J Pharmacol Sci 121, 157 – 171 (2013) Journal of Pharmacological Sciences © The Japanese Pharmacological Society Full Paper Stimulatory and Inhibitory Roles of Brain 2-Arachidonoylglycerol in Bombesin-Induced Central Activation of Adrenomedullary Outflow in Rats

Takahiro Shimizu1,*, Kenjiro Tanaka1, and Kunihiko Yokotani1 1Department of Pharmacology, School of Medicine, Kochi University, Nankoku, Kochi 783-8505, Japan

Received September 9, 2012; Accepted December 19, 2012

Abstract. 2-Arachidonoylglycerol (2-AG) is recognized as a potent endocannabinoid, which

reduces synaptic transmission through CB1 receptors, and is hydrolyzed by mono- acylglycerol (MGL) to (AA), a substrate. We already reported that centrally administered MGL and cyclooxygenase inhibitors each reduced the intra- cerebroventricularly (i.c.v.) administered bombesin-induced secretion of adrenal catecholamines,

while a centrally administered CB1-antagonist potentiated the response, indirectly suggesting bidirectional roles of brain 2-AG (stimulatory and inhibitory roles) in the bombesin-induced response. In the present study, we separately examined these bidirectional roles using 2-AG and 2-AG ether (2-AG-E) (stable 2-AG analog for MGL) in rats. 2-AG (0.5 μmol/animal, i.c.v.), but not 2-AG-E (0.5 μmol/animal, i.c.v.), elevated basal plasma catecholamines with JZL184 (MGL inhibitor)- and indomethacin (cyclooxygenase inhibitor)-sensitive brain mechanisms. 2-AG-E (0.1 μmol/animal, i.c.v.) effectively reduced the bombesin (1 nmol/animal, i.c.v.)-induced elevation of

plasma catecholamines with (CB1 antagonist)-sensitive brain mechanisms. Immuno- histochemical studies demonstrated the bombesin-induced activation of diacylglycerol lipase α (2-AG-producing )-positive spinally projecting neurons in the hypothalamic paraventricular nucleus, a control center of central adrenomedullary outflow. These results directly indicate bidirectional roles of brain 2-AG, a stimulatory role as an AA precursor and an inhibitory role as an endocannabinoid, in the bombesin-induced central adrenomedullary outflow in rats.

Keywords: 2-arachidonoylglycerol, , cyclooxygenase,

cannabinoid CB1 , central adrenomedullary outflow

Introduction splanchnic nerves, but intravenous administration of bombesin did not affect the nerve activities in rats (3). Bombesin, a tetradecapeptide originally isolated from We also reported that i.c.v. administered bombesin the skin of the European frog Bombina bombina (1), is a elevated plasma levels of catecholamines (noradrenaline for bombesin BB receptors. This peptide is not and adrenaline) and the elevations were both abolished expressed in mammals, while the mammalian counter- by acute bilateral adrenalectomy in rats (4). Furthermore, parts and the receptors are widely distributed in the we reported that the peptide-induced elevations of plasma mammalian brain (2). Using bombesin, we have been catecholamines were reduced by central pretreatment examining central regulatory mechanisms of sympatho- with (an inhibitor of cyclooxygenase) (5) and adrenomedullary outflow. Previously, we reported that that central pretreatment with indomethacin (another intracerebroventricularly (i.c.v.) administered bombesin inhibitor of cyclooxygenase) reduced the centrally increased nerve activity of the adrenal branch of the administered arachidonic acid (a representative substrate of cyclooxygenase)-induced elevation of plasma cate- *Corresponding author. [email protected] cholamines in rats (6). In addition, we reported that Published online in J-STAGE on February 2, 2013 (in advance) centrally administered , produced from doi: 10.1254/jphs.12208FP arachidonic acid by cyclooxygenase-mediated mecha-

157 158 T Shimizu et al nisms, elevated plasma levels of catecholamines in rats uptake-inhibitor of endocannabinoid) potentiated or (7, 8). These results suggest that the cyclooxygenase- reduced the centrally administered bombesin–induced mediated production of active arachidonic acid metabo- elevation of plasma catecholamines, respectively (16). lites (prostanoids) in the brain is involved in the These results suggest that endogenously generated brain bombesin-induced activation of central adrenomedullary endocannabinoid (probably 2-AG) plays bidirectional outflow in rats. roles (a stimulatory role as a precursor of prostanoids Arachidonic acid is found in cellular membrane and an inhibitory role as an endocannabinoid) in the phospholipids in position 2. A2 (PLA2) bombesin-induced elevation of plasma catecholamines hydrolyzes this phospholipid sn-2 ether bond, thereby in rats. directly releasing arachidonic acid (9, 10). However, we In the present study, we attempted to separately reported that central pretreatment with mepacrine (an examine the bidirectional roles of 2-AG by examining inhibitor of PLA2) had no effect on the centrally admin- the effect of centrally administered 2-AG itself on the istered bombesin–induced elevation of plasma catechol- basal plasma catecholamines levels and by examining amines (11), indicating the involvement of a brain the effect of centrally administered 2-AG ether (a stable phospholipase other than PLA2 in the bombesin-induced analog of 2-AG for monoacylglycerol lipase) on the response. A (PLC)-mediated arachidonic centrally administered bombesin–induced elevation of acid producing pathway is also reported: 1) PLC cleaves plasma catecholamines in rats. the phosphodiester bond of membrane phospholipids, resulting in the formation of diacylglycerol (12); 2) Materials and Methods diacylglycerol containing arachidonic acid in position 2 can be hydrolyzed to 2-arachidonoylglycerol (2-AG) by Animals sn-1 selective diacylglycerol lipase (13); and 3) 2-AG All animal experiments were conducted in compliance can be further hydrolyzed by monoacylglycerol lipase to with the guiding principles for the care and use of free arachidonic acid (14, 15). Recently, using inhibitors laboratory animals approved by Kochi University, which of each lipase described above, we indirectly indicated are in accordance with the “Guidelines for Proper the involvement of brain arachidonic acid generated Conduct of Animal Experiments” from the Science by the brain PLC-, diacylglycerol lipase-, and mono- Council of Japan. Male Wistar rats weighing about 350 acylglycerol lipase–mediated pathway in the bombesin- g were used (Japan SLC, Inc., Hamamatsu). Two rats induced responses in rats (11, 16). Actually, this PLC- were kept in one cage, and they were maintained in an mediated pathway has been reported to be a source of air-conditioned room at 22°C – 24°C under a constant arachidonic acid from 2-AG in many cells such as bovine day–night rhythm (14/10 h light–dark cycle, lights on at coronary endothelial cells (17), rabbit aorta (18), and 05:00) for more than 2 weeks and given food (laboratory murine melanoma cells (19). These findings suggest that chow, CE-2; Clea Japan, Hamamatsu) and water ad brain 2-AG generated by the PLC- and diacylglycerol libitum. lipase–mediated pathway functions as a source of brain arachidonic acid during the bombesin-induced elevation Experimental procedures for intracerebroventricular of plasma catecholamines in rats. administration Interestingly, this arachidonic acid source 2-AG has In the morning (09:00 – 10:00), the femoral vein was recently been recognized as a major brain endocannabi- cannulated for infusion of saline (1.2 ml/h) and the noid for cannabinoid CB receptors (20, 21). In the ner- femoral artery was cannulated for collecting blood vous system, 2-AG produced on demand from membrane samples, under urethane anesthesia (1.2 g/kg, i.p.). After phospholipids of postsynaptic neurons acts on pre- these procedures, the rat was placed in a stereotaxic synaptic cannabinoid CB1 receptors, thereby inhibiting apparatus for the brain until the end of each experiment, the release of neurotransmitters (21, 22). The 2-AG as shown in our previous papers (30, 31). The skull was released into the synaptic cleft seems to be rapidly drilled for intracerebroventricular administration of test inactivated by uptake into neurons (23). This 2-AG– reagents using a stainless-steel cannula (0.3 mm in outer induced retrograde signaling process seems to be involved diameter). The stereotaxic coordinates of the tip of the in widespread effects, including synaptic plasticity (24), cannula were as follows (in mm): AP −0.8, L 1.5, V 4.0 analgesia (25), neuroprotection and neurotoxicity (26, (AP, anterior from the bregma; L, lateral from the 27), food intake (24, 28), and also in the baroreflex- midline; V, below the surface of the brain), according to evoked central sympathetic modulation (29). Recently, the rat brain atlas (32). Three hours were allowed to we reported that central pretreatment with AM 251 (an elapse before the application of reagents. antagonist of cannabinoid CB1 receptors) or AM 404 (an Brain 2-AG in Adrenomedullary Outflow 159

