Constitutive histamine H2 receptor activity regulates serotonin release in the substantia nigra

Authors


Address correspondence and reprint requests to Sarah Threlfell, Department of Neuroscience, Rosalind Franklin University of Medicine and Science, 3333 Green Bay Road, North Chicago, IL 60064, USA.
E-mail: sarah.threlfell@rosalindfranklin.edu

Abstract

The substantia nigra pars reticulata (SNr) forms a principal output from the basal ganglia. It also receives significant histamine (HA) input from the tuberomammillary nucleus whose functions in SNr remain poorly understood. One identified role is the regulation of serotonin (5-HT) neurotransmission via the HA-H3 receptor. Here we have explored regulation by another HA receptor expressed in SNr, the H2-receptor (H2R), by monitoring electrically evoked 5-HT release with fast-scan cyclic voltammetry at carbon-fiber microelectrodes in SNr in rat brain slices. Selective H2R antagonists (inverse agonists) ranitidine and tiotidine enhanced 5-HT release while the agonist amthamine suppressed release. The ‘neutral’ competitive antagonist burimamide alone was without effect but prevented ranitidine actions indicating that inverse agonist effects result from constitutive H2R activity independent of HA tone. H2R control of 5-HT release was most apparent (from inverse agonist effects) at lower frequencies of depolarization (≤ 20 Hz), and prevailed in the presence of antagonists of GABA, glutamate or H3-HA receptors. These data reveal that H2Rs in SNr are constitutively active and inhibit 5-HT release through H2Rs on 5-HT axons. These data may have therapeutic implications for Parkinson’s disease, when SNr HA levels increase, and for neuropsychiatric disorders in which 5-HT is pivotal.

Abbreviations used
[5-HT]o

extracellular concentration of 5-HT

5-HT

5-hydroxytryptamine or serotonin

CFM

carbon fiber microelectrode

FCV

fast-scan cyclic voltammetry

H2R

histamine H2 receptor

H3R

histamine H3 receptor

HA

histamine

M-WU

Mann–Whitney U test

PD

Parkinson’s disease

R-mHA

(R)-(-)-α-methylhistamine dihydrobromide

(S)-MCPG

(S)-α-methyl-4-carboxyphenylglycine

SN

substantia nigra

SNr

substantia nigra pars reticulata

The substantia nigra pars reticulata (SNr) is a key output nucleus of the basal ganglia. Projections from multiple nuclei converge at the SNr to regulate nigrothalamic output (Alexander and Crutcher 1990). The SNr is also a target of a variety of monoamine pathways, including dopamine released somatodendritically from dopaminergic neurons as well as 5-hydroxytryptamine (5-HT) from raphé inputs. However, a less discussed but nonetheless significant amine input to the SNr is a projection from histamine (HA)-containing neurons of the tuberomammillary nucleus (Panula et al. 1989; Threlfell et al. 2004). There is growing evidence for an important role for HA and HA receptor ligands in the substantia nigra (SN) in the regulation of motor function and its manipulation in Parkinson’s disease (PD) (Garcia-Ramirez et al. 2004; Chen et al. 2005; Gomez-Ramirez et al. 2006). In PD, the density of histaminergic fibers in the SN is increased (and their morphology altered) (Anichtchik et al. 2000) and furthermore, the density of histamine H3 receptor (H3R) binding sites is elevated (Anichtchik et al. 2001). Moreover, H3R agonists reduce 3,4-dihydroxy-L-phenylalanine (l-DOPA)-induced chorea in experimental parkinsonism (Gomez-Ramirez et al. 2006). HA systems may therefore offer an effective target for alleviation of side effects associated with long-term dopamine replacement therapy in PD.

The cellular functions of HA in the SN documented to date include a powerful action at H3Rs to regulate 5-HT release (Threlfell et al. 2004). The release of 5-HT in turn plays a key role in regulating dopamine and GABA transmission in SNr (Rick et al. 1995; Stanford and Lacey 1996; Di Matteo et al. 2001; Cobb and Abercrombie 2003). The neuromodulation of 5-HT systems has consequently been suggested to offer opportunities for novel therapeutics in PD (Nicholson and Brotchie 2002); HA receptor modulation of 5-HT may be one such candidate mechanism (Gomez-Ramirez et al. 2006).

The H3R is not the only HA receptor subtype in the SNr: the H2–type receptor is also found at high densities in the SNr, as well as in other nuclei in the basal ganglia (Vizuete et al. 1997; Pillot et al. 2002). The H2-receptor (H2R) is best known as a G-protein coupled receptor involved in stimulating gastric acid secretion (Black et al. 1972). The role of the H2R in the central nervous system however is less well understood, but has been proposed to include modulation of motor circuits and even neuroprotection (Hamami et al. 2004; Malagelada et al. 2004; Adachi et al. 2005; Chen et al. 2005). H2Rs in the brain are located pre-and postsynaptically (Poli et al. 1994; Timm et al. 1998; Chen et al. 2005) where their actions include inhibition (Haas and Bucher 1975; Haas and Wolf 1977) or excitation (Haas and Konnerth 1983; Haas and Greene 1986; Chen et al. 2003) of target neurons. Rat and human H2Rs have been identified to be constitutively active (i.e. can spontaneously activate intracellular signalling pathways), when expressed in non-neuronal cell lines (Smit et al. 1996; Alewijnse et al. 1998, 2000; Esbenshade et al. 2003). It has remained unidentified however, whether H2Rs are constitutively active in the CNS.

