Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

HA is a cerebral neurotransmitter exerting its actions on target cells via three classes of molecularly and/or pharmacologically well defined receptors designated H₁, H₂ and H₃ (reviewed by Hill et al., 1997; Schwartz et al., 1991, 1995). The H₃ receptor is a presynaptic receptor regulating the synthesis and/or release of HA itself (Arrang et al., 1983) as well as a variety of other aminergic or peptidergic neurotransmitters (reviewed by Schlicker et al., 1994). It was initially defined by the design of two selective ligands: (R)-MeHA, a full agonist, and thioperamide, an antagonist with nanomolar potency (K_i  4 nM). Thioperamide has become the prototypical H₃-receptor antagonist, used in a large number of neurochemical, electrophysiological and behavioral studies because it is one of the few agents able to markedly enhance cerebral histaminergic transmissions in vivo via a selective mechanism. In agreement, few other actions of thioperamide were described, e.g., inhibition of P450 cytochromes (La Bella et al., 1992) and 5-HT₃-receptor blockade (Leurs et al., 1995), which require higher drug concentrations than H₃-receptor blockade and are therefore not relevant for in vivo studies.

Nevertheless thioperamide has several drawbacks: 1) its in vivo potency is rather low compared with its in vitro potency, suggesting that drug bioavailability, particularly its brain penetration, is restricted, 2) more importantly it displays a distinct liver toxicity on repeated administration which has prevented it being submitted to human clinical trials.

Because H₃-receptor antagonists represent a novel class of agents with potentially interesting therapeutic applications, namely in psychiatry (Schwartz et al., 1995) sustained efforts have been devoted to the design of drugs more potent and safer than thioperamide (reviewed by Stark et al., 1996b).

As with thioperamide, all highly effective compounds obtained so far contain a monosubstituted imidazole ring, but the thiourea moiety of the latter, to which hepatotoxicity might be attributable, is replaced by numerous polar functionalities such as amine or carbamate, etc.

Recently, we have designed 3-(1H-imidazol-4-yl)propanol derivatives as a novel series of potent in vitro H₃-receptor antagonists with a high capacity for oral absorption and brain penetration in some compounds (Stark et al., 1996a; Hüls et al., 1996). In these novel chemical classes of compounds, we have recently identified [¹²⁵I]iodoproxyfan, i.e., [¹²⁵I]3-(1H-imidazol-4-yl)propyl-(4-iodophenyl)methyl ether as a new probe for a sensitive assay and localization of the H₃ receptor in brain (Ligneau et al., 1994).

We describe the biological properties of ciproxifan (fig. 1), a highly potent and selective H₃-receptor antagonist belonging to this novel chemical class of compounds which suggest its potential therapeutic interest as a waking and procognitive agent.

View larger version (6K): [in this window] [in a new window]   Fig. 1.   Chemical structure of ciproxifan, i.e., cyclopropyl-(4-(3-(1H-imidazol-4-yl)propyloxy)phenyl) ketone.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

[³H]Histamine release from synaptosomes. [³H]HA release experiments were performed according to Garbarg et al. (1992). Briefly, a crude synaptosomal preparation from rat cerebral cortex was preincubated for 30 min with [³H]L-histidine (0.4 µM) at 37°C. After extensive washing synaptosomes were resuspended in fresh 2 mM K⁺-Krebs-Ringer's medium and in the presence of the appropriate drugs. After 5-min incubation synaptosomes were depolarized bringing the K⁺-concentration to 30 mM for 2 min. Incubations were ended by a rapid centrifugation and [³H]HA levels in the supernatant were determined after an ion-exchange chromatography purification. Release was expressed as the percent fraction of total [³H]HA initially present in the synaptosomal preparation. Typically total [³H]HA represented about 3,500 dpm/mg protein and total radioactivity about 100,000 dpm/mg protein in the test tube.

Assay of t-MeHA in brain. Male Swiss mice (18-20 g) or male Wistar rats (140-160 g) (Iffa-Credo, L'Arbresle, France) were fasted for 24 hr before p.o. administration. After treatment animals were killed by decapitation, the brain was dissected out and homogenized in 10 volumes (w/v) of ice-cold perchloric acid (0.4 N). The clear supernatant obtained after centrifugation (2000 × g, 30 min, + 4°C) was stored at 20°C before measuring the t-MeHA level by radioimmunoassay as described (Garbarg et al., 1992). Changes were evaluated statistically by the Student's t test.

