Histamine

Jean-Charles Schwartz, Jean-Michel Arrang and Monique Garbarg

In a certain way, histaminergic systems have had a great, although indirect, historical importance in the development of neuropsychopharmacology. Indeed, the discovery of both the neuroleptic and tricyclic antidepressant drugs in the 1950s was derived from the clinical study of behavioral actions of "antihistamines", a class of antiallergic drugs now designated H1-receptor antagonists.

Nevertheless, the histaminergic neuronal system in brain, although already unraveled by the mid-1970s, has remained largely unexploited in drug design. Thus, only the traditional brain-penetrating H1-receptor antagonists, used as over-the-counter sleeping pills, are known to interfere with histaminergic transmissions in the central nervous system (CNS). This contrasts with the emergence, during the last decade, of a detailed knowledge of the system, which has revealed that it shares many biological and functional properties with other aminergic systems overexploited in CNS drug design.

Histamine and its receptors in the brain have recently been the subject of two comprehensive reviews (25,75). Therefore, in order to limit the length of the present chapter, we have deliberately chosen to summarize the detailed information that can be found in these reviews, adding only more recent information and major references.

One decade after Garbarg et al. obtained the first evidence of an ascending histaminergic pathway by lesions of the medial forebrain bundle (21), the exact localization of corresponding perikarya in the posterior hypothalamus was revealed immunohistochemically (61,82,94). Since then, the distribution, morphology and connections of histamine and histidine decarboxylase-immunoreactive neurons have been determined. These data were recently reviewed (60,81,86,96) and will be summarized only briefly here.

All known histaminergic perikarya constitute a continuous group of mainly magnocellular neurons (about 2,000 in the rat), located in the posterior hypothalamus and collectively named the tuberomammillary nucleus (a, b). It can be subdivided into medial, ventral and diffuse subgroups extending longitudinally from the caudal end of the hypothalamus to the midportion of the third ventricle. A similar organization was described in humans, except that histaminergic neurons are more numerous (~64,000) and occupy a larger proportion of the hypothalamus (2). Besides their large size (25-35 mm), tuberomammillary neurons are characterized by few thick primary dendrites, with overlapping trees, displaying few axodendritic synaptic contacts. Another characteristic feature is the close contact of dendrites with glial elements in a way suggesting that they penetrate into the ependyma and come in close contact with the cerebrospinal fluid, perhaps to secrete or receive still unidentified messengers. Neurons expressing mRNAs for histidine decarboxylase (EC 4.1.1.22), the enzyme responsible for the one-step histamine formation in the brain (75), were found by in situ hybridization in the tuberomammillary nucleus, but not in any other brain area (6). Tuberomammillary neurons possess the monoamine transporter 2 (62), which accounts for the histamine-releasing effect of reserpine (75).

The histaminergic neurons are characterized by the presence of an unusually large variety of markers for other neurotransmitter systems: glutamic acid decarboxylase, the gamma-aminobutyric acid (GABA)-synthesizing enzyme; adenosine deaminase, a cytoplasmic enzyme possibly involved in adenosine inactivation; galanin, a peptide co-localized with all other monoamines; (Met5)enkephalyl-Arg6Phe7, a product of the proenkephalin A gene; and other neuropeptides such as substance P, thyroliberin or brain natriuretic peptide. Tuberomammillary neurons also contain monoamine oxidase B, an enzyme responsible for deamination of tele-methylhistamine, a major histamine metabolite in brain. Finally a subpopulation of histaminergic neurons is able to take up and decarboxylate exogenous 5-hydroxytryptophan, a compound that they do not synthesize, however (81). Unraveling the functions of such a high number of putative co-transmitters in the same neurons remains an exciting challenge.

Analogous to other monoaminergic neurons, histaminergic neurons constitute long and highly divergent systems projecting in a diffuse manner to many cerebral areas (a). Immunoreactive, mostly unmyelinated, varicose or non-varicose fibers are detected in almost all cerebral regions, particularly limbic structures. Individual neurons have been confirmed to project to widely divergent areas (43). Ultrastructural studies suggest that these fibers make few typical synaptic contacts (85).