In this study we explored whether H2Rs regulate 5-HT release in the SNr. We used fast-scan cyclic voltammetry (FCV) at carbon-fiber microelectrodes (CFMs) as published previously (Threlfell et al. 2004) in order to monitor in real time the modulation of electrically evoked 5-HT release in the SNr by the H2R.

Materials and methods

Slice preparation for FCV

Male Wistar rats (150–300 g) were anaesthetized with isoflurane by inhalation and decapitated and their brains removed. Midbrain slices were prepared according to previously published methods (Cragg et al. 1997; Threlfell et al. 2004). A block of midbrain was mounted onto a specimen plate with cyanoacrylate adhesive and placed in the double walled buffer tray in the Leica VT1000S vibratome. The buffer tray was then surrounded by ice and filled with ice-cold HEPES ringer containing (in mM): NaCl (120), KCl (5), NaHCO3 (20), HEPES acid (6.7), HEPES salt (3.3), CaCl2 (2), MgSO4 (2), KH2PO4 (1.2) and glucose (10), saturated with 95% O2/5% CO2. Slices 350 μm-thick were cut and maintained in individual vials in HEPES ringer for ≥ 1 h at room temperature (20°C) before transferal to the recording chamber. The coordinates of midbrain slices containing the SNr correspond to A3.2–A4.2 mm of adult rat, according to the atlas of Paxinos and Watson (1982). The slices were equilibrated with the superfusion medium of the recording chamber at 30–32°C for an additional 30 min prior to the beginning of the experiment. The superfusion medium was a bicarbonate-buffered artificial cerebrospinal fluid maintained at 30–32°C and containing (in mM): NaCl (124), KCl (3.7), NaHCO3 (26), CaCl2 (2.4), MgSO4 (1.3), KH2PO4 (1.3), and glucose (10) saturated with 95% O2/5% CO2. Superfusion flow rate was approx 1.5 mL/min.

Voltammetry and carbon-fiber microelectrodes

All recordings were made using a Millar Voltammeter (P.D. Systems, West Molesey, UK) as previously described (Cragg and Greenfield 1997; Cragg et al. 1997; Threlfell et al. 2004). The instrument was used as a three-electrode potentiostat with an Ag/AgCl wire as reference electrode and the stainless steel flow outlet of the bath as the auxiliary. Voltammograms were obtained in the simple triangle wave mode with a voltage ramp of −0.7 to +1.3 V then back to −0.7 V versus Ag/AgCl at a scan rate of 800 V/s. The scan was switched out of circuit between scans. Sampling frequency was 8 Hz. Evoked extracellular concentrations of 5-HT ([5-HT]o) were measured using FCV with carbon fiber microelectrodes (CFMs) (fiber tip diameter ∼10 μm; exposed fiber length, ∼100 μm; World Precision Instruments, Stevenage, UK). CFMs were positioned 100 μm into the tissue with the aid of a binocular microscope.

Electrode calibrations were performed post-experiment in the recording chamber at 32°C using 0.25 μM 5-HT made up in all experimental solutions. Calibration solutions were made immediately prior to use from stock solutions of 5-HT in 0.1 M HClO4. Sensitivity to 5-HT varied with electrode from 50–200 nA/μM. The minimum detection limit for 5-HT was ∼5–10 nM (2 × noise). Peak [5-HT]o detected in the SNr in control conditions after 20 pulse, 50 Hz stimulation was typically 40–90 nM.

Electrical stimulation

Concentric bipolar stimulating electrodes (FHC, Bowdoin, ME, USA) were used to electrically evoke 5-HT release in SNr. The stimulating electrode was positioned on the tissue surface approx 50–100 μm from the CFM with the aid of a binocular microscope. Stimulus pulses (0.2 ms duration; 0.6–0.7 mA) were applied in trains of 40 pulses at 20 Hz unless otherwise stated, at inter-stimuli intervals of 5 min; by this time release is fully reproducible and can be maintained for > 3 h. The evoked 5-HT release is Ca2+ and Na+-channel dependent, and sensitive to depolarization frequency and pulse number (Threlfell et al. 2004). Experiments to deduce the activity-dependence of H2R effects were performed using stimuli of (i) varying duration at a fixed frequency (0.4–4 s, 20 Hz), or (ii) varying frequency at a fixed pulse number (10–100 Hz, 40 pulses), or (iii) varying frequency at fixed duration (10–100 Hz, 2 s). Each variable was delivered in a random order.

Drugs

d-(–)-2-Amino-5-phosphonopentanoic acid (d-AP5), GYKI 52466 hydrochloride, saclofen, (S)-α-methyl-4-carboxyphenylglycine [(S)-MCPG], (R)-(-)-α-methylhistamine dihydrobromide (R-mHA), ranitidine hydrochloride (ranitidine), thioperamide maleate (thioperamide) and tiotidine were obtained from Tocris Cookson Ltd (Bristol, UK). Burimamide was a generous gift from Glaxo SmithKline (Harlow, UK). All other compounds were obtained from Sigma-Aldrich (Poole, UK). Drugs were dissolved in water, dilute NaOH [(S)-MCPG], dilute HCl (GYKI 52466 hydrochloride), ethanol (picrotoxin) or dimethyl sulfoxide (tiotidine) to make stock aliquots at 1000 or 10 000 × final concentrations, and stored at −20°C until required. Stock aliquots were diluted with artificial cerebrospinal fluid to final concentration immediately prior to use, and perfused for > 7 min when drug effects were stable before data were included in drug analysis.

