Cannabinoids exert complex actions on neurotransmitter systems involved in cognition, locomotion, appetite, but no information was available so far on the interactions between the endocannabinoid system and histaminergic neurons that command several, similar behavioural states and memory. In this study, we investigated the effect of cannabimimetic compounds on histamine release using the microdialysis technique in the brain of freely moving rats. We found that systemic administration of the cannabinoid receptors 1 (CB1-r) agonist arachidonyl-2′chloroethylamide/N-(2chloroethyl)-5Z,8Z,11Z,14Z-eicosatetraenamide (ACEA; 3 mg/kg) increased histamine release from the posterior hypothalamus, where the histaminergic tuberomamillary nuclei (TMN) are located. Local infusions of ACEA (150 nM) or R(+)-methanandamide (mAEA; 1μM), another CB1-r agonist, in the TMN augmented histamine release from the TMN, as well as from two histaminergic projection areas, the nucleus basalis magnocellularis and the dorsal striatum. When the endocannabinoid uptake inhibitor AM404 was infused into the TMN, however, increased histamine release was observed only in the TMN. The cannabinoid-induced effects on histamine release were blocked by co-administrations with the CB1-r antagonist AM251. Using double-immunofluorescence labelling and confocal laser-scanning microscopy, CB1-r immunostaining was found in the hypothalamus, but was not localized onto histaminergic cells. The modulatory effect of cannabimimetic compounds on histamine release apparently did not involve inhibition of γ-aminobutyric acid (GABA)ergic neurotransmission, which provides the main inhibitory input to the histaminergic neurons in the hypothalamus, as local infusions of ACEA did not modify GABA release from the TMN. These profound effects of cannabinoids on histaminergic neurotransmission may partially underlie some of the behavioural changes observed following exposure to cannabinoid-based drugs.

CB1-r activation modulates neuronal activity by hyperpolarizing the cell and/or inhibiting the release of neurotransmitters such as γ-aminobutyric acid (GABA). As histaminergic neurons receive prominent innervations from sleep active GABAergic neurons of the ventrolateral preoptic nucleus ( Sherin et al., 1998 ), we also measured the effect of cannabinoids on GABA release from the hypothalamus. Finally, we used immunohistochemical techniques to establish whether there is a similar distribution of CB1-r and histaminergic fibres in the hypothalamus.

We therefore explored the possibility that the cannabinoid system exerts some of its behavioural effects by modulating, or acting in concert with, the activity of histaminergic neurons. Using the microdialysis technique in freely moving rats, selective, directly acting cannabinoid CB1-r agonists or an indirectly acting inhibitor of endocannabinoid uptake were given locally into the TMN, and their effect on histamine release was monitored in the TMN itself, in the NBM and striatum. We show that cannabinoids, at the low concentrations used in this study, have an excitatory effect on the histaminergic system.

The endocannabinoid system plays an important neuromodulatory role in brain physiology, fine-tuning information flow in neuronal networks associated with cognitive processing, emotion, pain perception and appetite ( Di Marzo & Matias, 2005 ). Endocannabinoids are generally made on demand, bind with high affinity to cannabinoid receptors 1 (CB1-r), and are rapidly eliminated by a carrier-mediated transport followed by intracellular enzymatic metabolism ( Piomelli, 2003 ). For example, endocannabinoids release in the hypothalamus regulates appetitive behaviour ( Kirkham et al., 2002 ), whereas in the basal ganglia it counteracts the stimulation of movement induced by dopamine agonists ( Beltramo et al., 2000 ). Furthermore, recent studies indicated that endocannabinoids content rises in restricted brain areas engaged in the processing of emotional information ( Marsicano et al., 2002 ). CB1-r mediate not only the physiological effects of endocannabinoids, but also the psychotropic effects of Δ 9 -tetrahydrocannabinol (Δ 9 -THC). Exogenous cannabinoids induce a broad array of behavioural responses that include catalepsy, reduced movement and hypothermia ( Pertwee, 1997 ). Furthermore, behaviours ranging from relaxation and sedation to anxiety and panic attacks have been observed ( Zuardi et al., 1982 ).