Fibers arising from the tuberomammillary nucleus constitute two ascending pathways: one laterally, via the medial forebrain bundle, and the other periventricularly. These two pathways combine in the diagonal band of Broca to project, mainly in an ipsilateral fashion, to many telencephalic areas; for example, in all areas and layers of the cerebral cortex, the most abundant projections are to the external layers. Other major areas of termination of these long ascending connections are the olfactory bulb, the hippocampus, the caudate putamen, the nucleus accumbens, the globus pallidus and the amygdaloid complex. Many hypothalamic nuclei exhibit a very dense innervation; for example, the suprachiasmatic, supraoptic, arcuate or ventromedial nuclei.

Finally, a long descending histaminergic subsystem arises also from the tuberomammillary nucleus to project to a variety of mesencephalic and brainstem structures such as the cranial nerve nuclei (e.g., the trigeminal nerve nucleus), the central gray, the colliculi, the substantia nigra, the locus coeruleus, the mesopontine tegmentum, the dorsal raphe nucleus, the cerebellum (sparse innervation) and the spinal cord.

Several anterograde tracing studies by Wouterlood and colleagues (96) established the existence of afferent connections to the histaminergic perikarya, namely from the infralimbic division of the prefrontal cortex, the septum-diagonal band complex, the medial preoptic area and the hippocampal area (subiculum).

Three histamine receptor subtypes (H1, H2 and H3) have been defined by means of functional assays and, subsequently, design of selective agonists and antagonists (25,76). All three seem to belong to the superfamily of receptors with seven transmembrane domains and coupled to guanylnucleotide-sensitive G proteins (). In addition, histamine affects the N-methyl-D-aspartate (NMDA) receptor (7,93).

The H1 receptor was initially defined in functional assays (e.g., smooth muscle contraction) and by the design of potent antagonists, the so-called "antihistamines" (e.g., mepyramine), most of which have prominent sedative properties.

Biochemical and localization studies of the H1 receptor were made feasible with the design of reversible and irreversible radiolabeled probes such as [3H] mepyramine, [125I]iodobolpyramine and [125I]iodoazidophenpyramine (22,63,69).

Initial biochemical studies indicated that the cerebral guinea pig H1 receptor was a glycoprotein of apparent molecular size of 56 kDa with critical disulfide bonds. Agonist binding was regulated by guanyl nucleotides, implying that the receptor belonged to the superfamily of receptors coupled to G proteins. In addition, various intracellular responses were found to be associated with H1-receptor stimulation: inositol phosphate release, increase in Ca2+ fluxes, cyclic AMP or cyclic GMP accumulation in whole cells and arachidonic acid release (25). It was not known, however, whether such a variety of responses corresponded to a single receptor or to distinct isoreceptors. Indeed, several photoaffinity-labeled proteins of slightly different sizes, but similar H1 pharmacology, were detected in some tissues (69).

In spite of preliminary purification attempts using affinity columns with a mepyramine derivative, the H1 receptor was never purified to homogeneity. Nevertheless, the deduced amino acid sequence of a bovine H1 receptor was disclosed after expression cloning of a corresponding cDNA. The latter was based upon the detection of a Ca2+-dependent Cl- influx into microinjected Xenopus oocytes. Following the transient expression of the cloned cDNA into COS-7 cells, the identity of the protein as an H1 receptor was confirmed by binding studies (98).

Starting from the bovine sequence, the H1 receptor DNA was also cloned in the guinea pig (26,87), a species in which the pharmacology of the receptor is better established. Although marked species differences in H1-receptor pharmacology had been reported (75), the sequence homology between the putative transmembrane domains (TMs) of the two proteins is rather high (90%).

The "anatomy" of the H1 receptor, with a long I3 (third intracellular domain) and a short C-terminal tail, is similar to that of other receptors positively coupled to phospholipases A2 and C. Amino acid sequence homology between the TMs of the H1 and of the muscarinic receptors (~45%) is higher than between those of H1 and H2 receptors (~40%). H1-receptor antagonists often display significant anti-muscarinic activity but only limited H2-receptor antagonist properties.

A single intronless gene seems to encode the guinea pig H1 receptor, and mRNAs of similar size were detected in brain areas and peripheral tissues (87). Thus the two pharmacologically indistinguishable isoforms of the H1 receptor of 56 and 68 kDa, detected after photolabeling and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in some tissues (69), may correspond either to the same protein with different degrees of glycosylation or to the product of another, similar gene revealed by Southern blot analysis (26).

