Adenosine in the Tuberomammillary Nucleus Inhibits the Histaminergic System Via A1 Receptors and Promotes Non-Rapid Eye Movement Sleep

Home , Histaminergic, Tuberomammillary nucleus

Adenosine in the tuberomammillary nucleus inhibits the histaminergic system via A1 receptors and promotes non-rapid eye movement sleep

Yo Oishia,b, Zhi-Li Huanga,c,1, Bertil B. Fredholmd, Yoshihiro Uradea,b, and Osamu Hayaishia,1

aDepartment of Molecular Behavioral Biology, Osaka Bioscience Institute, Suita, Osaka 565-0874, Japan; bDepartment of Aging Science, Graduate School of Medicine, Osaka University, Suita, Osaka 565-0871, Japan; cState Key Laboratory of Medical Neurobiology and Department of Pharmacology, Shanghai Medical College of Fudan University, Shanghai 200032, China; and dDepartment of Physiology and Pharmacology, Karolinska Institute, S-171 77 Stockholm, Sweden

Contributed by Osamu Hayaishi, October 30, 2008 (sent for review October 6, 2008)

Adenosine has been proposed to promote sleep through A1 re- promotes sleep through adenosine acting at adenosine A2A ceptors (A1R’s) and/or A2A receptors in the brain. We previously receptors (A2AR’s) (11), followed by activation of sleep-active reported that A2A receptors mediate the sleep-promoting effect of neurons in the ventrolateral preoptic (VLPO) area (12, 13). The prostaglandin D2, an endogenous sleep-inducing substance, and somnogenic effect of PGD2 is mimicked by an A2AR agonist, but that activation of these receptors induces sleep and blockade of not by an adenosine A1 receptor (A1R) one, and is blocked by them by caffeine results in wakefulness. On the other hand, A1R an A2AR antagonist (11). These findings indicate that the has been suggested to increase sleep by inhibition of the cholin- PGD2–A2AR system is involved in both circadian and homeo- ergic region of the basal forebrain. However, the role and target static sleep regulation (14). Among the four subtypes of adenosine receptors, A1,A2A, sites of A1R in sleep–wake regulation remained controversial. In this study, immunohistochemistry revealed that A R was ex- A2B, and A3 (15), A1R and/or A2AR subtypes have been reported 1 to mediate the sleep-promoting effect of adenosine. Although pressed in histaminergic neurons of the rat tuberomammillary A R’s involvement in the PGD –VLPO system is clearly es- nucleus (TMN). In vivo microdialysis showed that the histamine 2A 2 tablished, adenosine via A R has been proposed to induce sleep release in the frontal cortex was decreased by microinjection into 1 6 by inhibiting the cholinergic region of the BF (16). For example, the TMN of N -cyclopentyladenosine (CPA), an A1R agonist, aden- the unilateral infusion of the BF with an A1R-selective antag- osine or coformycin, an inhibitor of adenosine deaminase, which onist increased waking and decreased sleep (17). Single unit catabolizes adenosine to inosine. Bilateral injection of CPA into the recording of BF neurons in conjunction with in vivo microdialysis rat TMN signiﬁcantly increased the amount and the delta power of an A1R-selective agonist decreased, and an A1R antagonist, density of non-rapid eye movement (non-REM; NREM) sleep but increased the discharge activity of the neurons in the BF (18). did not affect REM sleep. CPA-promoted sleep was observed in WT Moreover, perfusion of A1R antisense oligonucleotides into the mice but not in KO mice for A1R or histamine H1 receptor, indicating BF reduced NREM sleep and EEG delta power (19). However, that the NREM sleep promoted by A1R-speciﬁc agonist depended infusion of an A1R agonist into the lateral ventricle of mice did on the histaminergic system. Furthermore, the bilateral injection of not alter the amounts of NREM and REM sleep (20). Caffeine, adenosine or coformycin into the rat TMN increased NREM sleep, an antagonist for both A1R and A2AR, increased wakefulness in which was completely abolished by coadministration of 1,3- A1R KO mice and in WT mice, but not in A2AR KO mice (21). dimethyl-8-cyclopenthylxanthine, a selective A1R antagonist. Therefore, the role of A1R in sleep–wake regulation has re- These results indicate that endogenous adenosine in the TMN mained uncertain. In the brain parenchyma, adenosine deaminase (ADA), an suppresses the histaminergic system via A1R to promote NREM sleep. enzyme which catabolizes adenosine to inosine, is dominantly localized in the tuberomammillary nucleus (TMN) of the posterior hypothalamus (22) and is colocalized with histidine de- Adenosine deaminase ͉ histamine ͉ knockout mouse ͉ rat carboxylase (HDC) (23), the key enzyme for histamine synthesis. Histaminergic neurons project from the TMN to most of the n 1954, Feldberg and Sherwood (1) showed that intraventric- central nervous system and have been shown to promote wake- Iular injection of micromole quantities of adenosine into cats fulness through histamine H1 receptors (H1R’s) (3, 24). How- caused a state resembling natural sleep of 30-min duration. ever, the functional significance of adenosine and high expres- Subsequent pharmacological studies from several laboratories sion of ADA in the TMN has not been elucidated so far. demonstrated that adenosine and its receptor agonists pro- In the present study, we found that A1R was coexpressed with moted, but antagonists such as caffeine, inhibited both non-rapid ADA in rat TMN and that activation of A1R or inhibition of eye movement (non-REM; NREM) and REM sleep (for review, ADA in the TMN inhibited histaminergic systems to promote see refs. 2, 3). However, the exact molecular mechanisms NREM sleep without affecting REM sleep, clearly indicating underlying sleep–wake regulation by adenosine still remained that adenosine in the TMN promotes NREM sleep via A1R’s. unclear. Since 1983, we have reported that prostaglandin (PG) Results D2 is an endogenous sleep-inducing substance in rodents (4) and Localization of A1R in Histaminergic Neurons of the Rat TMN. Immu- monkeys (5). PGD2-induced sleep was indistinguishable from natural physiological sleep as judged by several electrophysio- nohistochemical staining with polyclonal and monoclonal (25) logical and behavioral criteria (5). Subsequent studies have