[¹²⁵I]Iodoproxyfan binding assays. The procedure for binding assays to rat striatal brain membranes was that described by Ligneau et al. (1994). Aliquots of membrane suspension [100 µl containing 15 to 20 µg of protein determined according to Lowry et al. (1951) using bovine serum albumin as standard] were incubated for 60 min at 25°C with 25 pM [¹²⁵I]iodoproxyfan (K_d = 65 ± 4 pM) alone or together with competing drugs dissolved to give a final volume of 200 µl in a phosphate buffer medium (Na₂HPO₄/KH₂PO₄ 50 mM, pH 6.8). Incubations performed in triplicate were stopped by four additions of 5 ml ice-cold medium followed by rapid filtration through glass microfiber filters (GF/B, Whatman, Maidstone, U.K.) presoaked in a 0.3% polyethylene imine ice-cold buffer. Radioactivity trapped on filters was measured on a gamma counter.

Histamine H₁ receptor assay on guinea pig ileum. The procedure used was that described by Pertz and Elz (1995).

Histamine H₂ receptor assay on guinea pig right atrium. The procedure used was that described by Pertz and Elz (1995).

Muscarinic M₃ receptor assay on guinea pig ileum. The procedure used was that described by Pertz and Elz (1995).

Adrenergic _1D receptor assay on rat aorta. The procedure used was that described by Hirschfeld et al. (1992).

Adrenergic ₁ receptor assay on guinea pig right atrium. The procedure used was that described by Pertz and Elz (1995).

Serotoninergic 5-HT_1B receptor assay on guinea pig iliac artery. The procedure used was that described by Pertz (1993).

Serotoninergic 5-HT_2A receptor assay on rat tail artery. The procedure used was that described by Pertz and Elz (1995).

Serotoninergic 5-HT₃ receptor assay on guinea pig ileum. The procedure used was that described by Elz and Keller (1995).

Serotoninergic 5-HT₄ receptor assay on rat esophagus. The procedure used was that described by Elz and Keller (1995).

Histamine H₃ receptor assay on guinea pig ileum. The procedure used was that described by Ligneau et al. (1994). Briefly, longitudinal muscle strips from guinea pig small intestine were dissected out and incubated in a gassed O₂/CO₂ (95%/5%) modified Krebs-Ringer's bicarbonate medium at +37°C in presence of 1 µM mepyramine to block the H₁ receptor. After equilibration, contractile activity under stimulation (rectangular pulses of 15 V, 0.5 msec, 0.1 Hz) was recorded. Concentration-response curves of the effect of (R)-MeHA alone or together with the antagonist were established.

Radioreceptor assay of H₃-receptor ligands in serum. Male Swiss mice (18-20 g, Iffa-Credo, L'Arbresle, France) were fasted for 24 hr before ciproxifan administration. At various times thereafter, mice were decapitated. Blood was collected at +4°C, serum collected after centrifugation (100 × g, 10 min, +4°C), and stored at 20°C. Ciproxifan levels were measured with the following radioreceptor assay derived from the [³H](R)-MeHA binding assay (Garbarg et al., 1992).

(R)-MeHA-induced water consumption in rats. The procedure used was that described by Clapham and Kilpatrick (1993) with slight modifications. Briefly, male Lister hooded rats (280-320 g, Charles River, St Aubin lès Elbeuf, France), housed in cages of 10, were allowed free access to food and water. Experiments were carried out between 10 A.M. and 1 P.M. Drugs in 0.9% NaCl solution were administered i.p., each rat receiving two injections [vehicle or (R)-MeHA 10 mg/kg/p.o., and vehicle or ciproxifan] of 250 µl each. After drug treatments, rats were returned to their home cage with food (standard pellet) and without water for 30 min. Then they were placed in individual cages with only a water supply and 10 min later, the amount of water consumed was recorded by weighing. Statistical evaluation of results was performed by Student's t test.