All values are expressed as means ± SEM, and the number of rats used in each experiment is also indicated. The presence of significant treatment effects was first determined by a one-way ANOVA followed by Fisher’s PSLD test or with two-way ANOVA. For all statistical tests, P < 0.05 was considered significant. For clarity purposes we reported in figures and figure legends only the significant differences vs. the last sample before drug treatment. However, differences were significant vs. all baseline samples. Statistical analysis was performed using StatView (Abacus Concepts, Berkley, CA, USA).

Young adult male Sprague–Dawley rats were deeply anaesthetized with chloral hydrate and perfused transcardially with 50 mL physiological saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PB), pH 7.4. Brains were then postfixed in 4% paraformaldehyde in PB for 2 h at 4 °C. Brains were cryoprotected in 30% sucrose in PB, and 40-μm-thick sections were cut on a cryostat microtome and collected in PB. All immunostaining procedures were performed on free-floating sections. For double-labelling of CB1-r and histaminergic cells, sections were preincubated with 5% normal donkey serum (NDS, Jackson Immunoresearch, West Grove, PA, USA), 2.5% bovine serum albumin (BSA, Sigma) and 0.5% Triton X-100 in PB for 1 h at room temperature. Sections were then incubated in a cocktail of goat anti-CB1-r antibodies (1 : 1000, directed against the C-terminal; Hájos et al., 2000 ) and rabbit antihistidine decarboxylase (HDC, 1 : 1000; Acris, Bad Nauheim, Germany) primary antibodies in PB containing 0.5% Triton X-100, 0.1% BSA and 1% NDS for 48 h at 4 °C. After thorough rinsing in PB, sections were incubated in Cy3-conjugated donkey anti-goat IgG (1 : 200; Jackson Immunoresearch) for 2 h at room temperature. After extensive rinses in PB, sections were incubated in Alexa Fluor 488-labelled, donkey anti-rabbit IgG (Molecular Probes, Eugene, OR, USA) diluted 1 : 200 in PB containing 2% BSA for 1 h at room temperature. Sections were then mounted on glass slides and coverslipped with 70% glycerol in PB. Preadsorption of the CB1-r antibodies with glutathione-S-transferase-conjugated fusion protein (1 μg/mL), or elimination of the primary antibodies resulted in no immunostaining. Observations were performed with a Bio-Rad MCR 1024 ES confocal laser-scanning microscope (Bio-Rad, Hercules, CA, USA) equipped with a Krypton/Argon laser source 15 mW for fluorescence measurements. Series of optical sections (512 × 512 pixels) were taken through the depth of the specimens with a thickness of 1 μm at the intervals of 0.8 μm by using a Nikon Planapo × 60 (× 40) 1.4 oil immersion objective. Twenty optical sections for each sample were examined and projected as a single composite image by superimposition. To avoid bleed-through, dual channel scanning of signals from Cy3 and Alexa Fluor 488 were recorded separately and saved in two different files.

GABA analysis was carried out by HPLC with fluorimetric detection after OPA derivatization as previously described ( Bianchi et al., 1999 ). The OPA-derivatives were separated on a 5 μm reverse-phase Nucleosil C18 column (200 × 4.6 mm i.d., Machery-Nagel, Duren, Germany) at room temperature, using as mobile phase methanol and potassium acetate (0.1 M, pH adjusted to 5.52 with glacial acetic acid), eluting at 1 mL/min flow rate in a three-step linear gradient, from 25% to 90% methanol.

Histamine contents in the dialysates were determined by HPLC-fluorimetry using a modified version of the protocol by Yamatodani et al. (1985) . Briefly, the column (Hypersil ODS, 3 μm, 2.1 × 100 mm; Thermo Electron Corporation, Bellefonte, PA, USA) was eluted with 0.25 M potassium dihydrogen phosphate containing 5% octanesulphonic acid (Sigma) at a flow rate of 0.4 mL/min. The eluate from the column was mixed first with 0.1% o-phthalaldehyde (OPA) solution at a flow rate of 0.1 mL/min and then to a solution containing 4 M sodium hydroxide and 0.2 M boric acid (flow rate 0.137 mL/min) to adjust the reaction mixture to pH 12.5. The reaction took place at 45 °C. Then 17% orthophosphoric acid was added to the solution (flow rate 0.137 mL/min) to reach a final reaction mixture at pH 3. The fluorescent intensity was measured with a spectrofluorometer (Agilent series 1100, Waldbronn, Germany) at 450 nm with excitation at 360 nm.