Drug administration detector, +450 mV potential against an Ag/AgCl refer- 2-AG or 2-AG ether dissolved in 100% N,N-dimethyl- ence electrode; column, Eicompack CA-50DS, 2.1 formamide (DMF) was slowly administered into the mm × 150 mm (Eicom); mobile phase, 0.1 M NaH2PO4− right lateral ventricle in a volume of 2.5 μl/animal using Na2HPO4 buffer (pH 6.0) containing 50 mg/l disodium the cannula connected to a 10-μl Hamilton syringe, and EDTA, 0.75 g/l sodium 1-octanesulfonate, and 15% the cannula was retained until the end of the experiment methanol at a flow rate of 0.18 ml/min; injection volume, (Figs. 1 and 4). 40 μl. The amount of catecholamines in each sample was JZL184 (a selective inhibitor of monoacylglycerol calculated using the peak height ratio relative to that of lipase) dissolved in 3 μl DMF/animal or water-soluble 3,4-dihydroxybenzylamine. By this assay, coefficients of indomethacin-Na (indomethacin) (an inhibitor of cyclo- variation for the intra- and inter-assay were 3.0% and oxygenase) dissolved in 5 μl sterile saline/animal was 3.7%, respectively, and 0.5 pg of noradrenaline and i.c.v. administered using the cannula connected to a 10-μl adrenaline was accurately determined. Hamilton syringe, which was retained in the ventricle for 15 min to avoid the leakage of these reagents and then Immunohistochemical study on the spinally projecting removed from the ventricle (Figs. 2 and 3). 2-AG was neurons of the hypothalamic paraventricular nucleus then slowly administered as described above 30 min after (PVN) the application of the reagents (Figs. 2 and 3). For labeling PVN neurons innervating the spinal cord, Bombesin dissolved in 10 μl sterile saline/animal was a mono-synaptic retrograde tracer Fluoro-Gold was slowly administered into the ventricle using the cannula microinjected into the intermediolateral cell column connected to a 50-μl Hamilton syringe, and the cannula (IML) of the thoracic spinal cord according to a pre- was retained until the end of the experiment (Figs. 5 – 7). viously reported method (34, 35). Briefly, under pento-

Rimonabant (an antagonist of cannabinoid CB1 recep- barbital anesthesia (50 mg/kg, i.p.), the rats were placed tors) and/or 2-AG ether, dissolved in 3 μl DMF/animal, in a stereotaxic apparatus for the spinal cord until the end were i.c.v. administered 30 min before the application of the surgery. The spinal cord was exposed by dorsal of bombesin, using the cannula connected to a 10-μl laminectomy through a back midline incision with an Hamilton syringe, which was retained for 15 min to aseptic surgical procedure. Fluoro-Gold (4% in sterile avoid the leakage of these reagents and then removed saline) was microinjected bilaterally into the IML (0.5- from the ventricle (Figs. 5 and 6). mm lateral from the midline and 1.0-mm below the surface of the spinal cord) at the T8 level in a volume of Measurement of plasma catecholamines 200 nl on each side using a 30-G stainless-steel cannula Blood samples (250 μl) were collected through an (0.3 mm in outer diameter) at a rate of 40 nl/min. Then arterial catheter and were preserved on ice during experi- the muscle and skin were closed, and the rats were ments. Plasma was prepared immediately after the final returned to their home-cages. The exact location of the sampling. Catecholamines in the plasma were extracted spinal cord injection was verified by Nissl staining at by the method of Anton and Sayre (33) with a slight the end of each experiment described below. modification and were assayed electrochemically with Fourteen days after the Fluoro-Gold injection, the rats high performance liquid chromatography (HPLC) (30). were anesthetized with urethane (1.2 g/kg, i.p.) and the Briefly, after centrifugation (1500 × g for 10 min, at 4°C), femoral vein was cannulated for infusion of saline (1.2 the plasma (100 μl) was transferred to a centrifuge ml/h). Then, the rats were placed in a stereotaxic tube containing 30 mg of activated alumina, 2 ml of apparatus for the brain, as shown in our previous papers water deionized in a MilliQ water purification system (34, 35). The skull was drilled for intracerebroventricular (Millipore, Billerica, MA, USA), 1 ml of 1.5 M Tris administration of bombesin or vehicle into the right buffer (pH 8.6) containing 0.1 M disodium EDTA, and 1 lateral ventricle using a stainless-steel cannula (0.3 mm ng of 3,4-dihydroxybenzylamine as an internal standard. in outer diameter) in the same manner as described The tube was shaken for 10 min and the alumina was above. Three hours were allowed to elapse before the washed three times with 4 ml of ice-cold deionized water. start of the intracerebroventricular administration of Then, catecholamines adsorbed onto the alumina were bombesin or vehicle. eluted with 300 μl of 4% acetic acid containing 0.1 mM At 1 h after the administration of bombesin (1 nmol/ disodium EDTA. A pump (EP-300; Eicom, Kyoto), a animal, i.c.v.) or vehicle (10 μl sterile saline/animal, sample injector (Model-231XL; Gilson, Villiers-le-Bel, i.c.v.), the rats were perfused through the left cardiac France), and an electrochemical detector (ECD-300, ventricle with 100 ml of 0.1 M phosphate-buffered saline Eicom) equipped with a graphite electrode were used (pH 7.4) followed by 350 ml of ice-cold 4% paraformal- with HPLC. Analytical conditions were as follows: dehyde in 0.1 M phosphate buffer. Brains and spinal 160 T Shimizu et al

Fig. 1. Effect of centrally administered 2-arachidonoylglycerol (2-AG) on plasma catecholamines. Vehicle (2.5 μl DMF/animal) or 2-AG (0.1 or 0.5 μmol/animal) was i.c.v. administered. A) Increments of plasma catecholamines above the basal level. ΔNoradrenaline and ΔAdrenaline: increments of noradrenaline and adrenaline above the basal level. Arrow indicates the administration of vehicle or 2-AG. The actual values for noradrenaline and adrenaline at 0 min were 177 ± 19 and 99 ± 11 pg/ml (n = 15). B) The area under the curve (AUC) of the elevation of plasma catecholamines above the basal level for each group is expressed as pg/1 h. Each point represents the mean ± S.E.M. *P < 0.05, when compared with the Bonferroni method to the vehicle-treated group.

cords were immediately removed, post fixed in the (1:1000, respectively) for 2 h at room temperature in the same fixative overnight, equilibrated in 0.1 M phosphate dark and washed in 0.05 M Tris-buffered saline again. buffer containing 20% sucrose at 4°C, coronally cut on a The sections were then mounted on silane-coated slides freezing cryostat (Cryostat HM505E; Thermo Scientific, and cover slipped with VECTASHIELD® mounting Yokohama) at a thickness of 20 μm, and washed in 0.05 medium. All antisera were diluted in 0.05 M Tris-buffered M Tris-buffered saline (pH 7.4). saline containing 0.25% Triton X-100 and 0.3% bovine Immunohistochemical analysis was performed with a serum albumin. Photographs from brain sections were slight modification of previously reported methods captured using a digital camera (DP70; Olympus, Tokyo) (34 – 36). Free-floating sections were incubated in a attached to a fluorescent microscope (AX70, Olympus) mixed diluent of a rabbit polyclonal antibody against with appropriate filter sets that allow the separate diacylglycerol lipase α (1:200) and a goat polyclonal visualization of Cy3 (for diacylglycerol lipase α), FITC antibody against Fos (1:100) for 48 h at 4°C. After (for Fos), and ultraviolet excitation (for Fluoro-Gold). washing in 0.05 M Tris-buffered saline, the sections Fluoro-Gold–labeled neurons were visualized under were incubated in a mixed diluent of Cy3-conjugated ultraviolet illumination. donkey anti-rabbit IgG and fluorescein isothiocyanate (FITC)-conjugated donkey anti-goat IgG antibodies Brain 2-AG in Adrenomedullary Outflow 161