Analysis of neocortical EEG power spectral density and sleep-wake in the cat. Cats, a species in which the role of the H₃ receptor in sleep-wake control has been demonstrated, were used in this experiment (Lin et al., 1990). Briefly, five adult cats of both sexes weighing 2.7 to 3.8 kg were chronically implanted, under pentobarbital anesthesia (25 mg/kg, i.v.), with electrodes for polygraphic recordings of neocortical and hippocampal EEG, ponto-geniculooccipital activity, electromyogram and electrooculogram. In addition, a thermistor (10 K3 MCD2, Betatherm, 10 K at 25°C, outer diameter of 0.45 mm) was placed in the caudate nucleus to record brain temperature. After a recovery period of 7 days the cats were housed in a sound-attenuated and dimly illuminated cage set at 24 to 26°C and fed daily at 6 P.M. with a normal standard diet. Polygraphic recordings were performed for 4 days to obtain the basic qualitative and quantitative parameters of the sleep-wake cycle.

Five-choice task in rats. Male Lister hooded rats (Olac, Bicester, U.K.) were housed in pairs in a temperature-controlled (21°C) room that was illuminated in accordance with an alternating 12-hr light/dark cycle. Rats were food deprived and maintained at 85% of their free-feeding weight (MRC Diet 41B laboratory food) throughout the experiment while water was available ad libitum.

Analysis of data. Maximal effects, ED₅₀, EC₅₀ and IC₅₀ values were determined using an iterative computer least-squares method derived from that of Parker and Waud (1971) with the following nonlinear regression: effect of the drug = ([maximal effect of the drug].[drug dose-or-concentration])/([drug dose-or-concentration] + (ED₅₀ or EC₅₀ or IC₅₀)).

Radiochemicals and drugs. [¹²⁵I]Iodoproxyfan and [³H](R)-MeHA (specific activities at reference date of 2000 and 38.0 Ci/mmol, respectively) were from Amersham (Amersham, U.K.). All drug doses are expressed as free base of compound. Administration to animals was performed with drug preparation in 1% methylcellulose for the oral route, in 0.9% NaCl for i.v. and i.p. routes. The drugs and their sources were as follows: ciproxifan [cyclopropyl-(4-(3-(1H-imidazol-4-yl)propyloxy)phenyl) ketone] synthesis was described in the patent application (Schwartz et al., 1996); thioperamide and (R)-MeHA (Laboratoire Bioprojet, Paris, France); carboperamide [1-(heptanoyl)-4-(1H-imidazol-4-yl)piperidine] was from M. Robba (University of Caen, France); clobenpropit was provided by H. Timmerman (Vrije Universiteit, Amsterdam, The Netherlands); iodoproxyfan was synthesized at the Freie Universität Berlin; imetit was synthesized by S. Athmani (University College, London, U.K.). All other chemicals were obtained from commercial sources and were of the highest purity available.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Effects of ciproxifan on H₃-receptor functional models in vitro. The H₃-receptor agonist imetit inhibited the [³H]HA release with a maximal effect of 54 ± 2%. Pharmacological parameters were estimated using the H₃-receptor controlled inhibition curve of the [³H]HA release. Imetit concentration required for half inhibitory effect of its own maximal effect was 1.6 ± 0.3 nM and the corresponding pseudo-Hill coefficient (n_H) of this concentration response curve was close to unity (0.85 ± 0.15). Ciproxifan (15 nM) induced a parallel rightward shift of the concentration-response curve for imetit and tended by itself to increase (~20%) the K⁺-induced [³H]HA release; the antagonism developed by ciproxifan was entirely surmounted in the presence of the highest imetit concentrations tested (fig. 2). The half-maximal inhibitory concentration for imetit in the presence of ciproxifan (estimated considering the total H₃-receptor controlled inhibition curve) was 12.8 ± 1.8 nM, leading to an apparent K_i value of 1.9 ± 0.3 nM for the antagonist.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Effect of ciproxifan on the inhibition by imetit of the K+-induced [3H]HA release from synaptosomes of rat cerebral cortex. Synaptosomes preincubated with [3H]L-histidine were incubated for 5 min in the presence of imetit in increasing concentrations, alone or together with ciproxifan at a fixed (15 nM) concentration. They were subsequently depolarized for 2 min in the presence of 30 mM K+ (final concentration). The spontaneous efflux of [3H]HA (2 mM K+) represented 14 ± 2% of the total [3H]HA in synaptosomes. In the presence of added agents, [3H]HA release induced by 30 mM K+ (expressed in percent of total [3H]HA over spontaneous efflux) represented 17 ± 2%. Each point represents the mean ± S.E.M. of results from three different experiments with quadruplicate determinations each.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Effects of H3-receptor antagonists on the inhibition by HA of the K+-induced [3H]HA release from rat cerebral cortex synaptosomes. Synaptosomes preincubated with 1 µM HA alone or together with one of the compounds at increasing concentrations. They were depolarized for 2 min in the presence of 30 mM K+. Each point represents the mean result from three different experiments with quadruplicate determinations each.