shown that PGD2 is involved in both circadian (6, 7) and Author contributions: Y.O., Z.-L.H., Y.U., and O.H. designed research; Y.O. and Z.-L.H. homeostatic regulation of sleep (8). Further studies on the performed research; B.B.F. contributed new reagents/analytic tools; Y.O., Z.-L.H., Y.U., and molecular mechanisms of sleep–wake regulation by PGD2 dem- O.H. analyzed data; and Y.O., Z.-L.H., Y.U., and O.H. wrote the paper. onstrated that PGD2 is produced by the action of lipocalin-type The authors declare no conﬂict of interest. PGD synthase dominantly expressed in the leptomeninges (9), 1To whom correspondence may be addressed. E-mail: [email protected] or binds with PGD receptors (DPRs) exclusively localized in the [email protected]. arachnoid membrane of the basal forebrain (BF) (10), and © 2008 by The National Academy of Sciences of the USA

19992–19997 ͉ PNAS ͉ December 16, 2008 ͉ vol. 105 ͉ no. 50 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810926105 Downloaded by guest on September 30, 2021 demonstrated that A1R was colocalized with HDC or ADA in the TMN. These results indicate that A1R was expressed in HDC-positive neurons and in ADA-positive ones in the TMN.

Inhibitory Action of A1R in Histaminergic Systems of the Rat TMN. To monitor the regulation of histaminergic systems by A1R or ADA, we determined the histamine release from the frontal cortex (FrCx) after a bolus injection of a selective A1R agonist, N6-cyclopentyladenosine (CPA), its natural agonist, adenosine, or an ADA inhibitor, coformycin, into the TMN in urethane- anesthetized rats (Fig. 1J). The CPA administration into the TMN (0.17–1.5 nmol/side) induced a remarkable dose- dependent decrease in the histamine release from the FrCx for 2 h after the injection (Fig. 1K). The adenosine administration (2.7 or 4.5 nmol/side) also induced a significant decrease in the histamine release for1haftertheinjection in a dose-dependent manner (Fig. 1L), although the inhibition of the histamine release was weaker and shorter than in the case of CPA. The injection of coformycin (1.3 or 4 nmol/side) also decreased the histamine release for1hfrom1haftertheinjection (Fig. 1M). These results indicate that the stimulation of A1R’s in the TMN with CPA or adenosine and the inhibition of ADA with coformycin suppressed the activity of histaminergic systems in the TMN to decrease the histamine release in the FrCx.