Fig. 2. Effect of JZL184 on the centrally administered 2-AG–induced elevation of plasma catecholamines. JZL184 (JZL) (a selective inhibitor of monoacylglycerol lipase) (0.7 or 1.4 μmol/animal) or vehicle-1 (V-1) (3 μl DMF/animal) was i.c.v. admin- istered 30 min before the administration of 2-AG (0.5 μmol/animal, i.c.v.) or vehicle-2 (V-2) (2.5 μl DMF/animal, i.c.v.). A) Increments of plasma catecholamines above the basal level. Arrows indicate the administration of JZL/V-1 and 2-AG/V-2. The actual values for noradrenaline and adrenaline at 0 min were 154 ± 23 and 84 ± 15 pg/ml in the V-1–pretreated group (n = 10), 257 ± 77 and 149 ± 43 pg/ml in the JZL (0.7 μmol/animal)-pretreated group (n = 4), and 74 ± 13 and 126 ± 27 pg/ml in the JZL (1.4 μmol/animal)-pretreated group (n = 8), respectively. B) AUC of the elevation of plasma catecholamines above the basal level for each group. *P < 0.05, when compared with the Bonferroni method to the V-1– and 2-AG–treated group. Other condi- tions are the same as those of Fig. 1.

Treatment of data and statistics (Peptide Institute, Osaka); 2-AG (Enzo Life Science, Increments of plasma catecholamines above the basal Plymouth Meeting, PA, USA); water-soluble indo- level at each time period are expressed as pg/ml (Figs. methacin sodium trihydrate (indomethacin) (a kind gift 1 – 6). The area under the curve (AUC) is also expressed from Merck, Rahway, NJ, USA); 2-AG ether, JZL184 as pg/1 h (Figs. 1 – 4) or pg/2 h (Figs. 5 and 6). The [4-nitrophenyl 4-(dibenzo[d][1,3]dioxol-5-yl(hydroxy) number of animals in each group is shown in these methyl)piperidine-1-carboxylate], and rimonabant [5- figures (Figs. 1 – 6). All values are expressed as (4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl- means ± S.E.M. Statistical differences were determined N-1-piperidinyl-1H-pyrazole-3-carboxamide] (Cayman using repeated-measure (treatment × time) or one-way Chemical, Ann Arbor, MI, USA); anti-diacylglycerol analysis of variance (ANOVA), followed by post hoc lipase α rabbit polyclonal antibody (sc-133308) and analysis with the Bonferroni method. P values less than anti-Fos goat polyclonal antibody (sc-52-g) (Santa Cruz 0.05 were taken to indicate statistical significance. Biotechnology, Santa Cruz, CA, USA); Cy3-conjugated donkey anti-rabbit IgG and FITC-conjugated donkey Materials anti-goat IgG antibodies (Jackson ImmunoResearch The following materials were used: synthetic bombesin Laboratories, West Grove, PA, USA); VECTASHIELD® 162 T Shimizu et al

Fig. 3. Effect of indomethacin on the centrally administered 2-AG–induced elevation of plasma catecholamines. Indomethacin (IND) (an inhibitor of cyclooxygenase) (0.6 or 1.2 μmol/animal) or vehicle-1 (V-1) (5 μl saline/animal) was i.c.v. administered 30 min before the administration of 2-AG (0.5 μmol/animal, i.c.v.) or vehicle-2 (V-2) (2.5 μl DMF/animal, i.c.v.). A) Increments of plasma catecholamines above the basal level. Arrows indicate the administration of IND/V-1 and 2-AG/V-2. The actual values for noradrenaline and adrenaline at 0 min were 171 ± 27 and 148 ± 31 pg/ml in the V-1–pretreated group (n = 12), 104 ± 1 and 151 ± 41 pg/ml in the IND (0.6 μmol/animal)-pretreated group (n = 5), and 146 ± 29 and 181 ± 29 pg/ml in the IND (1.2 μmol/ animal)-pretreated group (n = 10), respectively. B) AUC of the elevation of plasma catecholamines above the basal level for each group. *P < 0.05, when compared with the Bonferroni method to the V-1– and 2-AG–treated group. Other conditions are the same as those of Figs. 1 and 2. mounting medium (Vector Laboratories, Burlingame, interaction between these two factors [noradrenaline, CA, USA); Fluoro-Gold (Fluorochrome, Denver, CO, F(8,47) = 9.00, P < 0.05; adrenaline, F(8,44) = 3.30, USA). All other reagents were of the highest grade avail- P < 0.05] (Fig. 1A). One-way ANOVA also revealed able (Nacalai Tesque, Kyoto). significant effects of the treatments on plasma catechol- amines [noradrenaline, F(2,12) = 19.77, P < 0.05; adrena- Results line, F(2,11) = 10.11, P < 0.05] (Fig. 1B). Treatment with vehicle (2.5 μl DMF/animal, i.c.v.) Effect of centrally administered 2-AG on plasma had no significant effect on the basal plasma levels of catecholamines noradrenaline and adrenaline (Fig. 1: A and B). 2-AG For plasma levels of catecholamines (noradrenaline [0.1 μmol (38 μg)/animal, i.c.v.] had little effect on and adrenaline), repeated-measure ANOVA showed the plasma levels of both catecholamines, but a larger significant effects of treatment with 2-AG [noradrena- dose of this reagent [0.5 μmol (189 μg)/animal, i.c.v.] line, F(2,12) = 21.72, P < 0.05; adrenaline, F(2,12) = significantly elevated the plasma levels of both catechol- 13.09, P < 0.05], time [noradrenaline, F(4,47) = 12.45, amines (adrenaline > noradrenaline) (Fig. 1: A and B). P < 0.05; adrenaline, F(4,44) = 4.00, P < 0.05], and the The responses of noradrenaline and adrenaline reached Brain 2-AG in Adrenomedullary Outflow 163