Histamine itself had pronounced electrochemical activity which greatly compromised electrode sensitivity to 5-HT and precluded its use (Threlfell, Cragg and Greenfield, unpublished observations). Ranitidine and tiotidine were used as H2R inverse agonists since alternative available compounds were found to have significant interfering electrochemical activity, including zolantidine and ICI 162,846 (not illustrated). After initial characterization of effects of the inverse agonists ranitidine and tiotidine (Fig. 1), subsequent experiments to explore H2R effects further used only ranitidine in preference to tiotidine: tiotidine desensitized the sensitivity of CFMs to 5-HT (tiotidine concentrations > 10 μM, not illustrated). Note that the reduced signal:noise ratio that accompanies this desensitization in tiotidine gives rise to the greater apparent noise observed in the 5-HT electrochemical signals at higher tiotidine concentrations (see Fig. 1b). It should also be noted that thioperamide is an antagonist at not only the H3R, but also the H4R. However, the H4R is present only in peripheral tissues and not in the brain (Oda and Matsumoto 2001) and therefore, in this study we refer to thioperamide as an H3R antagonist.

Figure 1.

 Inhibition of 5-HT release in SNr by H2Rs is indicated by effects of inverse agonists and an agonist. (a–c), Top, Mean profiles of [5-HT]o versus time evoked by 2 s, 20 Hz (solid bar) normalized to mean peak control. Bottom, Concentration-response curves for mean peak [5-HT]o normalized to mean peak control values for (a) ranitidine, (b) tiotidine, or (c) amthamine. (a) Inset voltammograms depict typical voltammograms for evoked 5-HT release in SNr in control and in the presence of 50 μM ranitidine. The dual reduction current profiles at potentials of ∼0 and −550 mV (dotted lines) confirm the identity of the electrochemical signals as 5-HT. (a) Ranitidine concentration dependently enhanced [5-HT]o (= 12–22, Kruskal–Wallis; Dunn’s t-test vs. control, **< 0.01, ***< 0.001). Curve-fit is a sigmoidal concentration-response curve, R= 0.52: IC50 = 4.4 μM, Hill slope = 1.0. (b) Tiotidine concentration dependently enhanced [5-HT]o (= 9–12, Kruskal–Wallis; Dunn’s t-test vs. control, **< 0.01). Curve-fit is a sigmoidal concentration-response curve, R= 0.33: IC50 = 0.8 μM, Hill Slope = 1.0. (c) Amthamine concentration-dependently suppressed evoked [5-HT]o (**< 0.01, Kruskal–Wallis and Dunn’s t-test vs. control, = 7–20). Curve-fit is a sigmoidal concentration-response curve, R= 0.88: IC50 = 1.9 μM.

Data analysis and statistics

Data were acquired and analyzed using Strathclyde Whole Cell Program (University of Strathclyde). Current sampled at the 5-HT oxidation peak potential was measured from the baseline of each voltammogram to provide profiles of [5-HT]o versus time. This procedure minimizes inclusion of contributions from other electroactive and non-electroactive species to the neurotransmitter oxidation current (Rice and Nicholson 1989; Venton et al. 2003). Illustrated data represented are means ± SEM, = number of observations. The number of animals used in each experiment was 3–5. Illustrated voltammograms have background current subtracted. Comparisons for differences in means were assessed for non-parametric or normalized data by unpaired Mann–Whitney U-tests (M-WU) or where appropriate by Kruskal–Wallis variance analysis followed by Dunn’s or Bonferroni t tests. Concentration-response curves were generated using non-linear regression to a sigmoidal curve and a cubic spline curve was generated for stimulus duration data. (GraphPad Prism, San Diego, CA, USA).

Results

Constitutively active H2Rs inhibit 5-HT release in the SNr

The release of 5-HT in rat SNr was readily evoked by 20 Hz stimuli (2 s) as reported previously in rat and mouse (Threlfell et al. 2004; John et al. 2006). Evoked voltammograms in SNr (inset in Fig. 1a, bottom) show typical electrochemical signals observed both in control and in the presence of 50 μM ranitidine; the characteristic dual reduction currents of electrochemical signals at approximately 0 mV and −500 mV, as seen with applied 5-HT (Threlfell et al. 2004), confirms their identity as 5-HT. This is consistent with previous reports that 5-HT is the predominant neuroamine detected in rat SNr (Cragg et al. 1997; Threlfell et al. 2004).

The extracellular concentration of 5-HT ([5-HT]o) evoked by a 20 Hz stimulus train (2 s) in the SNr was significantly enhanced by bath application of either of two selective H2R antagonists, the ‘inverse agonists’ ranitidine, or tiotidine (Fig. 1a and b). Each drug displayed a significant concentration-dependent enhancement of mean peak evoked [5-HT]o (Kruskal–Wallis tests; < 0.001) (ranitidine, IC50 4.4 μM; = 12–22; tiotidine, IC50 0.8 μM; = 9–12). Mean peak [5-HT]o was enhanced by ranitidine (200 μM) from 100 ± 5% in control to 208 ± 14% in ranitidine (= 12; Dunn’s post-hoc t-test vs. control, < 0.001) and by tiotidine (50 μM) from 100 ± 10% in control to 200 ± 30% in tiotidine (= 11; Dunn’s post-hoc t-test vs. control, < 0.01), indicating that H2Rs can limit 5-HT release. Furthermore, the H2R agonist amthamine reversibly suppressed evoked 5-HT release in a concentration-dependent manner (20 Hz, 2 s stimulus train) (Fig. 1c; < 0.001, Kruskal–Wallis test, = 7–20) to levels of [5-HT]o that were below the 5-HT detection limit.