The placement of microdialysis membranes was verified post mortem. Rats were overdosed with chloral hydrate, the brains removed and stored in 10% formalin for 10 days. Forty-micrometre sections were then sliced on a cryostat, mounted on gelatin-coated slides and then stained with Cresyl violet for light microscopic observation. Data from rats in which the membranes were not correctly positioned were discarded. Typical probe placements are shown in .

The microdialysis experiments were performed 24 h after surgery, during which rats, housed one per cage, recovered from surgery. The stylet was removed from the guide cannulae and the microdialysis probes (molecular weight cut-off, 6000 Da; Metalant) were inserted; the dialysing membrane protruded 2 mm from the tip of the cannula. Both probes were perfused with Ringer’s solution (in mM: NaCl, 147; CaCl 2 , 1.2; KCl, 4.0; pH 7.0) at a flow rate of 2 μL/min using a microperfusion pump (Carnegie Medicine, Sweden; Mod CMA/100). Histamine release stabilized 2 h after insertion of the microdialysis probes, and fractions were collected at 15-min intervals. Spontaneous release was defined as the average value of the four 15-min fractions collected during 1 h of perfusion with Ringer’s solution prior to drug treatment. All subsequent fractions were expressed as a percentage of this value. To prevent degradation of histamine, 1.5 μL of 5 mM HCl was added to each sample. The dialysates were kept at −80 °C until analysis. Drugs were supplied in 100% EtOH that was diluted 1 : 90 000 in the final solution. For arachidonyl-2′chloroethylamide/N-(2chloroethyl)-5Z,8Z,11Z,14Z-eicosatetraenamide (ACEA) i.p. administration, EtOH concentration was 60%. Control injections with saline contained 60% EtOH.

Rats, anaesthetized with chloral hydrate (400 mg/kg i.p.) and positioned in a stereotaxic frame (Stellar, Stoelting, Wood Dale, IL, USA), were implanted with one or two guide cannulae (Metalant, Sweden) according to the following coordinates from bregma ( Paxinos & Watson, 1998 ): TMN, AP = −4.3, L = −1.1, DV = −7.2; NBM, AP = −0.8; L = −2.8; DV = −6.5; dorsal striatum, AP = 0, L = −4, DV = −4. A surgical screw served as an anchor and the cannulae were fixed to the skull with acrylic dental cement.

Male, 8–9-week-old Sprague–Dawley rats (250–280 g/wt, Harlan, Italy) were housed in groups of three in a temperature-controlled room (20–24 °C), on a 12 h light : dark cycle, and were allowed free access to food and water. All the experiments were done in strict compliance with the EEC recommendations for the care and use of laboratory animals (86/609/CEE), and were approved by the Animal Care Committee of the ‘Dipartimento di Farmacologia Preclinica e Clinica’ of the ‘Universitá di Firenze’.

We investigated whether the increased histamine release elicited by CB1-r agonists involved modulation of intrahypothalamic GABA release. GABA content measured in dialysates collected from the TMN perfused with 1 μM mAEA for 60 min did not significantly change (spontaneous release, 0.80 ± 0.14 pmol/15 min; n = 7; data not shown). As such, it is unlikely that the excitatory effect of CB1-r agonists on histaminergic neurons is mediated by the reduction of an inhibitory GABAergic tone. However, the activity of histaminergic cells is controlled by a tonic release of GABA, as perfusion of the TMN with the GABA A receptor antagonist bicuculline (10 μM) for 75 min significantly increased histamine spontaneous release from the TMN up to 91 ± 16% ( , lower panel) but, surprisingly, not from the NBM ( , upper panel; n = 4). Co-administration in the TMN of bicuculline and ACEA (150 nM) increased significantly histamine release from the NBM (196 ± 87.7% of baseline; , upper panel), and induced an additional increase of TMN histamine release over that induced by 10 μM bicuculline alone ( , lower panel; n = 5). We compared the percentage of histamine release calculated by averaging histamine content in all samples collected during pharmacological stimulations, and found that ACEA increased histamine release from the TMN by 82 ± 14%, bicuculline by 65 ± 9% and both compounds together by 136 ± 21% ( ), therefore bicuculline did not occlude the effect of ACEA on histamine release. The administration of bicuculline likely induces strong local depolarization, which, in turn, might induce release of endocannabinoids. We tested this hypothesis co-administering AM251, at the concentration (200 nM) that fully blocked the effect of mAEA, together with bicuculline (10 μM). After collection of four 15-min baseline samples, AM251 was added to the perfusing medium for 15 min, and then it was administered in combination with bicuculline ( ). A significant increase (85 ± 17%) was achieved during the second, 15-min period of TMN perfusion with AM251 and bicuculline and returned to basal levels during wash out in the presence of AM251 (n = 6).