Fig. 4. Effect of centrally administered 2-AG ether on plasma catecholamines. Vehicle (2.5 μl DMF/animal) or 2-AG ether (2-AG-E) (a stable analog of 2-AG for monoacylglycerol lipase) (0.1 or 0.5 μmol/animal) was i.c.v. administered. A) Increments of plasma catecholamines above the basal level at each time point are expressed as pg/ml. Arrow indicates the administration of vehicle or 2-AG-E. The vehicle-treated group is the same as that in Fig. 1. The actual values for noradrenaline and adrenaline at 0 min were 181 ± 27 and 150 ± 23 pg/ml (n = 14). B) AUC of the elevation of plasma catecholamines above the basal level for each group. Other conditions are the same as those of Figs. 1 – 3. a maximum at 5 and 10 min, respectively, after the Treatments with vehicle-1 (3 μl DMF/animal, i.c.v.) administration of 2-AG (Fig. 1A). Intravenous admin- and vehicle-2 (2.5 μl DMF/animal, i.c.v.) had no effect istration of 2-AG (0.5 μmol/animal), however, had no on the plasma levels of noradrenaline and adrenaline effect on plasma levels of catecholamines (data not shown). (Fig. 2: A and B). Treatments with JZL184 [1.4 μmol (730 μg)/animal, i.c.v.] and vehicle-2 also had no effect Effect of JZL184 on the centrally administered 2-AG–in- on the plasma levels of both catecholamines (Fig. 2: A duced elevation of plasma catecholamines and B). For plasma levels of catecholamines (noradrenaline Pretreatment with JZL184 effectively reduced the and adrenaline), repeated-measure ANOVA showed 2-AG (0.5 μmol/animal, i.c.v.)-induced elevation of significant effects of treatments with JZL184 and 2-AG plasma noradrenaline and adrenaline in a dose-dependent [noradrenaline, F(4,17) = 6.40, P < 0.05; adrenaline, manner [0.7 and 1.4 μmol (365 and 730 μg)/animal, F(4,17) = 12.55, P < 0.05], time [noradrenaline, F(4,67) = i.c.v.] (Fig. 2: A and B). 2.99, P < 0.05; adrenaline, F(4,63) = 5.14, P < 0.05], and the interaction between these two factors [noradrenaline, Effect of indomethacin on the centrally administered F(16,67) = 2.01, P < 0.05; adrenaline, F(16,63) = 2.47, 2-AG–induced elevation of plasma catecholamines P < 0.05] (Fig. 2A). One-way ANOVA also revealed For plasma levels of catecholamines (noradrenaline significant effects of the treatments on plasma catechol- and adrenaline), repeated-measure ANOVA showed amines [noradrenaline, F(4,16) = 7.26, P < 0.05; adrena- significant effects of treatments with indomethacin line, F(4,16) = 11.92, P < 0.05] (Fig. 2B). and 2-AG [noradrenaline, F(4,22) = 10.62, P < 0.05; 164 T Shimizu et al adrenaline, F(4,22) = 5.24, P < 0.05], time [noradrena- on the plasma levels of noradrenaline and adrenaline line, F(4,83) = 2.63, P < 0.05; adrenaline, F(4,75) = 3.03, (Fig. 5: A and B). Treatments with 2-AG ether [0.1 μmol P < 0.05], and the interaction between these two factors (36 μg)/animal, i.c.v.] and vehicle-2 also had no effect [noradrenaline, F(16,83) = 3.93, P < 0.05; adrenaline, on the plasma levels of both catecholamines (Fig. 5: A F(16,75) = 2.94, P < 0.05] (Fig. 3A). One-way ANOVA and B). also revealed significant effects of the treatments on Since we previously reported that bombesin (0.1, 1, plasma catecholamines [noradrenaline, F(4,21) = 12.17, and 10 nmol/animal, i.c.v.) dose-dependently elevated P < 0.05; adrenaline, F(4,17) = 12.56, P < 0.05] (Fig. 3B). plasma levels of noradrenaline and adrenaline (37), we Treatments with vehicle-1 (5 μl saline/animal, i.c.v.) used a sub-maximum dose of 1 nmol/animal in the pres- and vehicle-2 (2.5 μl DMF/animal, i.c.v.) had no effect ent study. Administration of bombesin (1 nmol/animal, on the plasma levels of noradrenaline and adrenaline i.c.v.) gradually elevated the plasma levels of noradrena- (Fig. 3: A and B). Treatments with indomethacin [1.2 line and adrenaline (adrenaline > noradrenaline) (Fig. 5A). μmol (500 μg)/animal, i.c.v.] and vehicle-2 also had The responses of noradrenaline and adrenaline reached a no effect on the plasma levels of both catecholamines maximum at 60 and 30 min after the administration of (Fig. 3: A and B). this peptide, respectively. On the other hand, intravenous Pretreatment with indomethacin effectively reduced administration of bombesin (1 nmol/animal) had no the 2-AG (0.5 μmol/animal, i.c.v.)-induced elevation of effect on plasma levels of catecholamines (data not plasma catecholamines in a dose-dependent manner [0.6 shown). Pretreatment with 2-AG ether [0.01 and 0.1 and 1.2 μmol (250 and 500 μg)/animal, i.c.v.] (Fig. 3: A μmol (3.6 and 36 μg)/animal, i.c.v.] dose-dependently and B). reduced the bombesin (1 nmol/animal, i.c.v.)-induced elevation of plasma catecholamines (Fig. 5: A and B). Effect of centrally administered 2-AG ether on plasma catecholamines Effect of rimonabant on the 2-AG ether-induced inhibi- Repeated-measure ANOVA revealed a significant tion of the centrally administered bombesin–induced el- effect of the treatment with 2-AG ether on plasma evation of plasma catecholamines adrenaline, but not on noradrenaline [noradrenaline, For plasma levels of catecholamines (noradrenaline F(2,11) = 0.15, P = 0.86; adrenaline, F(2,11) = 5.78, and adrenaline), repeated-measure ANOVA showed P < 0.05] (Fig. 4A). However, one-way ANOVA revealed significant effects of treatments with (rimonabant and/or no significant effects of the treatments on both plasma 2-AG ether) and bombesin [noradrenaline, F(5,26) = catecholamines [noradrenaline, F(2,10) = 0.24, P = 0.79; 20.93, P < 0.05; adrenaline, F(5,26) = 32.78, P < 0.05], adrenaline, F(2,9) = 1.29, P = 0.32] (Fig. 4B). time [noradrenaline, F(5,126) = 34.87, P < 0.05; adrena- 2-AG ether [0.1 and 0.5 μmol (36 and 182 μg)/animal, line, F(5,123) = 28.89, P < 0.05], and the interaction i.c.v.] had no significant effect on the plasma levels of between these two factors [noradrenaline, F(25,126) = noradrenaline and adrenaline at each time point with the 3.61, P < 0.05; adrenaline, F(25,123) = 4.15, P < 0.05] Bonferroni method (Fig. 4A). (Fig. 6A). One-way ANOVA also revealed significant effects of the treatments on plasma catecholamines Effect of 2-AG ether on the centrally administered [noradrenaline, F(5,25) = 19.37, P < 0.05; adrenaline, bombesin–induced elevation of plasma catecholamines F(5,23) = 20.35, P < 0.05] (Fig. 6B). For plasma levels of catecholamines (noradrenaline Treatments with rimonabant [90 nmol (42 μg)/animal, and adrenaline), repeated-measure ANOVA showed, i.c.v.]/2-AG ether [0.1 μmol (36 μg)/animal, i.c.v.] and significant effects of treatments with 2-AG ether and vehicle-2 had no effect on the plasma levels of nor- bombesin [noradrenaline, F(4,21) = 20.47, P < 0.05; adrenaline and adrenaline (Fig. 6: A and B). adrenaline, F(4,21) = 24.25, P < 0.05], time [noradrena- As shown in Fig. 5, A and B, 2-AG ether (0.1 μmol/ line, F(5,100) = 31.26, P < 0.05; adrenaline, F(5,96) = animal, i.c.v.) significantly reduced the bombesin (1 33.81, P < 0.05], and the interaction between these two nmol/animal, i.c.v.)-induced elevation of plasma cate- factors [noradrenaline, F(20,100) = 5.05, P < 0.05; cholamines (Fig. 6: A and B). Pretreatment with a adrenaline, F(20,96) = 6.90, P < 0.05] (Fig. 5A). One- smaller dose of rimonabant [30 nmol (14 μg)/animal, way ANOVA also revealed significant effects of the i.c.v.] had no effect on the 2-AG ether–induced inhibi- treatments on plasma catecholamines [noradrenaline, tion of the bombesin-induced response, while pretreat- F(4,19) = 16.61, P < 0.05; adrenaline, F(4,20) = 15.22, ment with a higher dose of rimonabant (90 nmol/animal, P < 0.05] (Fig. 5B). i.c.v.) significantly attenuated the 2-AG ether–induced Treatments with vehicle-1 (3 μl DMF/animal, i.c.v.) inhibitory response (Fig. 6: A and B). and vehicle-2 (10 μl saline/animal, i.c.v.) had no effect Brain 2-AG in Adrenomedullary Outflow 165

Fig. 5. Effect of 2-AG ether on the centrally administered bombesin–induced elevation of plasma catecholamines. 2-AG ether (2-AG-E) (a stable analog of 2-AG) (0.01 or 0.1 μmol/animal) or vehicle-1 (V-1) (3 μl DMF/animal) was i.c.v. administered 30 min before the administration of bombesin (BB) (1 nmol/animal, i.c.v.) or vehicle-2 (V-2) (10 μl saline/animal, i.c.v.). A) Increments of plasma catecholamines above the basal level. Arrows indicate the administration of 2-AG-E/V-1 and BB/V-2. The actual values for noradrenaline and adrenaline at 0 min were 193 ± 28 and 143 ± 19 pg/ml in the V-1–pretreated group (n = 12), 91 ± 18 and 113 ± 44 pg/ml in the 2-AG-E (0.01 μmol/animal)-pretreated group (n = 5), and 142 ± 41 and 176 ± 43 pg/ml in the 2-AG-E (0.1 μmol/animal)-pretreated group (n = 9), respectively. B) AUC of the elevation of plasma catecholamines above the basal level for each group is expressed as pg/2 h. *P < 0.05, when compared with the Bonferroni method to the V-1– and BB-treated group. Other conditions are the same as those of Figs. 1 – 4.