Effect of ciproxifan and other H₃-receptor antagonists on [¹²⁵I]iodoproxyfan binding. At 25 pM [¹²⁵I]iodoproxyfan the specific binding to rat striatal membranes represented 8.4 ± 0.4 fmol/mg of protein, i.e., 46 ± 1% of the total, and was completely and monophasically displaced by all compounds (fig. 4). From these displacement curves IC₅₀ values for the compounds were deduced leading to the K_i values presented in table 1.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Inhibition of [125I]iodoproxyfan binding to rat striatal membranes by various histaminergic agents. Membranes were incubated for 60 min at 25°C with 25 pM [125I]iodoproxyfan and unlabeled H3-receptor antagonists in increasing concentrations. Specific binding, defined as that inhibited by 1 µM (R)-MeHA, represented 46 ± 1% of the total binding. Results are expressed as percentages of specific [125I]iodoproxyfan binding in the absence of unlabeled agents. Each point and vertical bars represent the mean ± S.E.M. of results from three different experiments with triplicate determinations each.

Receptor selectivity of ciproxifan. The compound displayed low apparent affinity at other receptor subtypes as evaluated in functional tests on isolated organs (histamine H₁ and H₂, muscarinic M₃, adrenergic _1D and ₁, serotonin 5-HT_1B, 5-HT_2A, 5-HT₃ and 5-HT₄) (fig. 5). In addition, ciproxifan displayed a K_i value higher than 1 µM in a large variety of radioligand binding tests (Panlabs screen), except at [³H]pirenzepine binding to rat cerebral cortex membranes where its K_i was about 1 µM (data not shown).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Receptor selectivity profile of ciproxifan. The affinity of the compound at the H3 receptor (pKi from [3H]HA release assay) is compared to corresponding values obtained in functional tests in isolated organs. Other values represent mean ± S.E.M. pA2 (Arunlakshana and Schild, 1959) from n = 3 to 12 preparations except for the H2- and 1-receptor assay (pD'2 according to Van Rossum, 1963).

Changes in serum drug concentration in animals treated with ciproxifan. After i.v. administration of 1 mg/kg ciproxifan to mice (fig. 6B), the H₃-receptor ligand concentration in serum decreased progressively, fitting a typical biexponential decay model with half-times (t_1/2) of 13 and 87 min for the distribution and elimination phases, respectively. The quality of this fit is given by an R² value of 0.985. At 6 hr, the serum ligand concentration was still detectable with a value of 23 ± 6 nM. When ciproxifan was given orally, also at 1 mg/kg (fig. 5A), serum ligand level rose rapidly, being maximal at 30 min with a maximal concentration (C_max) value of 420 ± 40 nM; then, the ligand concentration decreased but still remained measurable at 6 hr (27 ± 5 nM). The AUCs were 1425 and 890 nM.hr after i.v. and p.o. administrations, respectively, leading to an oral bioavailability coefficient (AUCp.o./AUCi.v..100%) of 62%.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Serum drug concentration in mice receiving ciproxifan. Mice were killed after p.o. (A) or i.v. (B) administration of ciproxifan (1 mg/kg), and drug concentrations were evaluated in serum by a radioreceptor assay. Means ± S.E.M. of values from six mice.

Changes in brain t-MeHA level after administration of ciproxifan or other H₃-receptor antagonists. After the administration of ciproxifan (1 mg/kg, p.o.), brain t-MeHA level rose rapidly, being already significantly increased after 30 min, reaching a plateau between 90 and 180 min and still remaining enhanced after 270 min (fig. 7). In mice receiving pargyline, a monoamine oxidase inhibitor, t-MeHA level increased linearly with time at a rate of 55 ng/g/hr, whereas coadministration of pargyline and ciproxifan enhanced this rate to 120 ng/g/hr.