NREM Sleep Promotion by Activation of A1R in the Rat TMN with CPA. We then examined the sleep–wake profile after a bilateral injection of CPA into the TMN of freely moving rats at 22:00, when such animals spend most of their time in wakefulness. Typical examples of EEG, electromyogram (EMG), and hypnograms from rats given vehicle or CPA at a dose of 1.5 nmol/side are shown in Fig. 2A. The CPA (1.5 nmol/side) injection remarkably increased NREM sleep for3hfrom22:30 to 01:30, as compared with the case of the vehicle injection. As shown in Fig. 2B, CPA at 1.5 nmol/side significantly increased the hourly NREM sleep time by 4.6-, 2.1-, 2.5-, and 2.1-fold during the first, second, third, and fourth hour, respectively, after the injection as compared with the vehicle injection. Enhancement of NREM sleep by the CPA injection concomi- Fig. 1. Immunohistochemistry of A1R, A2AR, and HDC in the rat TMN (A–H) and in vivo microdialysis to measure histamine release in the rat FrCx (I–M). (A) tantly decreased wakefulness, but did not affect REM sleep. No Low- and high- (inset to A) magnification views of rat TMN immunostained additional disruption of sleep architecture was observed during with polyclonal A1R antibody. (B) Low- and high- (inset to B) magnification the subsequent period. Similar time course profiles were ob- views of rat TMN after incubation with anti-A2AR antibody. (C–E) Double served with the high (4.5 nmol/side) and middle (0.5 nmol/side) immunofluorescence staining for A1R(C, red) and HDC (D, green) in rat TMN doses of CPA. When CPA at a dose of 0.17 nmol/side was and their merged view (E). (F–H) Double immunofluorescence staining for A1R administered, no significant change was found in the sleep–wake (F, red) and ADA (G, green) in rat TMN and their merged view (H). (I) Coronal profile. section of a rat injected with coformycin into the TMN and immunostained The dose-dependency of NREM sleep promotion by the CPA with anti-HDC antibody to identify the TMN (stained brown and indicated by the arrowhead). The arrow indicates the trace of the injection cannula. 3V, administration is summarized in Fig. 2C. When the vehicle third ventricle. (J) Schematic representation of the implantation sites for the solution or CPA at 0.17 nmol/side was administered into the guide cannulae 3 mm above the TMN and injection sites (Left) and microdi- TMN, no difference was found in the sleep–wake cycle as alysis probe in the FrCx (Right) of rats. Figures of coronal sections are from the compared with the vehicle injection. CPA given at 0.5 and 1.5 stereotaxic atlas of Paxinos and Watson (37). (K–M) Histamine release from the nmol/side significantly increased the total amount of NREM FrCx after administration with CPA (K), adenosine (Ado, L), or coformycin (CF, sleep during4haftertheinjection by 2.1- and 3.1-fold in M) into the rat TMN. The mean value of histamine output during 1 h before comparison with the vehicle injection, whereas REM sleep was injection was defined as the basal output, and the subsequent fractions were not affected. When CPA at 4.5 nmol/side was administered, Ϯ ϭ expressed as percentages of this value. Values are means SEM (n 5–7). *, NREM sleep during4haftertheinjection was increased by P Ͻ 0.05; **, P Ͻ 0.01, significantly different from the vehicle injection, assessed by two-way ANOVA followed by Fisher’s probable least-squares 3.0-fold without affecting REM sleep, similar to the case of the difference test. 1.5 nmol/side, suggesting that these doses of CPA almost completely activated A1R’s in the TMN. When we determined the NREM sleep bout distribution for 4 h after the CPA injection (Fig. 2D), the CPA treatments anti-A1R antibodies revealed that A1R was predominantly lo- (0.5–4.5 nmol/side) decreased the number of the shorter NREM calized in the TMN in the posterior hypothalamus of rats (Fig. sleep bouts (60–120 sec) and increased the number of the 1A). On the other hand, no positive staining was observed in the prolonged ones (Ͼ240 sec) in a dose-dependent manner as TMN with anti-A2AR antibody (Fig. 1B) or without the addition compared with the vehicle injection. On the other hand, the CPA of primary antibodies (data not shown). Double immunofluo- administration did not affect the numbers of stage transition

rescence staining with antibodies against A1R and HDC (Fig. 1 from NREM sleep to wakefulness, wakefulness to NREM sleep, NEUROSCIENCE C–E), a marker for histaminergic neurons, or ADA (Fig. 1 F–H) NREM to REM sleep, and REM sleep to wakefulness. These

Oishi et al. PNAS ͉ December 16, 2008 ͉ vol. 105 ͉ no. 50 ͉ 19993 Downloaded by guest on September 30, 2021 Fig. 3. NREM sleep increase after administration of CPA (0.3 nmol/side) into the TMN of WT, but not A1RorH1R KO, mice. (A and C) Time courses of NREM and REM sleep after CPA treatment of WT and KO mice for A1R(A)orH1R(C). (B and D) NREM and REM sleep during a 4-h period after the injection of CPA into WT and KO mice for A1R(B)orH1R(D). Values are means Ϯ SEM (n ϭ 5–6). *, P Ͻ 0.05; **, P Ͻ 0.01, signiﬁcantly different from the baseline.

curve in the delta range of 0.5–4.0 Hz as compared with the vehicle. The low dose of CPA (0.17 or 0.5 nmol/side) showed a tendency to increase the power density curve of the delta range, although the difference was not statistically significant. These results clearly indicate that activation of A1R in the dose- dependent TMN increased the amount and the delta power density of NREM sleep.

NREM Sleep Promotion by Activation of A1R in the TMN of WT but not Fig. 2. Increase in NREM sleep after CPA injection into the rat TMN. (A) KO Mice for A1RorH1R. To investigate the specificity of CPA Typical examples of EEG, EMG, and hypnograms after administration of against A1R in vivo, we bilaterally injected CPA into the TMN vehicle (Upper panel) or CPA at a dose of 1.5 (Lower panel) nmol/side. (B) Time of WT and A1R KO mice of the same littermates (Fig. 3A). CPA courses of NREM and REM sleep in rats treated with CPA at 1.5 nmol/side. (C) at 0.3 nmol/side increased NREM sleep in the WT mice by 1.8-, NREM and REM sleep over a 4-h period after injection of CPA. (D) Number of 3.4-, 1.7-, and 2.4-fold during the first, second, third, and fourth NREM sleep bouts for 4 h after injection of CPA (Upper panel) and the numbers hour, respectively, after the injection as compared with that of of stage transition from NREM (N) to wakefulness (W), wakefulness to NREM, the baseline day. However, A R KO mice did not display any NREM to REM (R), and REM to wakefulness in 4 h after the injection of CPA 1 significant changes in NREM sleep after the CPA administra- (Lower panel). *, P Ͻ 0.05; **, P Ͻ 0.01, signiﬁcantly different from the vehicle injection. (E) EEG power density curve during NREM sleep for 4 h after tion. The total amount of NREM sleep for4hafteradminis- injection of CPA. The power of each 0.25-Hz bin was averaged across the sleep tration of CPA at 0.3 nmol/side increased by 2.1-fold in the WT stages and normalized as a group by calculating the percentage of each bin mice from the baseline, whereas no difference was observed in from the total power (0–24.75 Hz). The horizontal bars indicate where there the A1R KO mice (Fig. 3B), indicating that A1R is crucial for the is a statistical difference (P Ͻ 0.05) between vehicle and CPA at 1.5 (red) or 4.5 increased NREM sleep by CPA injected into the TMN. When Ϯ ϭ (purple) nmol/side. Values are means SEM (n 5–7). CPA (0.3 nmol/side) was administered to the TMN of WT and histamine H1R KO mice, the NREM sleep was significantly increased in WT mice by 2.9-, 4.9-, and 2.4-fold during the first, results suggest that CPA did not increase the total number of second, and fourth hour, respectively, after the injection (Fig. NREM sleep bouts but extended the duration of NREM sleep. 3C), giving a 2.8-fold increase during4haftertheinjection (Fig. Also, CPA administration into the TMN increased the delta 3D) from the baseline, but not in the H1R KO mice at all. power density of NREM sleep (Fig. 2E). CPA at 1.5 or 4.5 However, REM sleep was not affected by the treatment of CPA nmol/side significantly increased the peak of the power density in WT, A1R KO, or H1R KO mice. These results clearly indicate