Effects of bombesin on the spinally projecting PVN part (Fig. 7C) of the PVN in both vehicle (10 μl saline/ neurons exhibiting diacylglycerol lipase a immuno reactivity animal, i.c.v.)- and bombesin (1 nmol/animal, i.c.v.)-treated In the rats microinjected with a retrograde tracer animals. In the dorsal cap and ventral part of the PVN, (Fluoro-Gold) into the spinal cord, Fluoro-Gold–labeled immunoreactivity of Fos was observed in parts of the neurons exhibited gold fluorescent granules in the diacylglycerol lipase α–immunoreactive cells after cytoplasm with a heterogeneous distribution in the treatment with bombesin, but was almost absent after PVN. These labeled neurons were abundantly observed treatment with the vehicle (Fig. 7: B and C). Treatment in the dorsal cap and ventral part of the PVN (Fig. 7A). with bombesin obviously increased the number of However, Fluoro-Gold–labeled neurons were not de- triple-labeled neurons with Fluoro-Gold, diacylglycerol tected in other subnuclei such as the medial and lateral lipase α, and Fos in the dorsal cap (Fig. 7B) and ventral parts of the PVN. part (Fig. 7C) of the PVN. Diacylglycerol lipase α-immunoreactive cells were clearly observed in the dorsal cap (Fig. 7B) and ventral 166 T Shimizu et al

Fig. 6. Effect of rimonabant on the 2-AG ether–induced inhibition of the centrally administered bombesin-induced elevation of

plasma catecholamines. Rimonabant (Rimo) (an antagonist of cannabinoid CB1 receptors) (30 or 90 nmol/animal) and/or 2-AG ether (2-AG-E) (a stable analog of 2-AG) (0.1 μmol/animal), dissolved in a volume of 3 μl DMF/animal, were i.c.v. administered 30 min before the administration of bombesin (BB) (1 nmol/animal, i.c.v.) or vehicle-2 (V-2) (10 μl saline/animal, i.c.v.). A) Increments of plasma catecholamines above the basal level. Arrows indicate the administration of (Rimo and/or 2-AG-E)/vehicle-1 (V-1) and BB/V-2. The V-1–treated groups are the same as those in Fig. 5. The actual values for noradrenaline and adrenaline at 0 min were 140 ± 65 and 204 ± 70 pg/ml in the 2-AG-E–pretreated group (n = 5), 327 ± 37 and 210 ± 22 pg/ml in the [2-AG-E and Rimo (30 nmol)]-pretreated group (n = 5), and 219 ± 36 and 110 ± 25 pg/ml in the [2-AG-E and Rimo (90 nmol)]-pretreated group (n = 10), respectively. B) AUC of the elevation of plasma catecholamines above the basal level for each group. *P < 0.05, #P < 0.05, when compared with the Bonferroni method to the V-1– and BB-treated group or the 2-AG-E– and BB-treated group, respectively. Other conditions are the same as those of Figs. 1 – 5.

Discussion results suggest that brain arachidonic acid produced from 2-AG (as a precursor of prostanoids) is involved in 2-AG generated by the PLC- and diacylglycerol the bombesin-induced response. In the first experiment, lipase–mediated pathway can be hydrolyzed by mono- therefore, we examined the effect of exogenous 2-AG acylglycerol lipase to free arachidonic acid. Actually, on the basal plasma levels of catecholamines. Centrally blockage of monoacylglycerol lipase activity by inhibi- administered 2-AG effectively elevated plasma levels of tors or gene knockout resulted in increased 2-AG and catecholamines, while peripherally administered 2-AG decreased arachidonic acid levels in the mouse brain (38, had no effect. Furthermore, the 2-AG–induced elevation 39). Previously, we reported that central pretreatment was abolished by central pretreatment with JZL184, a with monoacylglycerol lipase inhibitor effectively potent and selective inhibitor of monoacylglycerol lipase reduced the centrally administered bombesin–induced (38). JZL184 actually elevated brain 2-AG level in the elevation of plasma catecholamines in rats (16). These mouse brain (38, 39). These results suggest that the Brain 2-AG in Adrenomedullary Outflow 167

Fig. 7. Effects of bombesin on the spinally projecting neurons of the hypothalamic paraventricular nucleus (PVN) with the immunoreactivity of diacylglycerol lipase α. A) The left panel shows a schematic illus- tration of the PVN (−1.80 mm anterior from the bregma) based on the rat brain atlas of Paxinos and Watson (32). Enlargement of the box in the schema is shown in the right panel as a representa- tive photograph of Fluoro-Gold (FG)- labeled neurons in the PVN. The FG- labeled neurons were abundantly distri- buted in the PVN subdivisions, dorsal cap (PaDC) and ventral part (PaV). Scale bar = 100 μm. B and C) Photographs show FG (left)-, diacylglycerol lipase α (DGLα) (middle)-, and Fos (right)-positive cells in PaDC (B) and PaV (C) of the PVN in the vehicle (10 μl saline/animal, i.c.v.) (up- per)- and bombesin (BB) (1 nmol/animal, i.c.v.) (lower)-treated rats. Arrows indicate triple-labeled neurons. Scale bar = 50 μm.