Notably, the ‘neutral’ competitive rat H2R antagonist burimamide (Smit et al. 1996) when applied alone (1–50 μM) was without effect on evoked [5-HT]o (Fig. 2a; = 21–23) indicating, as expected and as previously reported, that there is little spontaneous HA tone at HA receptors in coronal midbrain slices (Threlfell et al. 2004). However, burimamide (50 μM) could at least partially reverse the effects of the H2R agonist amthamine (1 μM) (Fig. 2b; = 14–17; Bonferroni post-hoc t-test, < 0.01) confirming that the burimamide compound was pharmacologically active and that amthamine effects are via the H2R. More importantly however, the striking disparity between the lack of effect of the ‘neutral’ competitive antagonist burimamide alone (Fig. 2a) versus the powerful enhancement of release by the ‘inverse agonists’ ranitidine and tiotidine (Fig. 1) suggests that H2R in SNr are constitutively active i.e. active in a manner that is independent of the presence of HA.

Figure 2.

 5-HT release in SNr is not modified by a competitive H2R antagonist. (a–b) Mean profiles of [5-HT]o versus time evoked by 2 s, 20 Hz (solid bar) normalized to mean peak control in control and in the presence of (a) burimamide (b) amthamine, or amthamine and burimamide combined or (c) ranitidine, or ranitidine and burimamide combined. (a) Burimamide alone (1–50 μM) did not modify evoked [5-HT]o (> 0.05, Kruskal–Wallis test, = 21–23). (b) Burimamide (50 μM) could partially reverse the inhibition of 5-HT release by amthamine (1 μM) (††< 0.01, Bonferroni post-hoc t-test, = 14–17). (c) Ranitidine alone (5 μM) significantly enhanced evoked [5-HT]o (< 0.001, Dunn’s t-test). Burimamide (50 μM) reversed the enhancement of 5-HT release by ranitidine (5 μM) (> 0.05, Dunn’s t-test, = 15–27). **< 0.01, ***< 0.001.

We further confirmed the hypothesis that the effects of the H2R inverse agonists appear to be because of constitutive activity at the H2R by assessing whether the effect of ranitidine could be prevented by application of the neutral antagonist burimamide. Ranitidine (5 μM) enhanced release by 139% and subsequent application of burimamide (50 μM) completely reversed the effects of the H2R inverse agonist ranitidine by reducing release to levels not different from control (Fig. 2c; n = 17–23; Dunn’s t-test control vs. burimamide plus ranitidine, > 0.05).

Control of 5-HT release by H2R is frequency dependent

To assess the impact of H2R activity on the control of 5-HT release during a full range of possible activities of 5-HT neurons, the effect of an H2R inverse agonist were explored during stimulus trains which were varied in duration and frequency. Firstly, the duration of a 20 Hz stimulus train was varied from 0.4 s to 4 s. In control conditions, evoked [5-HT]o varied significantly with train length (Kruskal–Wallis, p < 0.001) up to ∼2 s when longer trains did not enhance [5-HT]o further (Dunn’s t test; 2 s vs. 4 s, > 0.05, Fig. 3aleft panel). The inverse agonist ranitidine (50 μM) significantly enhanced mean peak [5-HT]o compared to control for all stimuli (Fig. 3aleft panel, M-WU; < 0.001–0.05) in a manner that varied significantly with train duration (Kruskal–Wallis; < 0.001): There was an increasing enhancement of [5-HT]o by ranitidine with increasing duration of stimulus that ranged from ∼150% release compared to controls at short stimuli (0.4 s) up to approx 200% release at longer stimuli (2–4 s) (Fig. 3aright panel; = 11–13). This increasing effect of ranitidine appeared to be maximal by ∼ 2 s (Dunn’s t test; 2 s vs. 4 s, > 0.05).

Figure 3.

 Degree of regulation of 5-HT release in SNr by H2Rs depends upon membrane activity (duration and frequency). (a) Left, Plot of mean peak [5-HT]o versus stimulus train duration. Data are normalized to lowest control release (eight pulses, 20 Hz). Ranitidine (50 μM) significantly enhanced [5-HT]o evoked by stimuli of duration ranging from 0.4 s to 4 s at 20 Hz compared to control (= 11–13; M-WU, *< 0.05,**< 0.01,***< 0.001). Right, Plot of percentage of mean peak [5-HT]o in ranitidine versus control (ranitidine:control %). The percentage increases with duration of stimulus train. (b) Left, Plot of mean peak [5-HT]o versus stimulus frequency for trains of fixed number of pulses. Data are normalized to lowest control release (4 s, 10 Hz). Ranitidine (50 μM) significantly enhanced [5-HT]o evoked by stimulus trains of 40 pulses of frequency ranging from 10 to 100 Hz compared to control (= 11–13; M-WU, ***< 0.001). Right, Plot of the percentage of mean peak [5-HT]o in ranitidine versus control. The percentage increases with decreasing stimulation frequency (and thus increasing duration of stimulus train). (c) Left, Plot of mean peak [5-HT]o versus stimulus frequency. Data are normalized to lowest control release (2 s, 10 Hz). Stimulus duration was constant (2 s). Ranitidine (50 μM) significantly enhanced [5-HT]o evoked by 2 s stimulus trains of frequency 10–50 Hz compared to control (= 9–10; M-WU, **< 0.01,***< 0.001). Right, Plot of the percentage of mean peak [5-HT]o in ranitidine versus control. The percentage increases with decreasing stimulation frequency.