19994 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810926105 Oishi et al. Downloaded by guest on September 30, 2021 adenosine did not alter the total amount of NREM sleep at a dose of 1.5 nmol/side but increased the amount of NREM sleep by 1.5- and 1.8-fold at doses of 2.7 and 4.5 nmol/side, respectively, in a dose-dependent manner. On the other hand, coformycin increased the total amount of NREM sleep weakly at a dose of 1.3 nmol/side and 1.9- and 1.7-fold at 4 and 12 nmol/side, respectively, suggesting that the latter two doses of coformycin almost completely inhibited ADA in the TMN. However, the treatment with adenosine (1.5–4.5 nmol/side) or coformycin (1.3–12 nmol/side) did not change the amount of REM sleep. These results suggest that NREM sleep was promoted by the increased adenosine level in the TMN after the adenosine administration or the inhibition of ADA with coformycin.

Abolishment of Increased NREM Sleep by Adenosine or ADA Inhibitor with an A1R Antagonist. The increase in NREM sleep by administration of adenosine (4.5 nmol/side) or coformycin (4 nmol/ side) into the rat TMN was completely suppressed by the coadministration of it with 1,3-dimethyl-8-cyclopenthylxanthine (CPT; 0.2 nmol/side), a selective A1R antagonist (Fig. 4C). On the other hand, the injection of CPT per se at the same dose did not produce significant changes in NREM sleep. These results suggest that the NREM sleep, increased by the elevated adenosine level in the TMN, depended on A1R’s. Moreover, CPT at 0.4 nmol/side significantly decreased NREM sleep by 26% as compared with the vehicle injection, suggesting that A1Rinthe TMN is also involved in physiological sleep. Because of the poor solubility of CPT, we could not examine its effect on the sleep profile at concentrations Ͼ0.4 nmol/side. We did not find significant changes in the REM sleep by the CPT treatment. These results, all together, indicate that the increased adenosine Fig. 4. NREM and REM sleep after administration of adenosine or coformy- level by adenosine injection or by inhibition of ADA in the TMN cin into the rat TMN in the absence or presence of an A1R antagonist. (A–B) promoted the NREM sleep via A1R’s. Time courses of NREM and REM sleep in rats administered adenosine (Ado) at 4.5 nmol/side (A) or coformycin (CF) at 4 nmol/side (B). (C) NREM and REM sleep Discussion during a 4-h period after the injection of adenosine or coformycin with and In this study, we demonstrated that administration of exogenous without the A R antagonist CPT. Values are means Ϯ SEM (n ϭ 5–8). , P Ͻ 0.05; 1 * adenosine or inhibition of ADA in the TMN suppressed the **, P Ͻ 0.01, signiﬁcantly different from the vehicle injection. histaminergic arousal system and increased the amount of NREM sleep. This effect was mimicked by activation of A1R that the increased NREM sleep by activation of A1R’s in the with its agonist CPA and abolished with the antagonist CPT. TMN depended on the H1R-related arousal system. These findings clearly indicate that A1R mediates the inhibition of the TMN by adenosine to promote NREM sleep. Murillo- NREM Sleep Promotion by Administration of Adenosine or ADA Rodriguez et al. (26) reported that an A1R agonist increased Inhibitor, Coformycin, to the Rat TMN. We then examined the sleep after perfusion into the BF, and Strecker et al. (17) found effects of the bilateral injection of adenosine or coformycin into that the unilateral infusion of an A1R-selective antagonist into the TMN on the sleep–wake profile of rats. The time courses of the BF decreased sleep. We also confirmed that microinjected NREM and REM sleep after administration of adenosine and CPA at 1.5 nmol/side into the BF increased sleep to a lower coformycin are shown in Fig. 4A and B, respectively. Adenosine extent than that given into the TMN (data not shown). In at 4.5 nmol/side significantly increased the hourly NREM sleep contrast, Methippara et al. (27) reported that administration of an A R agonist into the lateral preoptic area induced wakeful- time by 2.2-, 2.6-, and 1.8-fold during the first, second, and third 1 ness. These findings indicated that the somnogenic or arousal hour, respectively, as compared with the vehicle injection. Co- effect via A R is a region-dependent one and are in good formycin at 4 nmol/side also induced a significant increase in 1 agreement with our previous finding that sleep amounts were not NREM sleep, by 2.0- and 1.8-fold during the third and fourth ϳ changed at all when an A1R agonist was infused into the lateral hour, respectively, with a lag time of 2 h after the injection, ventricle of mice (20). similar to the pattern of decrease in the histamine release. Our study showed that adenosine is an endogenous suppressor Enhancement of NREM sleep by the injection of adenosine or of the histaminergic arousal system. Sherin et al. (28) previously coformycin concomitantly decreased wakefulness, but neither reported the GABAergic and galaninergic innervations of his- affected the amount of REM sleep nor disrupted additional taminergic TMN neurons from the sleep-active VLPO neurons. sleep architecture during the subsequent period. Similar time We previously identified GABA release in the TMN to inhibit course profiles were observed with the middle dose (2.7 nmol/ the histaminergic system (29). Synaptic co-release of ATP and side) of adenosine or the low (1.3 nmol/side) and high (12 GABA has already been demonstrated (30, 31). Therefore, ATP nmol/side) doses of coformycin. No significant change was found co-released with GABA from the VLPO neurons may be a in the sleep–wake profile at a low dose of adenosine (1.5 source of adenosine acting as a suppressor in the TMN. On the nmol/side). other hand, we previously reported that the extracellular level of The dose-dependency of NREM sleep promotion during 4 h adenosine was increased in the subarachnoid space under the