We examined if CB1-r are localized on histaminergic neurons in the hypothalamus by performing double immunofluorescence labelling of hypothalamic slices using a combination of anti-CB1-r antibodies and anti-HDC antibodies to identify histaminergic neurons. The distribution profile of HDC-immunoreactive neurons in the E2–E3 subdivisions of the TMN is shown in . CB1-r immunostaining was sparse in this region, indicating the presence of very few CB1-r-expressing fibres. A higher density of CB1-r immunostaining was found in the E4–E5 subdivisions of the TMN ( ), where CB1-r immunostaining apparently surrounds clusters of HDC-immunonegative cells. Analysis of optical scan volumes showed that CB1-r immunostaining did not co-localize to HDC positive cell bodies ( , insets), nor dendrites (not shown). Despite the very low CB1-r immunostaining in the hypothalamus, CA3 hippocampal neurons were densely immunolabelled ( ), as previously reported by Tsou et al. (1998) . Analysis of optical scan volumes in this region as well failed to show co-localization of CB1-r and HDC immunostaining (not shown). The specificity of CB1-r immunostaining in the hypothalamus was further confirmed preabsorbing the anti-CB1-r antibodies with glutathione-S-transferase-conjugated fusion protein ( ).

As shown in , after collection of four 15-min baseline samples, 200 nM AM251 was added to the perfusing medium for 15 min, and then it was administered in combination with 100 μM AM404 for an additional 60 min. AM251 completely blocked the effect of AM404 on histamine release in the TMN (n = 3).

To confirm that the excitatory effects of cannabimimetic compounds were mediated through activation of CB1-r, the CB1-r antagonist AM251 was co-administered together with the CB1-r agonist mAEA in the TMN, and histamine release was measured both in the TMN and NBM. After collection of four, 15-min baseline samples, 200 nM AM251 was added to the perfusing medium for 15 min, and then it was administered in combination with the CB1-r agonist for an additional 60 min. Co-application of AM251 with mAEA in the TMN completely abolished the stimulatory effect of mAEA on histamine release, both in the TMN and NBM ( ; n = 6).

As endocannabinoid action is presumably terminated by cellular uptake via an endocannabinoid membrane transporter ( Piomelli, 2003 ) and by amidohydrolysis ( Deutsch & Chin, 1993 ; McKinney & Cravatt, 2005 ), we tested whether endocannabinoids reached a sufficient concentration to activate their receptors and modulate the activity of histaminergic cells by blocking their uptake and metabolism in the hypothalamus. Presumably, both mechanisms are blocked by AM404 ( Jarrahian et al., 2000 ). When AM404 (100 μM) was added to the TMN-perfusing medium for 60 min, histamine release in the TMN increased by a maximum of 165 ± 58%, but was not significantly changed in the NBM ( ; n = 5). In the TMN, a significant increase was achieved during the fourth, 15-min period of perfusion in the presence of AM404 and persisted for approximately 45 min during wash out, after which basal histamine levels were attained. Perfusion of the TMN for 60 min with the selective CB1-r antagonist AM251 at a concentration of either 200 nM ( ; n = 5) or 10 μM (not shown; n = 3) did not change spontaneous histamine release significantly from the TMN ( , lower panel) and consequently from the NBM ( , upper panel).