When stably expressed in transfected fibroblasts, the guinea pig H1 receptor was found to trigger a large variety of intracellular signals involving or not coupling to pertussis toxin-sensitive G proteins (Gi or Go)—namely Ca2+ transients, inositol phosphates or arachidonate release (39). H1 receptor stimulation potentiates cAMP accumulation induced by forskolin in the same transfected fibroblasts, a response which resembles the H1-potentiation of histamine H2- or adenosine A2-receptor-induced accumulation of cAMP in brain slices. All these responses mediated by a single H1 receptor were known to occur in distinct cell lines or brain slices but could have been due to stimulation of isoreceptors.

The H1 receptor mediates various excitatory responses in brain (24). In addition, this receptor was recently shown to be responsible for a slow depolarization due to a decrease in a K+ current in lateral geniculate relay neurons (46).

H1-receptor distribution in the guinea pig brain was established autoradiographically using [3H]mepyramine or the more sensitive probe [125I]iodobolpyramine [63] (); the information was complemented by in situ hybridization of the mRNA (26,87). For instance, the high density of H1 receptors in the molecular layers of cerebellum and hippocampus seems to correspond, respectively, to dendrites of Purkinje and pyramidal cells, in which the mRNA is highly expressed. H1 receptors are also abundant in guinea pig thalamus, hypothalamic nuclei (e.g., ventromedial nuclei), nucleus accumbens, amygdaloid nuclei, and frontal cortex, but not in neostriatum (63), whereas they are more abundant in the human neostriatum (44).

The H1 receptor was visualized in the primate and human brain by positron emission tomography using [11C]mepyramine (99).

The widespread distribution of the H1 receptor in cerebral areas involved in wakefulness and cognition presumably accounts for the sedative properties of "antihistamines" of the first generation.

Molecular properties of the H2 receptor have remained largely unknown for a long time. For instance, reversible labeling of the H2 receptor was achieved only recently using [3H]tiotidine or, more reliably, [125I]iodoaminopotentidine (75). Irreversible labeling, using a photoaffinity probe, followed by SDS-PAGE led to the identification of H2 receptor peptides from the guinea pig (69).

By screening cDNA or genomic libraries with homologous probes, the gene encoding the H2 receptor was first identified in dogs (20) and, subsequently, in other species, including humans (19,90). The H2 receptor is organized like other receptors positively coupled to adenylyl cyclase; that is, it displays a short third intracellular loop and a long C-terminal cytoplasmic tail.

Using transfected cell lines, the well established positive linkage of the H2 receptor with adenylyl cyclase (76) and the unexpected inhibition of arachidonate release (88) and stimulation of Ca2+ transients (17)) were confirmed. Hence, H2 receptor stimulation can trigger intracellular signals either opposite or similar to those evoked by H1 receptor stimulation. Parallel observations were made for a variety of biological responses mediated by the two receptors in peripheral tissues.

Helmut Haas and colleagues showed that, in hippocampal pyramidal neurons, H2-receptor stimulation potentiates excitatory signals by decreasing a Ca2+-activated K+ conductance, presumably via cAMP production (24). H2-receptor activation depolarizes thalamic relay neurons slightly, markedly increasing apparent membrane conductance, a response due to enhancement of the hyperpolarization-activated cation current Ih (46).

The sole selective H2-receptor antagonist known to enter the brain is zolantidine, a compound used sometimes in animal behavioral studies but never introduced in therapeutics (18). However a number of tricyclic antidepressants are known to block H2-receptor-linked adenylylcyclase quite potently and to interact with [125I]iodoaminopotentidine binding in a complex manner (89).

Autoradiographic localization of H2 receptors in guinea pig, performed using [125I]iodoaminopotentidine, revealed a heterogeneous distribution in a manner suggesting their major association with neurons (63)) (). H2 receptors are found in most areas of the cerebral cortex, with the highest density in the superficial layers, the piriform and occipital cortices, both with low H1-receptor densities. The caudate putamen, ventral striatal complex and amygdaloid nuclei (bed nucleus of the stria terminalis) are among the brain areas richest in H2 receptors. In the hippocampal formation, H2 receptors display a laminated pattern with labeling of lacunosum moleculare, radiatum and oriens layers; the partial overlap with H1 receptors may account for their synergistic interaction in cAMP accumulation. H2 receptors in human brain were characterized and localized using [125I]iodoaminopotentidine (44).

The H3 receptor was initially detected as an autoreceptor controlling histamine synthesis and release in brain. Thereafter it was shown to inhibit presynaptically the release of other monoamines in brain and peripheral tissues as well as the release of neuropeptides from unmyelinated C fibers (3,72).