after the administration of adenosine or coformycin is summa- rostral BF after stimulation of DPRs on the arachnoid mem- NEUROSCIENCE rized in Fig. 4C. As compared with the vehicle injection, brane with PGD2 (10). The highest DPR immunoreactivity was

Oishi et al. PNAS ͉ December 16, 2008 ͉ vol. 105 ͉ no. 50 ͉ 19995 Downloaded by guest on September 30, 2021 found in the arachnoid membrane at the basal aspect of the TMN Materials and Methods (10). Thus, the adenosine increased by stimulation of DPR in the Animals and Chemicals. Male Sprague–Dawley rats, weighing 280–320 g (8–10 arachnoid membrane near the TMN may also suppress the weeks old), were purchased from Japan SLC Inc. A1R KO mice (35) and H1RKO histaminergic system. mice (36) by T. Watanabe, RIKEN Institute, were maintained at Oriental ADA is dominantly localized in the TMN in the brain paren- Bioservice, Japan. Both male KO mice and their respective WT littermates of chyma (22) and has been used as a convenient immunohisto- inbred C57BL/6 strain, weighing 20–26 g (10–12 weeks old), were used in these Ϯ chemical marker of histaminergic neurons. However, the phys- experiments. They were housed at an ambient temperature of 22 0.5 °C Ϯ iological function of ADA has not yet been elucidated. Our study with a relative humidity of 60 2% on an automatically controlled 12:12 h light/dark cycle (light on at 08:00, illumination intensity Ϸ100 lux). They had clearly showed that ADA in the TMN is involved in the ad libitum access to food and water. The experimental protocols were ap- sleep–wake regulation, because treatment of the TMN with the proved by the Animal Care Committee of Osaka Bioscience Institute. ADA inhibitor coformycin decreased the histamine release in Coformycin (Calbiochem), CPA, CPT, and adenosine (Sigma) were dissolved the FrCx and increased NREM sleep. These findings also suggest in artificial cerebrospinal fluid (composition: 140 mM NaCl, 3 mM KCl, 1 mM that endogenous adenosine suppressed the TMN to maintain the MgCl2, 1.3 mM CaCl2,2mMNa2HPO4, 0.2 mM NaH2PO4, and 4.5 mM glucose; NREM sleep pressure under the physiological condition. ADA pH 7.4). Aliquots were prepared and frozen at Ϫ20 °C and thawed immedi- may decompose local adenosine in the TMN to avoid excessive ately before each experiment. suppression of the TMN by adenosine. EEG and EMG Recordings. Under pentobarbital anesthesia (50 mg/kg, i.p.), rats We found that the A1R-mediated suppression of the histaminergic system increased both the duration (Fig. 2D) and the or mice were chronically implanted with EEG and EMG electrodes for poly- delta-power density (0.5–4 Hz) of NREM sleep (Fig. 2E). somnographic recordings (29, 33) and two guide cannulae (outer diameter, 0.55 mm) for drug application. The two guide cannulae were inserted ster- Similar effects were reported by Retey et al. (32), who found that eotaxically at positions of anteroposterior (AP) Ϫ4.52 mm; left–right (LR) Ϫ0.8 humans with a genetic variant of ADA, which is associated with mm; dorsoventral (DV) Ϫ6.2 mm from bregma; 3 mm above the TMN for rats the reduced metabolism of adenosine to inosine, exhibited as shown in Fig. 1J or AP Ϫ2.46 mm; LR Ϫ1.0 mm; DV Ϫ3.5 mm from bregma; enhanced deep sleep and slow-wave activity (power within 2 mm above the TMN for mice according to the atlases (37, 38). The EEG Ϸ0.5–5 Hz) during sleep. These observations suggest that the electrodes, both cannulae, and two stainless steel screws for anchorage were regulation of the histaminergic system by adenosine may be fixed to the skull with dental cement. Two stainless steel wires were placed involved in the duration and slow-wave activity of NREM sleep. into neck muscles as EMG electrodes. After 10 days for recovery, each animal was connected to an EEG/EMG recording cable in a soundproof recording We found that the A1R-mediated suppression of the histaminergic system extended NREM sleep duration without affect- chamber and habituated for 3 days before polygraphic recording. The EEG/ ing the total number of NREM sleep bouts in rats (Fig. 2D), EMG signals were amplified, filtered (EEG, 0.5–30 Hz; EMG, 20–200 Hz), digitized at a sampling rate of 128 Hz, and recorded by using SLEEPSIGN suggesting that the histaminergic systems were involved in the software ver. 3 (Kissei Comtec, Japan) (39). termination of NREM sleep by inducing wakefulness. This interpretation is consistent with our previous finding that H1R Pharmacological Treatment. Baseline recordings were taken for 24 h from KO mice showed prolonged duration of NREM sleep episodes 20:00. At 22:00 on the next day, CPA, adenosine, coformycin, CPT, or