Perfusion of the TMN for 60 min with either 150 nM ACEA (n = 3; ) or 1 μM mAEA (n = 4; ) increased histamine release in the TMN as well as the dorsal striatum. Histamine levels returned to basal values during wash out of the compounds. ACEA increased histamine release significantly, up to 91 ± 35% in the TMN and up to 173 ± 65% of basal levels in the striatum during the third, 15-min period of pharmacological stimulation. Similarly, 1 μM mAEA infused in the TMN augmented histamine release locally to a maximum of 88 ± 40% of basal values ( ), whereas histamine release from the dorsal striatum was augmented up to a maximum of 145 ± 63%.

Using a double-probe microdialysis protocol, ACEA was infused locally in the TMN and histamine release was monitored from the TMN and NBM. As shown in , 60 min perfusion with 150 nM ACEA induced a significant increase in histamine release from both the TMN and the NBM. In the TMN a significant increase of histamine release was present during the first 15-min fraction of ACEA administration and reached a maximum level of 122 ± 7% of spontaneous release (n = 5). Histamine release returned gradually to baseline values during ACEA wash out. Histamine release also increased in the NBM when ACEA was infused into the TMN and it reached a maximal level of approximately 400 ± 146% within the first 30 min of TMN perfusion. A similar, significant increase of histamine release from both the TMN and NBM was also observed following perfusion of the TMN for 60 min with 1 μM mAEA, a synthetic analogue of the endocannabinoid anandamide resistant to degradation ( Abadji et al., 1994 ). In the TMN, the maximal effect was of similar magnitude and time course to that observed with ACEA (the maximal increase was 155 ± 78%; n = 4; ). In the NBM, histamine release increased significantly during the first 15-min fraction of TMN perfusion with 1 μM mAEA, and the maximal effect was 98 ± 34% of spontaneous, baseline release (n = 4). Histamine output tended to remain elevated for the duration of mAEA application to the TMN and then returned quickly to baseline values in both the TMN and NBM after mAEA perfusion ended.

The effect of a single intraperitoneal injection of the selective and potent CB1-r agonist ACEA ( Hillard et al., 1999 ) on histamine release from the TMN is shown in . ACEA (3 mg/kg, i.p.; n = 8) significantly increased histamine efflux from the TMN 60 min after the injection compared with vehicle treatment (n = 4), up to a peak value of 84 ± 30%. Histamine release slowly returned to basal levels over the next 60 min.

Discussion

In summary, cannabinoids activate the histaminergic system in the rat brain in vivo. Acute i.p. administration of the CB1-r agonist ACEA augmented histamine release from the TMN, where histaminergic somata are located. This effect was mimicked when the selective CB1-r agonists ACEA or mAEA were infused locally into the TMN of freely moving rats via a microdialysis probe, indicating that modulation of histamine release is a local phenomenon mediated within the hypothalamus. Supposedly, histamine is released by short histaminergic projections within the posterior hypothalamus, as TMN neurons have extensive axonal arborizations within this brain region (Haas & Panula, 2003). TMN perfusion with CB1-r agonists increased histamine release not only from the TMN, but also from two histaminergic projection areas, the NBM and striatum. Modulation of histamine release in these regions presumably results from the activation of ascending histaminergic projections. Furthermore, we found that exogenous and endogenous cannabinoids differed significantly in the regulation of histaminergic cells activity: CB1-r agonists affected histamine release not only at the site of perfusion, the TMN, but also in the NBM and striatum, whereas the endocannabinoid membrane uptake blocker AM404 when administered in the hypothalamus increased histamine release only in the TMN.

Specificity of cannabinoid action The CB-r ligands used in this study are selective for the CB1-r and the concentrations used in the microdialysis experiments were consistent with the range of concentrations considered selective for this receptor. In our study, cannabimimetic compounds were used at similar or lower concentrations than those effective to modulate synaptic activity in vitro (e.g. Melis et al., 2004; Marcaggi & Attwell, 2005). The CB1-r antagonist AM251 effectively blocked both mAEA- and AM404-induced increase of histamine release at a concentration (200 nM) well below the range that decreases glutamate release from striatal synaptosomes (Köfalvi et al., 2003), and that inhibits the depolarization of synaptoneurosomes induced by the sodium channel, site 2-specific neurotoxin veratridine (Liao et al., 2004). Furthermore, the observation that AM251 blocked AM404-induced increase of histamine release in the TMN argues against the possibility that AM404 effect may be explained by its affinity for the vanilloid VR1 receptor (De Petrocellis et al., 2000).

