GABAergic neurons in the preoptic area send direct inhibitory projections to orexin neurons

Abstract

Populations of neurons in the hypothalamic preoptic area (POA) fire rapidly during sleep, exhibiting sleep/waking state-dependent firing patterns that are the reciprocal of those observed in the arousal system. The majority of these preoptic "sleep-active" neurons contain the inhibitory neurotransmitter GABA. On the other hand, a population of neurons in the lateral hypothalamic area (LHA) contains orexins, which play an important role in the maintenance of wakefulness, and exhibit an excitatory influence on arousal-related neurons. It is important to know the anatomical and functional interactions between the POA sleep-active neurons and orexin neurons, both of which play important, but opposite roles in regulation of sleep/wakefulness states. In this study, we confirmed that specific pharmacogenetic stimulation of GABAergic neurons in the POA leads to an increase in the amount of non-rapid eye movement (NREM) sleep. We next examined direct connectivity between POA GABAergic neurons and orexin neurons using channelrhodopsin 2 (ChR2) as an anterograde tracer as well as an optogenetic tool. We expressed ChR2-eYFP selectively in GABAergic neurons in the POA by AAV-mediated gene transfer, and examined the projection sites of ChR2-eYFP-expressing axons, and the effect of optogenetic stimulation of ChR2-eYFP on the activity of orexin neurons. We found that these neurons send widespread projections to wakefulness-related areas in the hypothalamus and brain stem, including the LHA where these fibers make close appositions to orexin neurons. Optogenetic stimulation of these fibers resulted in rapid inhibition of orexin neurons. These observations suggest direct connectivity between POA GABAergic neurons and orexin neurons.

Keywords: GABA; hypothalamus; orexin; preoptic area; sleep; wakefulness.

Figure 1

Selective pharmacogenetic stimulation of POA neurons. (A) Gad67-Cre mice were crossed with Rosa26-tdTomato reporter mice (see method) to confirm GABAergic neuron-specific expression of Cre recombinase. A representative image of the POA of the Gad67-Cre; Rosa26-tdTomato mice is shown (Bregma-0.1 mm). Left panels, upper, tdTomato fluorescence. Lower, same section stained with Gad65/67 anti-body. Middle panel, merged image of rectangular region in the left panel. Right panel: High power view of rectangular region in central panel. Arrowheads show the colocalization of tdTomato fluorescence and Gad65/67 immunoreactivity. (B) Left, Virus injection sites are shown by dots. Injection site for right panel image is shown by a red dot. Right, Representative image of HA-like immunoreactivity observed in POA region of Gad67-Cre mice injected with AAV-DIO-HAhM3Dq to express hM3Dq fused with HA tag. (C) Activation of POA neurons in Gad67-Cre mice expressing hM3Dq by CNO. Left panels, upper, representative images of double-immunostaining with anti-Fos and anti-Gad 65/67 in the POA region after administration of saline (left) or CNO (right) at 21:00. The brain was fixed at 23:00. Left panels, lower, high power view of the rectangular regions shown in upper panels. Arrowheads show the colocalization of Fos (nuclei) and Gad65/67 (cytoplasm). Right panel, Number of fos-immunoreactive GABAergic neurons in POA after treatment with saline or CNO (N = 4 and 4, respectively).

Figure 2

Specific pharmacogenetic stimulation of GABAergic neurons in the POA increased NREM sleep amount. (A) Total time of wakefulness (WAKE), NREM sleep and REM sleep for 3 h after CNO (or saline) administration at 21:00 (upper panels, Saline n = 14, CNO n = 13) and at 13:00 (lower panels, Saline n = 7, CNO n = 7). (B) Episode duration of WAKE, NREM sleep, and REM sleep for 3 h after CNO (or saline) administration at 21:00 (upper panels, Saline n = 14, CNO n = 13) and at 13:00 (lower panels, Saline n = 7, CNO n = 7). (C) EEG power density of WAKE, NREM sleep and REM sleep in 1–2 h time window after the administration of CNO or saline at 21:00 (Saline n = 14, CNO n = 13). EEG power density is shown as the mean percentage of total EEG power ± s.e.m. (D) Representative images of fos expression in orexin neurons, as shown by double staining of the LHA of Gad67-Cre mice 2 h after injection of saline (n = 5) or CNO (n = 8) at 21:00. Scale bars, 50μm. Right panel, ratio of Fos-positive orexin neurons after injections of saline or CNO. Arrowheads show colocalization of Fos (nuclei) and orexin (cytoplasm).

Figure 3

Mapping of projection sites of the POA GABAergic neurons. (A) Expression of ChR2-eYFP mRNA (blue) and Gad67 mRNA (red) in the POA of Gad67-Cre mice after bilateral injection of AAV-DIO-hChR2(H134R)-eYFP into the POA. Almost all (>95%, n = 4) ChR2-eYFP mRNA-positive neurons also expressed Gad67 mRNA. Arrowheads show co-localization of ChR2-eYFP mRNA and Gad67 mRNA. (B) After injection of AAV-DIO-hChR2(H134R)-eYFP into the POA of Gad67-Cre mice, the brain was subjected to histological analysis. Representative images show localization of ChR2-positive fibers from rostral to caudal in Gad67-Cre mouse brain coronal sections revealed by immunohistochemical staining using anti-GFP antiserum. (C) Images of single confocal planes showing ChR2-positive axonal fibers observed in the LHA, VTA, LC, LDT, PPT, DR, and TMN. Slices were stained with anti-GFP (green) and designated antibodies (red), including anti-orexin, anti-TH, anti-ChAT, anti-TPH and anti-HDC antibodies. Scale bars: 40 μm.

Figure 4

Optogenetic stimulation of POA GABAergic axons rapidly decreases orexin neuron firing. (A) Strategy of this study. We simultaneously injected AAV-horexin-tdTomato into the LHA and AAV-DIO-hChR2(H134R)-eYFP into the POA. (B) Identification of orexin neurons in the LHA by expression of tdTomato in these neurons after injection of AAV-horexin-tdTomato. Left: Coronal section at bregma −2.1 mm showing distribution of orexin neurons stained by anti-orexin antiserum (green). Center: Cells with red fluorescence of tdTomato. Right: Merged image (C) Typical electrical current-clamp recording of orexin neuron. Left: Example of changes in firing of orexin neurons induced by optical stimulation in the absence (top trace) and presence (bottom trace) of bicuculline (20 μM). We observed similar responses in three out of eight cells. Right: Group data for response and recovery at 10 Hz stimulation, expressed at % change in firing before (B), during (S) and after (R) light stimulation. (D) Optogenetically induced IPSCs in orexin neurons with and without bicuculline (20 μM). Holding potential was −60 mV. Black bar indicates light stimulus (5 ms, 1 Hz). Individual responses without bicuculline are shown in gray, average without bicuculline is in black, and average with bicuculline is in red. (E) Times after light-on to IPSPs onset are plotted for individual stimulations.