The molecular structure of the H3 receptor remains to be defined. Reversible labeling of this receptor was first achieved using the highly selective agonist [3H](R)a-methylhistamine (75); [3H]Na-methylhistamine, a less selective agonist, was also proposed (22), as, more recently, were the two antagonists [125I]iodophenpropit and [125I]iodoproxyfan (31,40). The binding of [3H](R)a-methylhistamine is regulated by guanyl nucleotides, strongly suggesting that the H3 receptor, like the other histamine receptors, belongs to the superfamily of receptors coupled to G proteins (75). Constitutive H3 receptors in a gastric cell line appear to be negatively coupled to phospholipase C via a mechanism sensitive to both pertussis and cholera toxins (12). In contrast, H3 receptors in vascular smooth muscle mediate voltage-dependent Ca2+-channel stimulation via a pertussis-insensitive G protein (55).

Recently, several highly potent and selective H3-receptor agonists have been designed (). Like (R)a-methylhistamine, these compounds can decrease brain histamine synthesis and release after systemic administration in low dosage (3,75). Thioperamide, a systemically active H3 antagonist, markedly increases histamine turnover in brain (3) and, because no other class of drug is available for this purpose, is widely used in behavioral studies.

Functional studies have found evidence of inhibitory H3 receptors on nerve terminals not only of histaminergic (3,75) but also noradrenergic, serotoninergic, dopaminergic (72), glutamatergic (10) and peptidergic neurons (45).

Autoradiography of H3 receptors in rat (16,64) () and monkey brain (44) shows a high concentration in neostriatum, nucleus accumbens, cingulate and infralimbic cortices, bed nucleus of stria terminalis and substantia nigra pars lateralis. In contrast, their density is relatively low in the hypothalamus (including the tuberomammillary nucleus), which contains the highest density of histaminergic axons (and perikarya). These data indicate that the majority of H3 receptors are not autoreceptors. In agreement, intrastriatal administration of kainate strongly decreased H3 binding sites in forebrain (as well as in the substantia nigra, consistent with the presence of these receptors in striatonigral neurons) (16,64).

Histamine potentiates NMDA-evoked currents in acutely dissociated and cultured hippocampal neurons, an effect that could not be ascribed to activation of the known histamine receptors (7,93). On recombinant NMDA receptors expressed in Xenopus oocytes, histamine acts directly at a novel recognition site, specifically on receptors containing the subunits NR1A/NR2B (95). It is not known whether this effect is also triggered by endogenous histamine, inasmuch as conflicting data have been reported about the effect of histamine on NMDA receptors in more physiological systems, such as hippocampal slices (8,9,101).

Cortically projecting histaminergic neurons share with other aminergic neurons a number of electrophysiological properties evidenced by extracellular recording. They fire spontaneously slowly and regularly, and their action potentials are of long duration (24). Among the pacing events which may contribute to their spontaneous firing, tuberomammillary neurons exhibit a tetrodotoxin-sensitive persistent Na+ current (91) and a Ca2+ current, probably of the low-threshold type (84)). In addition, they exhibit inward rectification attributed to an Ih current that may increase whole-cell conductance and decrease the efficacy of synaptic inputs during periods of prolonged hyperpolarization that is, when histaminergic neurons fall silent (32).

The autoreceptor-regulated modulation of histamine synthesis in, and release from, brain neurons is now well documented (75). It was initially demonstrated in brain slices or synaptosomes after labeling the endogenous histamine pool using the 3H-precursor. Exogenous histamine decreases the depolarization-induced formation and release of [3H]histamine. Analysis of these responses led to the pharmacological definition of H3 receptors. Autoregulation was found in various brain regions known to contain histamine nerve endings, which suggested that all terminals were endowed with H3 autoreceptors. Regulation of histamine synthesis was also observed in the posterior hypothalamus (3), Tuberomammillary neurons themselves are sensitive to histamine and to an H3-receptor agonist which inhibits their firing by means of hyperpolarization accompanied by an increased input resistance (24). This may indicate the existence of autoreceptors at the level of histaminergic perikarya or dendrites.

Galanin, a putative co-transmitter of a subpopulation of histaminergic neurons, regulates histamine release only in regions known to contain efferents of this subpopulation; that is, in hypothalamus and hippocampus but not in cerebral cortex or striatum (4). In brain slices, galanin also hyperpolarizes and decreases the firing rate of tuberomammillary neurons (24). It is not known, however, whether these galanin "autoreceptors" modulate galanin release from histaminergic nerve terminals. Other putative co-transmitters of histaminergic neurons failed to affect [3H]histamine release from slices of rat cerebral cortex (74).