Exogenous vs. endogenous cannabinoid action The current theory indicates that endocannabinoids are released via activity-dependent cleavage of membrane lipid precursors and are immediately released from the cells and metabolized (Di Marzo et al., 1994; Freund et al., 2003; Piomelli, 2003). Endocannabinoid release for a single neuron is a relatively rare occurrence, but release probability may be increased by the convergence of synchronous synaptic events (Varma et al., 2001; Kim et al., 2002). Indeed, intrahypothalamic administration of AM251 did not modify histamine release, suggesting that under resting conditions endocannabinoids tone is insufficient to modulate the activity of histaminergic neurons. Furthermore, the lack of an effect on histamine release at both AM251 concentrations tested (200 nM and 10 μM) suggests that CB1-r are not constitutively active. However, AM404 increased histamine release from the TMN, strongly suggesting that AM404 inhibited local endocannabinoid clearance and allowed sufficient accumulation to activate CB1-r. The increased endocannabinoid tone produced by AM404 augmented histamine release only in the TMN presumably by activating a more restricted, or different population of CB1-r than those activated by the administration of direct-acting CB1-r agonists. In this regard, it was recently shown that constitutive release of endocannabinoids by hypothalamic, proopiomelanocortin neurons inhibits a GABAergic tone, whereas exogenous CB1-r agonists inhibit glutamate release as well (Hentges et al., 2005). Alternatively, endocannabinoids may increase histamine release in a subpopulation of TMN neurons and, in turn, the released histamine may activate autoinhibitory H 3 receptors on other parts of the TMN. Consistent with these observations, in vivo studies demonstrated that administrations of AM404 or URB597, an inhibitor of the metabolic pathway of the endocannabinoid anandamide, do not mimic the full spectrum of pharmacological responses produced by classical CB1-r agonists (Gaetani et al., 2003; Solinas et al., 2005). Therefore, understanding in what circumstances endocannabinoids are released and activate histaminergic cells warrants further investigations (see below).

CB1-r are not expressed on histaminergic neurons and do not attenuate GABA release in the hypothalamus Extensive analysis of the histaminergic neuropil within the hypothalamus and of histaminergic projections to the hippocampus did not reveal co-expression of CB1-r protein and HDC in the same cells. Therefore, our data strongly suggest that CB1-r are not localized onto histaminergic neurons, which makes a direct excitatory effect of CB1-r activation on histamine release unlikely. CB1-r immunoreactivity in the hypothalamus was overall very low, as previously reported (Moldrich & Wenger, 2000), although CB1-r mRNA is expressed in hypothalamic neurons that release neuropeptides known to modulate food intake (Cota et al., 2003). This study did not examine if histaminergic neurons express CB1-r mRNA, therefore this possibility cannot be completely excluded. A possible explanation for the excitatory effects of CB1-r agonists on histamine release is therefore the blockade of an inhibitory action exerted on histaminergic neurons. Four lines of evidence strongly suggest that the cannabinoids modulate the activity of histaminergic neurons independently of GABAergic neurotransmission in the hypothalamus: (1) CB1-r agonists did not decrease GABA release in the TMN; (2) ACEA, but not bicuculline, administered in the TMN induced a significant increase of histamine release in the NBM suggesting different modes of action; (3) bicuculline did not occlude the increment in histamine release induced by ACEA in the TMN; (4) the CB1-r antagonist AM251 did not block the effect of bicuculline on histamine release from the TMN. Direct inhibitory actions on TMN neurons have been found also for galanin and nociceptin (Eriksson et al., 2000). However, preliminary studies in our laboratory showed that intra-TMN administration of UFP-101, an antagonist of the nociceptin receptor ORL 1 , did not modify histamine release (unpublished observations), indicating that TMN neurons are not tonically inhibited by nociceptin. Therefore, it is unlikely that cannabinoids effect on histamine release depend on depression of nociceptin-mediated hyperpolarization of histaminergic neurons. The mechanism by which cannabinoids increase histamine release therefore remains to be elucidated.