Secondly, the frequency of stimuli within a train was varied across the physiological range of frequencies at which 5-HT neurons can fire (Lakoski and Aghajanian 1983; Hajos et al. 1995). Stimuli ranged from 10 to 100 Hz using trains with a fixed number of pulses (40 pulses: i.e. 10 Hz for 4 s, or 100 Hz for 0.4 s). In control conditions, evoked [5-HT]o varied significantly with frequency (Kruskal–Wallis, < 0.001) as reported previously (Iravani and Kruk 1997; Bunin et al. 1998; Threlfell et al. 2004; John et al. 2006) (Fig. 3b, left panel). However, ranitidine (50 μM) significantly enhanced mean peak [5-HT]o compared to control for all stimuli (Fig. 3bleft panel, M-WU; < 0.001), in a manner that varied inversely and significantly with frequency (Kruskal–Wallis; < 0.001). The enhancement of [5-HT]o by ranitidine compared to controls ranged from ∼150% release at 100 Hz up to nearly 300% release at the lowest frequency, 10 Hz (Fig. 3b, right panel; = 11–13).

To distinguish whether the effects of ranitidine on frequency (Fig. 3b) were because of the concomitant variations in duration (e.g. Fig. 3a) that accompany trains of variable frequency but fixed pulse number, a third protocol was used in order to vary stimulus frequency (from 10 to 100 Hz) during trains of constant duration (2 s, i.e. 20 pulses at 10 Hz, or 200 pulses at 100 Hz). In control conditions, evoked [5-HT]o varied significantly with frequency (Fig. 3c, left panel, Kruskal–Wallis, < 0.001). Ranitidine (50 μM) significantly enhanced mean peak [5-HT]o compared to control in a manner that varied inversely with frequency and only for frequencies ≤ 50 Hz (Fig. 3c, left panel; 10–50 Hz: M-WU, < 0.01–0.001; 100 Hz: M-WU, > 0.05). The enhancement of [5-HT]o by ranitidine compared to controls ranged from ∼120% release at 100 Hz to ∼200% release at 10 Hz (Fig. 3c, right panel; = 9–10). This observation of frequency-variable effects of ranitidine during stimulation trains of constant duration suggests a greater H2R control of 5-HT release during low frequency stimulations (Fig. 3b and c) that is independent from the greater control apparent during longer stimulus trains (Fig. 3a).

These data suggest that 5-HT release by low frequency neuron activity (≤ 20 Hz) is more limited by the H2R than release by high frequencies (> 20 Hz) but that the effects of this H2R limitation are most apparent during long periods, i.e. ongoing activity. The effects of the H2R appear to accumulate over time but not with increasing number of pulses in that time: H2R blockade is less apparent if the same number of pulses is presented over a shorter duration, e.g. high frequency. In other words, whereas H2Rs limit release by sustained low frequency activity, they pose less limitation on shorter, high frequency activity.

H2R control is distinct from the H3R

Inverse agonists but not neutral competitive antagonists at the H2R enhance [5-HT]o, and therefore these data suggest that some form of constitutive-like activity at the H2R can inhibit the release of 5-HT. Since previous work highlighted a potent inhibitory control of 5-HT release by another HA receptor subtype, the H3R, (Threlfell et al. 2004) we confirmed that the H2R ligand effects described in the current study were pharmacologically distinct from, and not attributable to, those previously reported at H3Rs (Threlfell et al. 2004). Firstly, the effect of ranitidine was explored in the presence of the H3R antagonist thioperamide. Thioperamide (10 μM) alone had no detectable effect on [5-HT]o (Fig. 4a, = 12) as previously reported (Threlfell et al. 2004) whereas the full effects of ranitidine (50 μM) continued to be observed in the presence of thioperamide: ranitidine enhanced [5-HT]o evoked by 20 Hz stimuli (2 s) by more than two-fold (212 ± 14%) compared to release in thioperamide alone (Fig. 4a, M-WU; < 0.001, = 12) i.e. to a level identical to that seen without thioperamide (e.g. Fig. 1a). Secondly, ranitidine was applied (50 μM) and then the effects of an H3R agonist, R-mHA (1 μM) were tested. Ranitidine did not prevent H3R-mediated inhibition of [5-HT]o by R-mHA. R-mHA effects were additive upon those of ranitidine. Ranitidine alone enhanced peak [5-HT]o evoked by 20 Hz stimuli (2 s) to 147 ± 8% (Fig. 4b) (M-WU; < 0.01). Subsequent application of R-mHA reduced evoked [5-HT]o to a level that (i) when compared to that seen in the presence of ranitidine alone (42 ± 3% of ranitidine; Fig. 4b, M-WU; < 0.01) was comparable to that seen in our prior experiments with R-mHA applied alone (Threlfell et al. 2004), and (ii) when compared to pre-ranitidine controls (62 ± 4% of pre-ranitidine controls) was consistent with additive but not competitive effects of ranitidine and R-mHA. Thus, these data reveal that H2R ligand effects occur separately from those of H3R ligands suggesting that the H2R control of 5-HT release reported here is pharmacologically distinct from that of the H3R.