vehicle as compared with WT mice (33). In contrast, Takahashi et al. solution (0.4 ␮l) was injected bilaterally through the cannula (outer diameter, (34) proposed that histaminergic neurons play an important role 0.25 mm) inserted into the TMN. To verify correct placement of injection sites, in the maintenance of the vigilance but not in the induction of we performed immunohistochemical staining for HDC after EEG/EMG record- wakefulness per se, because they found that these neurons exhibit ings, which staining indicated that the injection cannula reached the TMN a pronounced delay in firing during transitions from sleep to (Fig. 1I). wakefulness, as judged from the extracellular single-unit recordings for the histaminergic tuberomammillary neurons in mice. Vigilance State Analysis. The vigilance states were classified off-line in 10-sec Thus, further analyses are expected to reveal the contribution of epochs into REM sleep, NREM sleep, and wakefulness by SLEEPSIGN automatically, according to the standard criteria (21). As a final step, defined sleep– the histaminergic system to the switching from NREM sleep to wake stages were examined visually and corrected if necessary. wakefulness. Direct activation of A1R’s or inhibition of ADA in the TMN In Vivo Microdialysis Procedure for Measurement of Histamine Release. Under increased NREM sleep but not REM sleep, suggesting that the urethane anesthesia (1.2 g/kg, i.p.), rats were implanted with a microdialysis histaminergic system is closely related to the regulation of probe (outer diameter, 0.22 mm; Eicom) into the FrCx (AP ϩ 3.2 mm; LR Ϫ1.0 NREM sleep and wakefulness and consistent with the evidence mm; DV Ϫ3.5 mm; membrane length, 2 mm) for monitoring extracellular that H1R KO mice show a decreased number of transition states histamine; and two stainless steel cannulae were introduced into the TMN for between wakefulness and NREM sleep but not REM sleep (33). administration of CPA, adenosine, or coformycin, as described above. The Alternatively, we previously showed that a PGD or A R microdialysis probe was perfused with artificial cerebrospinal fluid at a flow 2 2A ␮ agonist suppresses the histaminergic system (12, 13, 29) and rate of 2 l/min. Two hours after insertion of the microdialysis probe, dialy- ␮ increases both NREM and REM sleep (4, 11). For example, an sates were continuously collected from the FrCx at 20-min intervals (40 l each) from 1 h before injection of the drug to the TMN through the cannula to 4 h A2AR agonist (2 pmol/min) increased slow-wave sleep (NREM after the injection. The dialysates were kept at Ϫ20 °C until the histamine sleep) by 71% and paradoxical sleep (REM sleep) by 96% in rats assay could be performed by HPLC fluorometry (33). (11), whereas an A1R one at 1.5 nmol/side increased NREM sleep by 3.1-fold but did not change the amount of REM sleep Immunostaining for Detection of Adenosine Receptors in the Rat TMN. Under at all (Fig. 2C). The fact that a PGD2 or A2AR agonist activates deep pentobarbital (50 mg/kg, i.p.) anesthesia, rats were perfused via the sleep-active neurons in the VLPO (12, 13), which innervates heart with PBS followed by 10% formalin solution. The brains were removed histaminergic neurons in the TMN and other monoaminergic and immersed in the same fixative overnight at 4 °C, embedded into paraffin, arousal nuclei, such as serotonergic neurons in the dorsal raphe and cut into 10-␮m sections with a microtome. The sections were incubated nucleus and noradrenergic neurons in the locus ceruleus (28), with rabbit anti-A1R antibody (1:100; Sigma), raised against amino acid 309- suggests that inhibition of these arousal neurons may be involved 326 of the rat A1R; mouse anti-A1R antibody (1:100) (25) provided by T. in the increase in both NREM and REM sleep. These findings Ochiishi, National Institute of Advanced Industrial Science and Technology, Japan; mouse anti-A2AR antibody (1:1000; Upstate); or guinea pig anti-HDC suggest that the histaminergic system relatively selectively reg- antibody (1:5000; Euro-Diagnostica). They were then incubated with biotin- ulates the transition between wakefulness and NREM sleep. ylated anti-rabbit IgG antibody (1:400; Vector), anti-mouse IgG antibody In conclusion, the activation of A1R’s expressed in the TMN (1:400; Vector) or anti-guinea pig IgG antibody (1:400; Vector), followed by led to inhibition of the histaminergic system and promotion of horseradish peroxidase-conjugated avidin (Vectastain kit; Vector). These im- NREM sleep. munoreactivities were visualized by incubation with 3,3Ј-diaminobenzidine.