brain arachidonic acid generated from 2-AG by mono- cholamines using indomethacin. Centrally administered acylglycerol lipase evokes the elevation of plasma indomethacin effectively reduced the 2-AG–induced catecholamines in rats. response, suggesting that cyclooxygenase-mediated Arachidonic acid produced from 2-AG can be further production of brain prostanoids are involved in the metabolized by cyclooxygenase to active metabolites, 2-AG–induced elevation of plasma catecholamines in the prostanoids (39). Previously, we reported that central rats. Therefore, in the present experiment, centrally pretreatment with indomethacin (an inhibitor of cyclo- administered 2-AG may be hydrolyzed by mono- oxygenase) reduced the centrally administered arachi- acylglycerol lipase to free arachidonic acid, which is donic acid–induced elevation of plasma catecholamines further metabolized to prostanoids by cyclooxygenase in in rats (6). In the next experiment, therefore, we examined the brain, thereby elevating plasma catecholamines in whether brain cyclooxygenase is involved in the centrally rats. Recently, another mechanism has been shown to administered 2-AG–induced elevation of plasma cate- generate prostanoids in addition to the mechanism 168 T Shimizu et al described above. 2-AG can first be metabolized by the bombesin-induced elevation of plasma catechol- cyclooxygenase and synthases to generate amines in rats. 2-AG and 2-AG ether respectively exhibit glycerol esters of prostanoids (40, 41), which are then Ki values of 472 and 21.2 nM at cannabinoid CB1 recep- hydrolyzed by monoacylglycerol lipase to free pros- tors (42, 47), indicating that 2-AG ether has a higher tanoids. Further studies are necessary to clarify the affinity for the receptors than 2-AG. 2-AG acts as a full mechanisms responsible for generating prostanoids agonist at cannabinoid CB1 receptors (21, 48), while from 2-AG in the rat brain. 2-AG ether is reported as a partial agonist at the receptors In addition to a source of arachidonic acid described expressed in Xenopus oocytes (49) and a full agonist at above, 2-AG has been also recognized as a major endo- the receptors in human neocortex (50). Taken together genous ligand (endocannabinoid) for cannabinoid CB with the present results, exogenously administered 2-AG receptors (20, 21). In order to examine the role of ether into the brain seems to act as a full agonist at brain brain 2-AG as an endocannabinoid, we used 2-AG ether cannabinoid CB1 receptors, thereby further inhibiting the (noladin ether), which functions as a ligand for cannabi- bombesin-induced elevation of plasma catecholamines in noid CB receptors (42) and has been recognized as a addition to the inhibition mediated by endogenous 2-AG stable 2-AG analog for monoacylglycerol lipase with an generated by bombesin. endogenous half-life of hours rather than minutes (43). 2-AG is produced from diacylglycerol containing These findings suggest that 2-AG ether seems to be a arachidonic acid in position 2 by sn-1 selective diacyl- cannabinoid CB receptor ligand lacking a function as a glycerol lipase (13). Two isoforms of this lipase, α and β source of arachidonic acid. Actually, in the present isoforms, have been identified and both are expressed in study, centrally administered 2-AG ether at the same the brain (13). In diacylglycerol lipase α–knockout mice, doses of 2-AG had no effect on the basal plasma levels endocannabinoid-mediated retrograde synaptic signaling of catecholamines. Therefore, we used 2-AG ether as a is totally absent in the cerebellum, hippocampus, striatum, “pure” agonist for cannabinoid CB receptors in the next and prefrontal cortex, while the signaling is intact in experiments. Two major cannabinoid receptors, CB1 and diacylglycerol lipase β–knockout mouse brain (51 – 53). CB2, have been identified (44, 45). In the central nervous Additionally, 2-AG content is markedly reduced in these system, cannabinoid CB1 receptors are mainly present, brain regions of diacylglycerol lipase α–knockout mice, while cannabinoid CB2 receptors are mainly distributed but not in diacylglycerol lipase β–knockout mice (52, in peripheral and immune cells (44, 45). 2-AG ether has 53). These findings suggest that brain diacylglycerol been shown to be a selective cannabinoid CB1–receptor lipase α is mainly responsible for the endocannabinoid agonist, exhibiting Ki values of 21.2 and > 3000 nM at 2-AG–mediated retrograde signaling. Interestingly, in cannabinoid CB1 and CB2 receptors, respectively (42). the diacylglycerol lipase α–knockout mouse brain, levels Recently, we reported that centrally administered of arachidonic acid decrease in parallel with those of bombesin–induced elevation of plasma catecholamines 2-AG (51). Taken together, brain diacylglycerol lipase α was potentiated by central pretreatment with a cannabi- seems to be involved in production of both 2-AG and noid CB1–receptor antagonist and was reduced by central arachidonic acid. In the final experiment, therefore, we pretreatment with an endocannabinoid uptake-inhibitor examined the distribution of diacylglycerol lipase α in (16). These results suggest that brain endocannabinoid the hypothalamic PVN, which has been considered as a (probably 2-AG) acts on the bombesin-activated control center of the central sympatho-adrenomedullary mechanisms, thereby inhibiting the bombesin-induced outflow (54). response through cannabinoid CB1 receptor–mediated Presympathetic neurons in the PVN send mono- mechanisms in the brain, in addition to its stimulatory and poly-synaptic projections to the sympathetic pre- role as a precursor of brain prostanoids. In the present ganglionic neurons residing in the IML of the spinal cord experiment, therefore, we examined the complicated (55). The released acetylcholine from the sympathetic inhibitory effect of 2-AG using exogenously admin- pre ganglionic neurons located in T7-9 (56) activates the istered 2-AG ether on the bombesin-induced response adrenal chromaffin cells, thereby evoking the secretion with regard to the brain cannabinoid CB1 receptors. of catecholamines from the adrenal medulla. Therefore, Centrally administered 2-AG ether effectively reduced we microinjected Fluoro-Gold, a mono-synaptic retro- the bombesin-induced elevation of plasma catechol- grade tracer, into the rat spinal cord at the T8 level for amines, and this inhibition was attenuated by central labeling presympathetic PVN neurons innervating the treatment with rimonabant (SR141716) [a selective adrenal medulla and then examined the distribution of antagonist of cannabinoid CB1 receptors (46)]. These diacylglycerol lipase α on the Fluoro-Gold-labeled results clearly suggest that endogenously generated brain neurons using immunohistochemical procedures. Fluoro- 2-AG plays an inhibitory role as an endocannabinoid in Gold–labeled neurons were localized in the dorsal cap Brain 2-AG in Adrenomedullary Outflow 169 and ventral part of the PVN, in accordance with previous 2-AG, 1) drugs that inhibit brain prostanoids production studies showing that these regions contain neurons such as cyclooxygenase inhibitors and 2) drugs that projecting to the spinal cord (57, 58). In these Fluoro- activate brain cannabinoid CB1–receptor signaling such Gold–labeled spinally projecting PVN neurons, diacyl- as agonists for the receptors can be considered to be glycerol lipase α was apparently expressed, and centrally valuable for treating diseases evoked by excess activa- administered bombesin actually activated these spinally tion of central sympatho-adrenomedullary outflow. projecting PVN neurons expressing diacylglycerol lipase α Actually, in spontaneous hypertensive rats, decreased in rats. These results suggest a possibility that centrally cannabinoid CB1–receptor signaling in the nucleus of administered bombesin activates, at least in part, the the solitary tract seems to attenuate baroreflex-evoked spinally projecting PVN neurons expressing diacyl- central sympathetic modulation, thereby inducing the glycerol lipase α, thereby generating 2-AG to modulate hypertensive phenotype (69). These results suggest that

(stimulate and inhibit) the bombesin-induced elevation drugs activating the CB1-receptor signaling may be of plasma catecholamines in rats. Further studies are beneficial for clinical application to hypertensive required to identify the diacylglycerol lipase α in the disorders. Further studies are needed to clarify the roles pre-sympathetic, poly-synaptic neurons in the central of endogenous 2-AG for treating the central sympatho- nervous system. adrenomedullary outflow–induced disorders. Previously, we reported the involvement of the brain In summary, we demonstrated here that 2-AG in the PLC-, diacylglycerol lipase-, monoacylglycerol lipase-, brain plays bidirectional roles, a stimulatory role as an and cyclooxygenase-mediated pathway in the elevation arachidonic acid precursor on the basal state and an of plasma catecholamines evoked by centrally admin- inhibitory role as an endocannabinoid in the bombesin- istered arginine–vasopressin and corticotropin-releasing induced activation of central adrenomedullary outflow in factor (other stress-related peptides in the brain) (30, rats. 59 – 63). These results suggest that endogenously generated brain 2-AG by the PLC- and diacylglycerol Acknowledgments lipase–mediated pathway plays a stimulatory role as a precursor of arachidonic acid and prostanoids through This work was supported in part by Grants-in-Aid for Young the monoacylglycerol lipase– and cyclooxygenase– Scientists (B) (No. 21790627 and 23790744 to T.S. and No. 23790745 to K.T.) and a Grant-in-Aid for Scientific Research (C) (No. 20590702 mediated pathway in the peptide-induced responses. In to K.Y.) from the Japan Society for the Promotion of Science; a grant addition, these peptide-induced elevations of plasma from The Smoking Research Foundation in Japan; a Discretionary catecholamines were potentiated by central pretreatment Grant of the President of the Kochi University and a Discretionary with a cannabinoid CB1–receptor antagonist (61, 62), Grant of the Head of the Kochi Medical School Hospital. suggesting that endogenously generated brain endo- cannabinoid (probably 2-AG) plays an inhibitory role References in these responses. Further studies are needed to clarify 1 Anastasi A, Erspamer V, Bucci M. Isolation and structure of the bidirectional roles of endogenous 2-AG as a pre- bombesin and alytesin, 2 analogous active peptides from the skin cursor of prostanoids and as an endocannabinoid in of the European amphibians Bombina and Alytes. Experientia. these peptide-induced elevations of plasma catechol- 1971;27:166–167. amines in rats. 2 Jensen RT, Battey JF, Spindel ER, Benya RV. International Union Plasma catecholamines are elevated by activation of of Pharmacology. LXVIII. Mammalian bombesin receptors: the sympatho-adrenomedullary system, which plays an nomenclature, distribution, pharmacology, signaling, and func- important role in physiological and pathophysiological tions in normal and disease states. Pharmacol Rev. 2008;60:1– 42. responses to stressors (64). Because this system modu- 3 Okuma Y, Yokotani K, Osumi Y. Centrally applied bombesin lates cardiovascular and immune functions (65, 66), increases nerve activity of both sympathetic and adrenal branch excess activation of the sympatho-adrenomedullary of the splanchnic nerves. Jpn J Pharmacol. 1995;68:227–230. system induced by high levels of stress may thus con- 4 Yokotani K, Okada S, Nakamura K, Yamaguchi-Shima N, tribute to the development of cardiovascular diseases Shimizu T, Arai J, et al. Brain prostanoid TP receptor-mediated such as hypertension and certain diseases of the immune adrenal noradrenaline secretion and EP3 receptor-mediated system (67, 68). Therefore, in order to clarify the mecha- sympathetic noradrenaline release in rats. Eur J Pharmacol. 2005; nisms of the stress-related diseases, we have been 512:29–35. 5 Lu L, Shimizu T, Nakamura K, Yokotani K. Brain neuronal/ examining the central regulatory mechanisms of sympa- inducible nitric oxide synthases and cyclooxygenase-1 are tho-adrenomedullary outflow using several stress-related involved in the bombesin-induced activation of central adreno- neuropeptides including bombesin. From the present medullary outflow in rats. Eur J Pharmacol. 2008;590:177–184. results indicating the bidirectional roles of endogenous 6 Yokotani K, Wang M, Murakami Y, Okada S, Hirata M. Brain 170 T Shimizu et al