[3H]Histamine synthesis and release are inhibited in various brain regions by stimulation of not only autoreceptors but also a2-adrenergic receptors, M1-muscarinic receptors and k-opioid receptors (75). Muscarinic receptors also inhibit endogenous histamine release in hypothalamus (59). Since these types of regulation are also observed with synaptosomes (74), all of these receptors presumably represent true presynaptic heteroreceptors. In contrast, histamine release is enhanced by stimulation of nicotinic receptors in rat hypothalamus (59) and by m-opioid receptors in mouse cerebral cortex (75).

Some molecular mechanisms regulating neuronal histamine dynamics remain unclear. No histamine transporter could be found, and a direct feedback inhibition of histidine decarboxylase by histamine has been excluded (75). Although the promoter of the histidine decarboxylase gene was analyzed (103), and the transcriptional mechanisms regulating its expression in stomach were studied (11), these processes have yet to be elucidated in neurons.

Both neurochemical and electrophysiological studies indicate that the activity of histaminergic neurons is high during arousal. In rat hypothalamus, histamine levels were low, whereas synthesis was high during the dark period, suggesting that neuronal activity was enhanced during the active phase (75). Histamine release from the anterior hypothalamus of freely moving rats, evaluated by in vivo microdialysis, gradually increased in the second half of the light period and was maintained at a maximal level during the active phase (49). Such state-related changes were also found in single-unit extracellular recordings performed in the ventrolateral posterior hypothalamus of freely moving cats. Neurons with properties consistent with those of histaminergic neurons exhibited a circadian rhythm in their firing rate, falling silent during deep slow-wave or paradoxical sleep (75). Tuberomammillary histaminergic neuron activity could be inhibited by a GABAergic pathway, originating in the ventrolateral preoptic area, that was activated during sleep (78).

A feeding-induced increase in the activity of histaminergic neurons has also been shown by microdialysis performed in the hypothalamus of conscious rats (30). Changes in the metabolism and release of histamine observed in vivo after occlusion of the middle cerebral artery in rats suggest that the histaminergic activity is also enhanced by cerebral ischemia (1).

Inhibition mediated by H3-autoreceptors constitutes a major regulatory mechanism for histaminergic neuron activity under physiological conditions. H1 and H2 receptors are apparently not involved. Administration of selective H3 receptor agonists reduces histamine turnover (75) and release, as shown by microdialysis (28). In contrast, H3-receptor antagonists enhanced histamine turnover (75) and release in vivo (29, 50) and histidine decarboxylase activity in various strains of mice (47), suggesting that these autoreceptors were under tonic stimulation by endogenous histamine.

Agents inhibiting histamine release in vitro, via stimulation of presynaptic a2-adrenergic or muscarinic heteroreceptors, reduce histamine release and turnover in vivo. However, systemic administration of antagonists of these receptors does not generally enhance histamine turnover, suggesting that heteroreceptors are not tonically activated under basal conditions.

Activation of central nicotinic (56), and 5-HT1A serotonergic (57) receptors inhibited histamine turnover and activation of D2 dopaminergic receptors enhanced histamine release in vivo (65), but the presynaptic location of these receptors remains to be demonstrated. Histamine turnover in the brain was also rapidly reduced after administration of various sedative drugs such as ethanol, D9-tetrahydrocannabinol, barbiturates and benzodiazepines (75). The effect of the latter compounds may result from their interaction in vivo with GABA receptors present on nerve endings of a subpopulation of histaminergic neurons containing GABA (81).

In contrast, stimulation of m-opioid (75) and NMDA receptors (58) enhanced histamine release and turnover in brain. Morphine increased histamine release in the periaqueductal gray (5).

The reported effects of reserpine on brain histamine turnover are inconsistent: both enhancement (75) and inhibition (53) have been reported.

Few physiological roles for histaminergic neurons have been adequately documented, although many different suggestions have been derived from the observation of responses to locally applied histamine.

Since our initial article was published in 1977 (73), a large body of experimental evidence has accumulated to indicate that histaminergic neurons play a critical role in cortical activation and arousal mechanisms (75).