Figure 4.

 H2R-mediated inhibition of 5-HT release is pharmacologically distinct from H3R. (a–b) Mean profiles of [5-HT]o versus time evoked by 2 s, 20 Hz (solid bar), normalized to mean peak control. (a) Evoked [5-HT]o was not significantly modified by application of H3R antagonist thioperamide (10 μM, M-WU; > 0.05; = 12). In the presence of thioperamide, ranitidine (50 μM) significantly enhanced mean peak [5-HT]o (M-WU; †††< 0.001; = 12). (b) Evoked [5-HT]o was significantly enhanced by application of ranitidine (50 μM, M-WU; **< 0.01; = 12–13). In the presence of ranitidine, H3R agonist R-mHA (1 μM) significantly diminished mean peak [5-HT]o (M-WU; ††< 0.01; = 12–13).

Direct or indirect H2R effects?

We explored whether H2Rs present on GABAergic or glutamatergic inputs to 5-HT terminals could be responsible for H2R control of 5-HT release, rather than H2Rs located on the 5-HT terminals themselves. The effects of the H2R inverse agonist ranitidine were investigated in the presence of cocktails of antagonists for either GABA or glutamate receptors (ionotropic and metabotropic). Application of GABA receptor antagonists picrotoxin (PTX, GABAA, 100 μM) and saclofen (GABAB, 50 μM) did not significantly alter mean peak [5-HT]o evoked by 20 Hz, 2 s stimulus (Fig. 5; M-WU; > 0.05). Moreover, in the presence of these blockers, significant control by H2Rs prevailed: subsequent application of ranitidine (50 μM) significantly enhanced mean peak [5-HT]o from 100 ± 4% in the presence of GABA receptor antagonists alone to 185 ± 12% (M-WU; < 0.001; = 12–13). The percentage enhancement of [5-HT]o by ranitidine in the presence of GABA receptor antagonists was not significantly different from that by ranitidine alone (M-WU; > 0.05).

Figure 5.

 H2R effects on 5-HT release persist during blockade of GABA synapses. Mean profiles of [5-HT]o versus time evoked by 2 s, 20 Hz (solid bar), normalized to mean peak control. Evoked [5-HT]o was not significantly modified by application of GABA receptor antagonists (PTX, 100 μM; saclofen, 50 μM; = 12–15; M-WU; > 0.05). In the presence of GABA receptor block, ranitidine (50 μM) significantly enhanced mean peak [5-HT]o (M-WU; †††< 0.001; = 12–13).

Application of glutamate receptor antagonists GYKI 52466 (non-competitive α-amino-5-hydroxy-3-methyl-4-isoxazole propionic acid/kainate antagonist, 10 μM), d-AP5 (NMDA receptor antagonist, 50 μM) and (S)-MCPG (non-selective group I/II metabotropic glutamate antagonist, 200 μM) did not significantly alter mean peak [5-HT]o evoked by 20 Hz, 2 s stimulus train from that in control (Fig. 6; M-WU; > 0.05). Moreover, in the presence of these blockers, significant control by H2Rs prevailed: ranitidine (50 μM) significantly enhanced mean peak [5-HT]o from 100 ± 5% in the presence of glutamate antagonists alone to 156 ± 11% (M-WU; < 0.05; = 11–13). The percentage enhancement of [5-HT]o by ranitidine in the presence of glutamate receptor antagonists was not significantly different from that by ranitidine alone (M-WU; > 0.05).

Figure 6.

 H2R effects on 5-HT release persist during blockade of glutamate synapses. Mean profiles of [5-HT]o versus time evoked by 2 s, 20 Hz (solid bar), normalized to mean peak control. Evoked [5-HT]o was not significantly modified by application of glutamate receptor antagonists [(S)-MCPG, 200 μM; d-AP5, 50 μM; GYKI 52466, 10 μM; = 11–13; M-WU; > 0.05]. In the presence of glutamate receptor block, ranitidine (50 μM) significantly enhanced mean peak [5-HT]o (M-WU; †< 0.05; = 11–13).

Discussion

This study has explored the histamine-serotonin interaction in the SNr, and revealed that the role of the histaminergic system in the SNr is not only powerful but complex. Previously, we identified that exogenous activation of H3Rs inhibits 5-HT release (Threlfell et al. 2004). The current study now reveals a potent role for an additional HA receptor, the H2R, which through an apparent constitutive activity inhibits 5-HT release. Together, these findings demonstrate that HA actions in SNr can be multifaceted and potent. Previous reports have suggested that H2Rs may play a neuromodulatory role in the basal ganglia (Prast et al. 1999; Hamami et al. 2004; Chen et al. 2005; Zhou et al. 2006); this report is the first to identify H2R-mediated control of 5-HT neurotransmission in the SN which furthermore, occurs through a mechanism involving constitutive receptor activity.