-arachidonic acid cascade is involved in the 45:405–446. activation of central sympatho-adrenomedullary outflow in rats. 22 Katona I, Freund TF. Multiple functions of endocannabinoid Eur J Pharmacol. 2000;398:341–347. signaling in the brain. Annu Rev Neurosci. 2012;35:529–558. 7 Murakami Y, Okada S, Nishihara M, Yokotani K. Roles of 23 Yates ML, Barker EL. Inactivation and biotransformation of

brain prostaglandin E2 and thromboxane A2 in the activation of the endogenous and 2-arachidonoyl- the central sympatho-adrenomedullary outflow in rats. Eur J glycerol. Mol Pharmacol. 2009;76:11–17. Pharmacol. 2002;452:289–294. 24 Crosby KM, Inoue W, Pittman QJ, Bains JS. Endocannabinoids 8 Shimizu T, Yokotani K. Effects of centrally administered prosta- gate state-dependent plasticity of synaptic inhibition in feeding

glandin E3 and thromboxane A3 on plasma noradrenaline and circuits. Neuron. 2011;71:529–541.

adrenaline in rats: comparison with prostaglandin E2 and 25 Olango WM, Roche M, Ford GK, Harhen B, Finn DP. The

thromboxane A2. Eur J Pharmacol. 2009;611:30–34. in the rat dorsolateral periaqueductal

9 Balsinde J, Winstead MV, Dennis EA. Phospholipase A2 regula- grey mediates fear-conditioned analgesia and controls fear tion of arachidonic acid mobilization. FEBS Lett. 2002;531:2–6. expression in the presence of nociceptive tone. Br J Pharmacol.

10 Kudo I, Murakami M. Phospholipase A2 . Prostaglandins 2012;165:2549–2560. Other Lipid Mediat. 2002;68–69:3–58. 26 Mulder J, Zilberter M, Pasquaré SJ, Alpár A, Schulte G, Ferreira 11 Shimizu T, Okada S, Yamaguchi N, Arai J, Wakiguchi H, SG, et al. Molecular reorganization of endocannabinoid signal- Yokotani K. Brain phospholipase C/diacylglycerol lipase are ling in Alzheimer’s disease. Brain. 2011;134:1041–1060.

involved in bombesin BB2 receptor-mediated activation of 27 Shohami E, Cohen-Yeshurun A, Magid L, Algali M, Mechoulam sympatho-adrenomedullary outflow in rats. Eur J Pharmacol. R. Endocannabinoids and traumatic brain injury. Br J Pharmacol. 2005;514:151–158. 2011;163:1402–1410. 12 Rebecchi MJ, Pentyala SN. Structure, function, and control of 28 Higuchi S, Irie K, Mishima S, Araki M, Ohji M, Shirakawa A, phosphoinositide-specific phospholipase C. Physiol Rev. 2000; et al. The cannabinoid 1-receptor silent antagonist O-2050 80:1291–1335. attenuates preference for high-fat diet and activated astrocytes 13 Bisogno T, Howell F, Williams G, Minassi A, Cascio MG, in mice. J Pharmacol Sci. 2010;112:369–372. Ligresti A, et al. Cloning of the first sn1-DAG points to 29 Brozoski DT, Dean C, Hopp FA, Seagard JL. Uptake blockade the spatial and temporal regulation of endocannabinoid signaling of endocannabinoids in the NTS modulates baroreflex-evoked in the brain. J Cell Biol. 2003;163:463–468. sympathoinhibition. Brain Res. 2005;1059:197–202. 14 Blankman JL, Simon GM, Cravatt BF. A comprehensive 30 Shimizu T, Okada S, Yamaguchi-Shima N, Yokotani K. Brain profile of brain enzymes that hydrolyze the endocannabinoid phospholipase C-diacylglycerol lipase pathway is involved in 2-arachidonoylglycerol. Chem Biol. 2007;14:1347–1356. vasopressin-induced release of noradrenaline and adrenaline 15 Dinh TP, Carpenter D, Leslie FM, Freund TF, Katona I, Sensi from adrenal medulla in rats. Eur J Pharmacol. 2004;499:99– SL, et al. Brain lipase participating in endo- 105. cannabinoid inactivation. Proc Natl Acad Sci U S A. 2002;99: 31 Yokotani K, Nishihara M, Murakami Y, Hasegawa T, Okuma Y, 10819–10824. Osumi Y. Elevation of plasma noradrenaline levels in urethane-

16 Shimizu T, Lu L, Yokotani K. Endogenously generated 2- anaesthetized rats by activation of central prostanoid EP3 arachidonoylglycerol plays an inhibitory role in bombesin- receptors. Br J Pharmacol. 1995;115:672–676. induced activation of central adrenomedullary outflow in rats. 32 Paxinos G, Watson C. In: Paxinos G, Watson C, editors. The rat Eur J Pharmacol. 2011;658:123–131. brain in stereotaxic coordinates. Burlington: Elsevier Academic 17 Gauthier KM, Baewer DV, Hittner S, Hillard CJ, Nithipatikom Press; 2005. K, Reddy DS, et al. Endothelium-derived 2-arachidonylglycerol: 33 Anton AH, Sayre DF. A study of the factors affecting the an intermediate in vasodilatory release in bovine aluminum oxide-trihydroxyindole procedure for the analysis of coronary arteries. Am J Physiol Heart Circ Physiol. 2005;288: catecholamines. J Pharmacol Exp Ther. 1962;138:360–375. H1344–H1351. 34 Tanaka K, Shimizu T, Lu L, Nakamura K, Yokotani K. Centrally 18 Tang X, Edwards EM, Holmes BB, Falck JR, Campbell WB. administered bombesin activates COX-containing spinally pro- Role of phospholipase C and diacylglyceride lipase pathway in jecting neurons of the PVN in anesthetized rats. Auton Neurosci. arachidonic acid release and acetylcholine-induced vascular 2012;169:63–69. relaxation in rabbit aorta. Am J Physiol Heart Circ Physiol. 35 Tanaka K, Shimizu T, Lu L, Yokotani K. Possible involvement of 2006;290:H37–H45. S-nitrosylation of brain cyclooxygenase-1 in bombesin-induced 19 Balogh G, Péter M, Liebisch G, Horváth I, Török Z, Nagy E, central activation of adrenomedullary outflow in rats. Eur J et al. Lipidomics reveals membrane lipid remodelling and Pharmacol. 2012;679:40–50. release of potential lipid mediators during early stress responses 36 Tanaka K, Osako Y, Yuri K. Juvenile social experience regulates in a murine melanoma cell line. Biochim Biophys Acta. 2010; central neuropeptides relevant to emotional and social behaviors. 1801:1036–1047. Neuroscience. 2010;166:1036–1042. 20 Sugiura T, Kondo S, Sukagawa A, Nakane S, Shinoda A, Itoh K, 37 Okuma Y, Yokotani K, Osumi Y. Brain prostaglandins mediate et al. 2-Arachidonoylglycerol: a possible endogenous cannabi- the bombesin-induced increase in plasma levels of catechol- noid receptor ligand in brain. Biochem Biophys Res Commun. amines. Life Sci. 1996;59:1217–1225. 1995;215:89–97. 38 Long JZ, Li W, Booker L, Burston JJ, Kinsey SG, Schlosburg JE, 21 Sugiura T, Kishimoto S, Oka S, Gokoh M. Biochemistry, et al. Selective blockade of 2-arachidonoylglycerol hydrolysis pharmacology and physiology of 2-arachidonoylglycerol, an produces cannabinoid behavioral effects. Nat Chem Biol. 2009; endogenous ligand. Prog Lipid Res. 2006; 5:37–44. Brain 2-AG in Adrenomedullary Outflow 171