19996 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810926105 Oishi et al. Downloaded by guest on September 30, 2021 For double labeling for A1R and HDC, the sections were incubated with mine release, total sleep amounts, and the effect of CPA on the number of rabbit anti-A1R antibody (1:100) and guinea pig anti-HDC antibody (1:5000) NREM sleep bouts, stage transition, or power density were assessed by one- or followed by biotinylated anti-rabbit IgG antibody, Alexa-594-conjugated two-way ANOVA followed by Fisher’s probable least-squares difference test. streptavidin (1:400; Invitrogen), and fluorescein-conjugated anti-guinea pig In all cases, P Ͻ 0.05 was taken as the level of significance. IgG antibody (1:400; Jackson ImmunoResearch). For double labeling for A1R and ADA, the sections were incubated with rabbit anti-A1R antibody (1:100) ACKNOWLEDGMENTS. We thank Ms. N. Matsumoto of Osaka Bioscience and sheep anti-ADA antiserum (1:1000) (40) provided by Rodney E. Kellems, Institute for excellent technical support. This study was supported in part by University of Texas-Houston Medical School, followed by biotinylated anti- grants-in-aid for scientific research from the Program for Promotion of Basic rabbit IgG antibody, Alexa-594-conjugated streptavidin, and fluorescein- Research Activities for Innovative Biosciences of Japan, the Genome Network conjugated anti-sheep IgG antibody (1:400; Jackson ImmunoResearch). These Project from the Ministry of Education, Culture, Sports, Science and Technol- ogy, Japan (Y.U.), the National Natural Science Foundation of China signals were observed under a DM IRE2 fluorescence microscope (Leica). (30570581, 30625021, 30821002), National Basic Research Program of China Grant (2009CB5220004), Shanghai Leading Academic Discipline Project Statistical Analysis. All data were expressed as the mean Ϯ SEM (n ϭ 5–8). The (B119), Takeda Pharmaceutical, Ono Pharmaceutical (O.H.), Takeda Science statistical significance of time course data for sleep–wake profiles and hista- Foundation, and Osaka City.