Intracerebral injection of histamine in the cat ventrolateral hypothalamus, where the density of histaminergic axons is high, increased wakefulness via stimulation of postsynaptic H1 receptors. Endogenous histamine presumably plays a similar role, since inhibition of its synthesis by an L-histidine decarboxylase inhibitor, inhibition of its release by an H3-receptor agonist, or inhibition of its action by an H1-receptor antagonist all increase deep, slow-wave sleep and decrease wakefulness in several animal species. Conversely, inhibitors of histamine methylation or H3-receptor antagonists, which facilitate histamine release, both increased arousal (42,75). The role of histaminergic neurons in arousal was also shown by the decreased wakefulness following lesions of the posterior hypothalamus, particularly those lesions aimed at destruction of the tuberomammillary nucleus (71).

Finally, histaminergic neurons share a number of electrophysiological properties—including increased activity during wakefulness (see preceding section)—with other cortically projecting aminergic neurons which control behavioral states.

The "arousing effect" of histamine may well be mediated by the cellular actions of H1 and H2 receptors. Acting through these receptors on thalamic relay neurons, histamine exerts a double depolarizing action. Neuronal activity changes from endogenous oscillation and poor responsiveness to sensory inputs (which predominates during sleep), to activity dominated by single spikes and a more accurate and faithful relay of sensory information (characteristic of arousal) (83). In addition, histamine facilitates further processing of sensory information in the neocortex through the reduction of spike frequency adaptation. This effect is a consequence of either a blockade of a voltage-independent potassium conductance mediated by the H1 receptor (67) or a block of a Ca2+-activated K+ current mediated by the H2 receptor (24). These various actions seem to be shared by other neurotransmitters (including acetylcholine, noradrenaline, serotonin and glutamate) active within the forebrain. These transmitters may mediate shifts in oscillatory states of thalamocortical networks and promote the characteristic changes in firing occurring between sleep and arousal (83). Histaminergic neurons may also promote cortical activation via extrathalamic pathways. Intracellular recordings have revealed H1- and H2-receptor mediated activation of cholinergic neurons within the substantia innominata (35. Histamine might also enhance wakefulness by exerting a tonic control over the sleep-generating mechanisms of the preoptic area. In addition, descending histaminergic inputs to the mesopontine tegmentum may promote cortical electroencephalographic (EEG) desynchronization, mainly via activation of H1 receptors located on cholinergic neurons (41).

All of these cellular modes of action of histamine, taken together with observations indicating that its release from activated tuberomammillary neurons is maximal during wakefulness, suggest that histaminergic systems make important contributions to the control of arousal, attention, sensory information processing and cognition.

Accordingly, in humans, many H1 receptor antagonists induce drowsiness, impair performances requiring attention and increase the tendency to sleep—effects which are stereoselective (54). As a consequence, these compounds are common ingredients in over-the-counter sleeping pills (51). A new generation of "antihistamines", those that do not block cerebral H1 receptors, are devoid of sedative properties. A rather large number of antidepressant (e.g., mianserin or doxepin) and antipsychotic agents (e.g., clozapine) display high H1-receptor antagonist potency, a property that presumably accounts for their sedative side effects (68!popup(ch37ref68)). In mutant mice lacking H1 receptors, complex behavioral changes were observed; the expected impairment of locomotor activity in the dark period was accompanied by increased locomotion in the light period (27).

Perhaps as a consequence of the role played by histaminergic neurons in arousal, administration of H3 antagonists, which induce histamine release, enhances the learning ability of rodents in a variety of paradigms, particularly when this ability is impaired by scopolamine or accelerated senescence (47,48).

Exogenous and, in some cases, endogenous histamine were shown to affect the secretion of both posterior and anterior pituitary hormones (38,75).

Supraoptic nucleus neurons are typically excited by application of histamine. Histamine induced c-fos expression within these neurons (92); their firing rate was increased during secretory bursts of activity, and the depolarizing afterpotential was enhanced (79). This effect, mediated by H1 receptors, caused circulating vasopressin levels to rise. The evidence for histaminergic neuronal participation in the physiological control of vasopressin secretion includes the following observations: inhibition of histamine synthesis impaired the vasopressin response to adrenalectomy, and stimulation of the tuberomammillary nucleus caused phasic neurons in the supraoptic nucleus to become more excitable (100). In addition, dehydration, a potent stimulus of vasopressin secretion, increases the histidine decarboxylase mRNA level in the tuberomammillary nucleus (36). This control, however, may not be a tonic one under basal conditions, because H1-receptor antagonists do not appear to modify vasopressin secretion. It is not known whether the increase in water consumption elicited by an H3-receptor agonist is related to an effect on supraoptic nucleus neurons (13).