H2R control of 5-HT release in the SNr

It is increasingly well appreciated that G-protein coupled receptors can couple to their signal transduction pathways in the absence of their agonist i.e. they can exhibit constitutive activity. The H2R is well-known to exhibit constitutive activity, at least when re-expressed in cell lines (Smit et al. 1996; Alewijnse et al. 1998; Monczor et al. 1998, 2003), although it is not known whether this also typically occurs in the brain in situ. Thus, the drugs chosen to elucidate H2R function in SNr this study included the drugs ranitidine and tiotidine, which although traditionally viewed as antagonists, have been more accurately re-classified as inverse agonists: unlike classical or ‘neutral’ antagonists (such as burimamide), they should be capable of reducing the constitutive activity of H2Rs (Smit et al. 1996; Alewijnse et al. 1998; Monczor et al. 1998, 2003). By exploring the effects of these inverse agonists, in addition to the effects of an H2R agonist and classical antagonist, we have probed the role of H2Rs (and constitutive activity) in the regulation of 5-HT release in SNr.

Release of 5-HT in SNr was potently enhanced by H2R inverse agonists at concentrations appropriate for H2Rs (Prast et al. 1999), while the H2R agonist amthamine strongly inhibited release. By contrast, the ‘neutral’ competitive antagonist burimamide when applied alone did not modify release. Together these data suggest not only that activation of H2Rs can inhibit 5-HT release, but also that such regulation of 5-HT can occur through constitutive activity of the H2R independently of the availability of HA itself. In vivo, H2R inhibition of 5-HT release could be stronger yet, where the endogenous agonist, HA, would be expected to be found at higher concentrations owing to ongoing activity in intact HA neurons. The hypothesis that the H2R effects detected here arise owing to the constitutive activity of H2Rs in the SNr was supported further by the finding that the neutral antagonist burimamide was able to prevent the enhancement of 5-HT release by the inverse agonist ranitidine.

Interestingly, the IC50 value for amthamine-mediated inhibition of 5-HT release identified here was found to be approximately 10 times lower than the EC50 previously identified in the isolated gastric fundus [1.9 μM vs. 18.9 μM (van der Goot and Timmerman 2000)]. This observation may also be consistent with the idea that constitutively active receptors can exhibit higher affinity for agonists (Lefkowitz et al. 1993) and may therefore provide yet further evidence that the H2Rs regulating 5-HT release exhibit constitutive activity.

We were also able to discount non-selective actions of H2R drugs at an alternative HA receptor, the H3R, as an alternative explanation since an H3R antagonist did not prevent the effects of the H2R inverse agonist, and an H2R inverse agonist did not prevent the effects of an H3R agonist. Thus the effects of H2R ligands on 5-HT release in SNr were pharmacologically distinct from those reported for the H3R (Threlfell et al. 2004). Interestingly, the H3R antagonist, thioperamide, when applied alone did not affect 5-HT release, as shown previously (Threlfell et al. 2004). This suggests that the H3R itself is not constitutively active in the SNr, which is of note given that in other areas of the brain such as the cortex, the H3R has been shown to be constitutively active (Morisset et al. 2000).

Control of 5-HT release by H2R is frequency dependent

We explored the degree of neuromodulation by H2Rs on 5-HT release by assessing the impact of H2R activity during a range of physiologically relevant frequencies (and durations) of membrane depolarization activity. The degree of neuromodulation by H2Rs depends on neuron activity. H2Rs most powerfully limit 5-HT release evoked by low frequency neuron activity (≤ 20 Hz), especially during more prolonged stimuli. This dampening effect at low frequencies is more prominent over a greater duration of activity rather than number of pulses alone. Since release of 5-HT by high frequency stimuli (> 20 Hz) appears less limited by H2Rs than release by low frequencies, H2Rs may be effecting a form of frequency filtering i.e. a high frequency pass. This filtering could have physiological relevance to 5-HT function since serotonergic neurons of the raphé nuclei exhibit two firing modes in vivo, a slow, regular firing (≤ 5 Hz) (Aghajanian and Vandermaelen 1982) and a high frequency, burst firing of 1–3 spikes at ∼ 100 Hz (Hajos et al. 1995, 1996; Hajos and Sharp 1996a,b). And yet it remains unknown how burst firing of 5-HT neurons correlates with physiological behaviors. In vivo studies that have modeled burst firing of 5-HT neurons however, suggest that burst firing effectively enhances both the release and the postsynaptic effect of 5-HT (Gartside et al. 2000). Therefore the ability of 5-HT neurons to burst fire suggests that 5-HT is involved in more complex neural processing. A key outcome of H2R activity could be to differentiate more clearly the 5-HT signals released by high frequency bursts compared to the more H2R-limited release by low frequency, tonic activity, which may therefore be of importance in the regulation of more complex neural processing involving 5-HT, which is yet to be discovered.

A number of mechanisms require consideration that could explain the dynamic and activity-dependent control of 5-HT release by H2Rs that have been revealed here by the actions of inverse agonists. It seems unlikely that this control is because of the release of endogenous agonist HA (which might otherwise accumulate extracellularly over time to act at the H2R during the stimulation to limit 5-HT release) since the competitive H2R antagonist burimamide did not modify 5-HT release. This conclusion is also supported by the observation that an antagonist for an alternate HA receptor, the H3R (thioperamide), also does not reveal any control of evoked 5-HT release by endogenous HA. If the activity-dependent H2R effect were mediated by endogenous HA, we would have expected HA effects to also operate at the H3R since the affinity of HA for H3R is greater than that of H2R [reviewed in (Brown et al. 2001)].