1. Feldberg W, Sherwood SL (1954) Injections of drugs into the lateral ventricle of the cat. 21. Huang ZL, et al. (2005) Adenosine A2A, but not A1, receptors mediate the arousal effect J Physiol 123:148–167. of caffeine. Nat Neurosci 8:858–859. 2. Stenberg D, Porkka-Heiskanen T (2008) in Neurochemistry of Sleep and Wakefulness, 22. Nagy JI, LaBella LA, Buss M, Daddona PE (1984) Immunohistochemistry of adenosine eds Monti JM, Pandi-Perumal SR, Sinton CM (Cambridge Univ Press, Cambridge), pp deaminase: implications for adenosine neurotransmission. Science 224:166–168. 337–362. 23. Senba E, Daddona PE, Watanabe T, Wu JY, Nagy JI (1985) Coexistence of adenosine 3. Huang ZL, Urade Y, Hayaishi O (2007) Prostaglandins and adenosine in the regulation deaminase, histidine decarboxylase, and glutamate decarboxylase in hypothalamic of sleep and wakefulness. Curr Opin Pharmacol 7:33–38. neurons of the rat. J Neurosci 5:3393–3402. 4. Ueno R, Honda K, Inoue S, Hayaishi O (1983) Prostaglandin D2, a cerebral sleep- 24. Lin JS (2000) Brain structures and mechanisms involved in the control of cortical inducing substance in rats. Proc Natl Acad Sci USA 80:1735–1737. activation and wakefulness, with emphasis on the posterior hypothalamus and hista- 5. Onoe H, et al. (1988) Prostaglandin D2, a cerebral sleep-inducing substance in monkeys. minergic neurons. Sleep Med Rev 4:471–503. Proc Natl Acad Sci USA 85:4082–4086. 25. Ochiishi T, et al. (1999) Cellular localization of adenosine A1 receptors in rat forebrain: 6. Matsumura H, Takahata R, Hayaishi O (1991) Inhibition of sleep in rats by inorganic immunohistochemical analysis using adenosine A1 receptor-specific monoclonal an- selenium compounds, inhibitors of prostaglandin D synthase. Proc Natl Acad Sci USA tibody. J Comp Neurol 411:301–316. 88:9046–9050. 26. Murillo-Rodriguez E, Blanco-Centurion C, Gerashchenko D, Salin-Pascual RJ, Shiromani 7. Qu WM, et al. (2006) Lipocalin-type prostaglandin D synthase produces prostaglandin PJ (2004) The diurnal rhythm of adenosine levels in the basal forebrain of young and D2 involved in regulation of physiological sleep. Proc Natl Acad Sci USA 103:17949– old rats. Neuroscience 123:361–370. 17954. 27. Methippara MM, Kumar S, Alam MN, Szymusiak R, McGinty D (2005) Effects on sleep 8. Hayaishi O, Urade Y, Eguchi N, Huang ZL (2004) Genes for prostaglandin D synthase and of microdialysis of adenosine A1 and A2a receptor analogs into the lateral preoptic area receptor as well as adenosine A2A receptor are involved in the homeostatic regulation of rats. Am J Physiol Regul Integr Comp Physiol 289:R1715–1723. of NREM sleep. Arch Ital Biol 142:533–539. 28. Sherin JE, Elmquist JK, Torrealba F, Saper CB (1998) Innervation of histaminergic 9. Urade Y, et al. (1993) Dominant expression of mRNA for prostaglandin D synthase in tuberomammillary neurons by GABAergic and galaninergic neurons in the ventrolat- leptomeninges, choroid plexus, and oligodendrocytes of the adult rat brain. Proc Natl eral preoptic nucleus of the rat. J Neurosci 18:4705–4721. Acad Sci USA 90:9070–9074. 29. Hong ZY, et al. (2005) An adenosine A2A receptor agonist induces sleep by increasing 10. Mizoguchi A, et al. (2001) Dominant localization of prostaglandin D receptors on GABA release in the tuberomammillary nucleus to inhibit histaminergic systems in rats. arachnoid trabecular cells in mouse basal forebrain and their involvement in the J Neurochem 92:1542–1549. regulation of non-rapid eye movement sleep. Proc Natl Acad Sci USA 98:11674–11679. 30. Jo YH, Schlichter R (1999) Synaptic corelease of ATP and GABA in cultured spinal 11. Satoh S, Matsumura H, Suzuki F, Hayaishi O (1996) Promotion of sleep mediated by the neurons. Nat Neurosci 2:241–245. A2a-adenosine receptor and possible involvement of this receptor in the sleep induced 31. Jo YH, Role LW (2002) Coordinate release of ATP and GABA at in vitro synapses of by prostaglandin D2 in rats. Proc Natl Acad Sci USA 93:5980–5984. lateral hypothalamic neurons. J Neurosci 22:4794–4804. 12. Scammell T, et al. (1998) Activation of ventrolateral preoptic neurons by the somnogen 32. Retey JV, et al. (2005) A functional genetic variation of adenosine deaminase affects prostaglandin D2. Proc Natl Acad Sci USA 95:7754–7759. the duration and intensity of deep sleep in humans. Proc Natl Acad Sci USA 102:15676– 13. Scammell TE, et al. (2001) An adenosine A2a agonist increases sleep and induces Fos in 15681. ventrolateral preoptic neurons. Neuroscience 107:653–663. 14. Hayaishi O, Urade Y (2008) in Neurochemistry of Sleep and Wakefulness, eds Monti JM, 33. Huang ZL, et al. (2006) Altered sleep-wake characteristics and lack of arousal response Pandi-Perumal SR, Sinton CM (Cambridge Univ Press, Cambridge), pp 363–386. to H3 receptor antagonist in histamine H1 receptor knockout mice. Proc Natl Acad Sci 15. Fredholm BB, IJzerman AP, Jacobson KA, Klotz KN, Linden J (2001) International Union USA 103:4687–4692. of Pharmacology. XXV. Nomenclature and classification of adenosine receptors. Phar- 34. Takahashi K, Lin JS, Sakai K (2006) Neuronal activity of histaminergic tuberomammil- macol Rev 53:527–552. lary neurons during wake-sleep states in the mouse. J Neurosci 26:10292–10298. 16. Basheer R, Strecker RE, Thakkar MM, McCarley RW (2004) Adenosine and sleep-wake 35. Johansson B, et al. (2001) Hyperalgesia, anxiety, and decreased hypoxic neuroprotec- regulation. Prog Neurobiol 73:379–396. tion in mice lacking the adenosine A1 receptor. Proc Natl Acad Sci USA 98:9407–9412. 17. Strecker RE, et al. (2000) Adenosinergic modulation of basal forebrain and preoptic/ 36. Inoue I, et al. (1996) Impaired locomotor activity and exploratory behavior in mice anterior hypothalamic neuronal activity in the control of behavioral state. Behav Brain lacking histamine H1 receptors. Proc Natl Acad Sci USA 93:13316–13320. Res 115:183–204. 37. Paxinos G, Watson C (1997) The Rat Brain in Stereotaxic Coordinates (Academic, San 18. Alam MN, Szymusiak R, Gong H, King J, McGinty D (1999) Adenosinergic modulation Diego). of rat basal forebrain neurons during sleep and waking: neuronal recording with 38. Paxinos G, Franklin K (2001) The Mouse Brain in Stereotaxic Coordinates (Academic, microdialysis. J Physiol 521:679–690. San Diego). 19. Thakkar MM, Winston S, McCarley RW (2003) A1 receptor and adenosinergic homeo- 39. Kohtoh S, et al. (2008) Algorithm for sleep scoring in experimental animals based on static regulation of sleep-wakefulness: effects of antisense to the A1 receptor in the fast Fourier transform power spectrum analysis of the electroencephalogram. Sleep cholinergic basal forebrain. J Neurosci 23:4278–4287. Biol Rhythms 6:163–171. 20. Urade Y, et al. (2003) Sleep regulation in adenosine A2A receptor-deficient mice. 40. Ingolia DE, et al. (1985) Purification and characterization of adenosine deaminase from Neurology 61:S94–96. a genetically enriched mouse cell line. J Biol Chem 260:13261–13267. NEUROSCIENCE

Oishi et al. PNAS ͉ December 16, 2008 ͉ vol. 105 ͉ no. 50 ͉ 19997 Downloaded by guest on September 30, 2021