Endogenous histamine may be involved in stress-, estrogen- or morphine-induced release of prolactin (38); the response is prevented by blockade of histamine synthesis or postsynaptic H1 or H2 receptors, or by activation of presynaptic H3 autoreceptors. The action of histamine may involve vasopressin, as it is prevented by antibodies to that neuropeptide. Endogenous histamine, which acts primarily at H1 receptors, may also be involved in the secretory responses of adrenocorticotropic hormone and b-endorphin induced by restraint, insulin hypoglycemia or endotoxin (37,80).

Although exogenous histamine predominantly inhibits the release of growth hormone and thyroid-stimulating hormone, the role of histaminergic neurons in the control of secretion of these hormones has not been established.

Weight gain is often experienced by patients receiving H1 antihistamines or tricyclic antidepressants that have potent H1-receptor antagonist properties. This may reflect an inhibition of feeding exerted by histamine neurons that project to the ventromedial and paraventricular hypothalamic nuclei, as shown by the effects of histamine synthesis inhibitors or H3-receptor ligands (70). In addition, the extracellular concentration of the amine in rat hypothalamus increases during feeding (30)).

The action of drugs affecting histamine synthesis or methylation, as well as of H1- and H3-receptor antagonists, suggests that endogenous histamine may restrict the manifestations of electrically- and pentylenetetrazole-induced seizures in rodents (102).

Some H1 antagonists are the most commonly used antimotion-sickness drugs, but it is not clear whether this effect is related to blockade of H1 receptors. Histamine depolarizes neurons in vestibular nuclei and modulates static vestibular reflexes quite effectively through interaction with both H2 and H3 receptors (77,97). A beneficial effect of H3-receptor antagonists in vertigo or motion sickness was suggested by these data.

Among the various approaches that tend to establish the implication of other aminergic neuronal systems in neuropsychiatric diseases, so far only a few have been (or could be) applied to histamine. However, elevated levels of t-methylhistamine in cerebrospinal fluid suggest increased central histaminergic activity in patients with chronic schizophrenia (66).

Post mortem studies of basal ganglia from patients with Parkinson's disease (75) or in a rodent model of this disease (14) showed no change in the activity of the histamine-synthesizing enzyme.

Patients with Alzheimer's disease show numerous neurofibrillary tangles and typical senile plaques in the tuberomammillary area. It is not clear, however, whether the number of histamine-immunoreactive neurons is decreased (2). Because conflicting data have been reported concerning histamine levels in such patients, a role of histamine in the etiology of Alzheimer's disease remains doubtful. In addition, it may be significant that 9-amino-1,2,3,4-tetrahydroacridine (THA), an anticholinesterase that was found to be useful in Alzheimer's disease, is also a rather potent inhibitor of histamine methylation (15,52). In addition, an H3-receptor antagonist improved learning deficits in senescence-accelerated mice (47).

The effects of antipsychotics at dopamine receptors strongly suggested the role of dopamine in schizophrenia. In contrast, the interactions of psychotropic drugs with histamine receptors are of limited help for deducing the role of histaminergic neurons in psychiatric illnesses. Over a decade ago, J.P. Green, P. Greengard and colleagues (23,33) proposed that the cerebral H2 receptor was an important target for most tricyclic and other antidepressant drugs that interact with relatively high affinity with the receptor coupled to the cyclase (68). The strong dependence of the apparent affinity of these compounds for the H2 receptor under the experimental conditions of the assay leaves some doubt, however, about the therapeutic significance of this observation (89). A number of side effects (e.g., sedation or weight gain) of several antidepressant drugs, as well as some neuroleptics, are attributable to the blockade of cerebral H1 receptors (68). The affinity of the atypical antipsychotic drug clozapine at the H3 receptor is in the same range as at D2/D3 dopamine receptors (34), but the therapeutic implications of this observation—if any—are unclear.

Unfortunately, the effects of drugs able to stimulate the three cerebral histamine receptor subtypes or to block H2 or H3 receptors in neuropsychiatric diseases are not known. Therefore, these receptors remain important potential targets for novel classes of psychotropic agents, particularly "cognition/arousal enhancers" acting via facilitation of histaminergic neurotransmission in brain.

published 2000