. Orexin neurons (green) display numerous VIAAT-positive inhibitory terminals (blue), some of which are associated with GlyRα immunoreactivity (red) on the surface of orexin neurons. Scale bars, 10 µm.

Figure 6. Immunoelectron microscopy showing glycinergic synapse formation on orexin neurons.

Consecutive images from double-labeling preembedding immunoelectron microscopy for GFP (silver-intensified immunogold) and GlyT2 (immunoperoxidase). Note that a GlyT2-positive terminal (filled with diffuse DAB precipitates) makes a symmetrical synaptic junction (arrowheads) onto the soma (A) and dendritic shaft (B) of a GFP-positive orexin neuron. Scale bar, 500 nm.

Figure 7. Postembedding immunogold showing glycine receptor localization at symmetrical synapses.

A. Double-labeling postembedding immunogold for GlyRα (small particles) and orexin A (large particles) shows that an orexin neuron, whose Golgi apparatus is heavily labeled with immunogold for orexin-A (A₃), expresses GlyR at a symmetrical synapse (black arrowheads, A₂). B,C. Triple-labeling immunoelectron microscopy for GFP (dark precipiates by preembedding immunoperoxidase), GlyRα (10 nm, postembedding immunogold) and VIAAT (20 nm, postembedding immunogold). GlyRα is localized to contact sites made between VIAAT-positive inhibitory terminals (In) and the soma (B₂) or dendritic shaft (Dn, C) of GFP-positive orexin neurons. Scale bars, 1 µm in B₁, 400 nm in A,B₂,C.