An alternative type of mechanism that could account for the frequency selectivity of H2R could be an intracellular e.g. metabotropic or transmembrane event that occurs and accumulates during ongoing neuronal activity and is modified by H2R activity to limit 5-HT release probability at low frequencies. This release-limiting process could be increasingly overcome by any of the other processes which can occur to facilitate 5-HT release at higher frequencies e.g. intracellular Ca2+ summation and elevated release probability at high versus low frequencies [e.g.(Thomson 2000)]. Therefore, H2R effects observed here may be because of a cumulative, H2 activity-dependent effect resulting from constitutive receptor activity, known to occur for the heterologously expressed H2R (Smit et al. 1996; Alewijnse et al. 1998).

Direct or indirect H2R effects?

H2Rs are detected in the SNr by radioligand binding (Vizuete et al. 1997) but to date, there is no anatomical information available regarding the identity of the neurons that express H2Rs in the SNr or the ultrastructural location of these receptors. Thus, the synaptic circuitry through which H2Rs can regulate 5-HT release is unidentified. Since in situ hybridization for H2R mRNA reveals only low expression within the SN (guinea pig) (Vizuete et al. 1997) the H2Rs found abundantly in SN are unlikely to be localized on dopamine neurons in substantia nigra pars compacta or on projection neurons located within SNr. However, since H2R mRNA expression has been identified in striatum, subthalamic nucleus and pedunculopontine nucleus in guinea pig brain (Vizuete et al. 1997), regions which send major non-5-HT projections to the SNr, it is plausible that such inputs (among others) to SNr express H2Rs and are thereby involved in H2R-mediated inhibition of 5-HT release. The involvement of H2Rs on GABA neurons in particular might help explain why H2R activation, which is believed to be positively coupled to adenylyl cyclase via a stimulatory G-protein, might inhibit 5-HT release. Indeed, inverse agonists at H2Rs can enhance acetylcholine release in the striatum via an action at H2Rs present on GABA neurons (Prast et al. 1999).

Thus, the possibility that H2Rs present on GABAergic terminals in SNr contribute to H2R-mediated regulation of 5-HT release was investigated in this study using receptor antagonists for GABA. However, H2R-mediated regulation of 5-HT release shown with the inverse agonist ranitidine persisted in the presence of GABA synaptic block, to a degree identical to that seen by ranitidine alone. Similarly, experiments in the presence of receptor antagonists for glutamate confirmed that the glutamatergic input to SNr was not involved in this H2R-mediated regulation of 5-HT. A role for H2Rs on either GABA or glutamate terminals in the regulation of 5-HT release seems unlikely. In any event, neither GABA nor glutamate antagonists alone significantly modified [5-HT]o as previously reported (Threlfell et al. 2004), and thus the control of 5-HT release in this paradigm does not appear to be under direct control by GABA nor glutamate synapses. The present results thus indirectly support the possibility that H2R-mediated inhibition of 5-HT release is via H2Rs located on 5-HT terminals. We previously concluded that regulation of 5-HT in SNr by another HA receptor subtype, the H3R, may similarly be mediated by receptors located on 5-HT inputs (Threlfell et al. 2004). Together, these findings suggest a multifaceted regulation of 5-HT in SNr, by diverse HA receptors expressed on 5-HT neurons.

Conclusions

In this study we have identified a H2R-mediated inhibition of 5-HT release in SNr. H2Rs can powerfully limit 5-HT release during ongoing synaptic activity. Release by high frequencies appears less limited by the H2R that release by lower frequencies, and thus H2Rs may help to maintain the contrast in 5-HT signals released by different frequencies of 5-HT neuron activity. This effect may be mediated in part by constitutive activity of the H2R, an effect which could be magnified by any actions of the endogenous ligand (HA) in vivo. H2Rs located directly on 5-HT axons appears to be the most likely candidate receptors for this effect. Together with the previously reported inhibition of 5-HT release in SNr by H3Rs (Threlfell et al. 2004), these data indicate a potent and complex neuromodulation by histaminergic systems in the SNr.

Moreover, the capacity for HA to regulate 5-HT transmission in SNr via multiple HA receptors i.e. H2Rs as well as H3Rs (Threlfell et al. 2004), may provide attractive non-dopaminergic targets for improving therapies for PD. Serotonergic inputs to SNr (Corvaja et al. 1993; Moukhles et al. 1997) have the potential to modify motor output signals (Jacobs and Fornal 1993; Rick et al. 1995) via excitatory 5-HT2C receptors present on SNr projection neurons (Rick et al. 1995). In addition, pharmacological manipulation of H1, H2 and H3 receptors in the SNr have also been shown to modify the activity of SNr projection neurons (Zhou et al. 2006), which may in part be because of actions via HA receptors mediating regulation of 5-HT release here. Moreover, 5-HT is a target for therapeutic intervention in side effects associated with long term dopamine replacement therapy in PD (Nicholson and Brotchie 2002). Identifying novel approaches that manipulate 5-HT neurotransmission, such as the histaminergic neuromodulation revealed here, may help to enhance the efficacy of treatments not only for PD, but also for disorders of mood in which serotonergic manipulation is pivotal (Meltzer 1989).

Acknowledgements

We would like to thank Henrike Hartung for assistance in tissue preparation in some studies, and Dr Chris Hille (GSK) for assistance with the supply of, and information about, burimamide. The work was supported by Novartis Pharma AG, and a Paton Fellowship, University of Oxford (SJC).

Ancillary