39 Nomura DK, Morrison BE, Blankman JL, Long JZ, Kinsey SG, prefrontal cortex. J Physiol. 2011;589:4857–4884. Marcondes MC, et al. Endocannabinoid hydrolysis generates 54 Jansen AS, Nguyen XV, Karpitskiy V, Mettenleiter TC, Loewy brain prostaglandins that promote neuroinflammation. Science. AD. Central command neurons of the sympathetic nervous 2011;334:809–813. system: basis of the fight-or-flight response. Science. 1995;270: 40 Kozak KR, Rowlinson SW, Marnett LJ. Oxygenation of the 644–646. endocannabinoid, 2-arachidonylglycerol, to glyceryl prosta- 55 Pyner S. Neurochemistry of the paraventricular nucleus of the glandins by cyclooxygenase-2. J Biol Chem. 2000;275:33744– hypothalamus: Implications for cardiovascular regulation. J 33749. Chem Neuroanat. 2009;38:197–208. 41 Rouzer CA, Marnett LJ. Endocannabinoid oxygenation by cyclo- 56 Pyner S, Coote JH. Evidence that sympathetic preganglionic oxygenases, lipoxygenases, and cytochromes P450: cross-talk neurones are arranged in target-specific columns in the thoracic between the eicosanoid and endocannabinoid signaling path- spinal cord of the rat. J Comp Neurol. 1994;342:15–22. ways. Chem Rev. 2011;111:5899–5921. 57 Sawchenko PE, Swanson LW. Immunohistochemical identifica- 42 Hanus L, Abu-Lafi S, Fride E, Breuer A, Vogel Z, Shalev DE, tion of neurons in the paraventricular nucleus of the hypothala- et al. 2-arachidonyl glyceryl ether, an endogenous agonist of the mus that project to the medulla or to the spinal cord in the rat. J

cannabinoid CB1 receptor. Proc Natl Acad Sci U S A. 2001;98: Comp Neurol. 1982;205:260–272. 3662–3665. 58 Swanson LW, Kuypers HG. The paraventricular nucleus of the 43 Laine K, Järvinen K, Mechoulam R, Breuer A, Järvinen T. hypothalamus: cytoarchitectonic subdivisions and organization Comparison of the enzymatic stability and intraocular pressure of projections to the pituitary, dorsal vagal complex, and spinal effects of 2-arachidonylglycerol and noladin ether, a novel cord as demonstrated by retrograde fluorescence double-labeling putative endocannabinoid. Invest Ophthalmol Vis Sci. 2002;43: methods. J Comp Neurol. 1980;194:555–570.

3216–3222. 59 Okada S, Murakami Y, Nakamura K, Yokotani K. Vasopressin V1 44 Howlett AC, Barth F, Bonner TI, Cabral G, Casellas P, Devane receptor-mediated activation of central sympatho-adrenomedul- WA, et al. International Union of Pharmacology. XXVII. lary outflow in rats. Eur J Pharmacol. 2002;457:29–35. Classification of cannabinoid receptors. Pharmacol Rev. 2002;54: 60 Okada S, Shimizu T, Yokotani K. Brain phospholipase C and 161–202. diacylglycerol lipase are involved in corticotropin-releasing 45 Mackie K. Cannabinoid receptors: where they are and what they hormone-induced sympatho-adrenomedullary outflow in rats. do. J Neuroendocrinol. 2008;20 Suppl 1:10–14. Eur J Pharmacol. 2003;475:49–54. 46 Rinaldi-Carmona M, Barth F, Héaulme M, Shire D, Calandra B, 61 Shimizu T, Yokotani K. Bidirectional roles of the brain 2-arachi- Congy C, et al. SR141716A, a potent and selective antagonist of donoyl-sn-glycerol in the centrally administered vasopressin- the brain cannabinoid receptor. FEBS Lett. 1994;350:240–244. induced adrenomedullary outflow in rats. Eur J Pharmacol. 47 Mechoulam R, Ben-Shabat S, Hanus L, Ligumsky M, Kaminski 2008;582:62–69. NE, Schatz AR, et al. Identification of an endogenous 62 Shimizu T, Lu L, Yokotani K. Possible inhibitory roles of endo- 2-monoglyceride, present in canine gut, that binds to cannabinoid genous 2-arachidonoylglycerol during corticotropin-releasing receptors. Biochem Pharmacol. 1995;50:83–90. factor-induced activation of central sympatho-adrenomedullary 48 Savinainen JR, Järvinen T, Laine K, Laitinen JT. Despite outflow in anesthetized rats. Eur J Pharmacol. 2010;641:54–60. substantial degradation, 2-arachidonoylglycerol is a potent full 63 Yokotani K, Murakami Y, Okada S, Hirata M. Role of brain

efficacy agonist mediating CB1 receptor-dependent G-protein arachidonic acid cascade on central CRF1 receptor-mediated activation in rat cerebellar membranes. Br J Pharmacol. activation of sympatho-adrenomedullary outflow in rats. Eur J 2001;134:664–672. Pharmacol. 2001;419:183–189. 49 Luk T, Jin W, Zvonok A, Lu D, Lin XZ, Chavkin C, et al. 64 Chrousos GP. Stressors, stress, and neuroendocrine integration of Identification of a potent and highly efficacious, yet slowly the adaptive response. The 1997 Hans Selye Memorial Lecture. desensitizing CB1 cannabinoid receptor agonist. Br J Pharmacol. Ann N Y Acad Sci. 1998;851:311–335. 2004;142:495–500. 65 Barron BA, Van Loon GR. Role of sympathoadrenomedullary 50 Steffens M, Zentner J, Honegger J, Feuerstein TJ. Binding system in cardiovascular response to stress in rats. J Auton Nerv affinity and agonist activity of putative endogenous cannabinoids Syst. 1989;28:179–187.

at the human neocortical CB1 receptor. Biochem Pharmacol. 66 Kuroki K, Takahashi HK, Iwagaki H, Murakami T, Kuinose M, 2005;69:169–178. Hamanaka S, et al. beta2-adrenergic receptor stimulation-induced 51 Gao Y, Vasilyev DV, Goncalves MB, Howell FV, Hobbs C, immunosuppressive effects possibly through down-regulation of Reisenberg M, et al. Loss of retrograde endocannabinoid signal- co-stimulatory molecules, ICAM-1, CD40 and CD14 on mono- ing and reduced adult neurogenesis in diacylglycerol lipase cytes. J Int Med Res. 2004;32:465–483. knock-out mice. J Neurosci. 2010;30:2017–2024. 67 McDougall SJ, Widdop RE, Lawrence AJ. Central autonomic 52 Tanimura A, Yamazaki M, Hashimotodani Y, Uchigashima M, integration of psychological stressors: focus on cardiovascular Kawata S, Abe M, et al. The endocannabinoid 2-arachidonoyl- modulation. Auton Neurosci. 2005;123:1–11. glycerol produced by diacylglycerol lipase alpha mediates 68 Vanitallie TB. Stress: a risk factor for serious illness. Metabo- retrograde suppression of synaptic transmission. Neuron. 2010; lism. 2002;51:40–45. 65:320–327. 69 Brozoski DT, Dean C, Hopp FA, Hillard CJ, Seagard JL. 53 Yoshino H, Miyamae T, Hansen G, Zambrowicz B, Flynn M, Differential endocannabinoid regulation of baroreflex-evoked Pedicord D, et al. Postsynaptic diacylglycerol lipase mediates sympathoinhibition in normotensive versus hypertensive rats. retrograde endocannabinoid suppression of inhibition in mouse Auton Neurosci. 2009;150:82–93.