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Changes in brain t-MeHA levels in control and pargyline-treated mice after administration of ciproxifan. Mice were killed at various times after simultaneous administration of vehicle or ciproxifan (1 mg/kg, p.o.) and vehicle or pargyline (65 mg/kg, i.p.). t-MeHA levels are expressed in ng/g of tissue. Means ± S.E.M. of values from 12 mice.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Changes in brain t-MeHA levels in mice receiving H3-receptor antagonists. Mice were killed 90 min after the p.o. administration of vehicle or drugs in increasing doses. t-MeHA levels in treated mice are expressed in percent increase as compared to levels in control mice (133 ± 7 ng/g). Means ± S.E.M. of values from 12 mice.

View larger version (19K): [in this window] [in a new window]   Fig. 9.   Changes in brain t-MeHA levels in mice receiving ciproxifan or thioperamide. Mice were killed at various times after the p.o. administration of vehicle, ciproxifan (0.3 mg/kg) or thioperamide (3 mg/kg). t-MeHA levels are expressed in percent increase as compared to levels in corresponding controls. Means ± S.E.M. of values from 18 mice.

Effect of ciproxifan on the water consumption induced by an H₃-receptor agonist. The H₃-receptor agonist (R)-MeHA (10 mg/kg, i.p.) markedly enhanced water consumption in rats, an effect that was progressively reversed by coadministration of ciproxifan in increasing dosage, the ID₅₀ of the antagonist being 0.09 ± 0.04 mg/kg (fig. 10). Ciproxifan alone (3 mg/kg) did not significantly modify water consumption (fig. 10).

View larger version (23K): [in this window] [in a new window]   Fig. 10.   Effect of ciproxifan on the (R)-MeHA-induced water consumption in rats. Water consumption was measured over a 10-min period, 30 to 40 min after i.p. administrations of (R)-MeHA and ciproxifan. Means ± S.E.M. of values from 5 to 10 rats per treatment group (data from two experiments). *P < .05 vs. vehicle.

Effect of ciproxifan in the five-choice task in rats. Analysis of variance revealed a significant drug × stimulus duration interaction [F(1,9) = 12.19, P < .01]. As shown in figure 11, reducing the duration of the visual stimulus to 0.25 sec resulted in a significant reduction in the accuracy of performance compared to the baseline (0.5 sec) stimulus condition. Newman Keuls post hoc comparisons revealed that this reduction in performance was significant and that choice accuracy significantly increased after administration of 3.0 mg/kg of ciproxifan under the shorter stimulus condition compared to performance after administration of the vehicle (P < .05). There was no significant effect of this manipulation of the stimulus duration or of ciproxifan on any of the other measures recorded (i.e., speed, anticipatory or perseverative responses and errors of omission).

View larger version (21K): [in this window] [in a new window]   Fig. 11.   Effects of ciproxifan in the five-choice task in rats. The drug was administered (3 mg/kg, i.p.) 1 hr before the test session. The accuracy of responding, expressed as the percentage of correct responses, was significantly (*P < .05) improved by treatment only when the stimulus duration was 0.25 sec instead of 0.50 sec.

Effects of ciproxifan on neocortical EEG power spectral density and sleep-wake cycle in cat. Administration of ciproxifan caused suppression or diminution (depending on the dose used) of neocortical slow activity (0.8-5 Hz) and spindles (8-15 Hz), resulting in a total cortical activation, i.e., low voltage electrical activity with dominant waves in the  and  bands (mainly 25-45 Hz). Furthermore, ciproxifan increased the power density of these neocortical fast rhythms (fig. 12). These effects, occurring within 25 min after administration (fig. 12) were detectable at a dose of 0.15 mg/kg and became evident at a dose of 0.3 mg/kg or more (fig. 13).

View larger version (37K): [in this window] [in a new window]   Fig. 12.   An example of the effects of oral administration of ciproxifan on cortical EEG and sleep-wake cycle in cat. Neocortical EEG power density (mV2) in different frequency bands and sleep-wake cycle 1 hr before and up to 5 hr after administration of 2 mg/kg ciproxifan (arrow) are given. Note, from bottom to top, that ciproxifan elicited a total waking state accompanied by a suppression of cortical slow frequency activity (0.8-5 Hz) and spindle (8-15 Hz) and a marked increase in fast rhythms (25-45 Hz). Abscissa, Time in hours; ordinate, sleep-wake stages (PS, paradoxical sleep; S2, deep slow wave sleep; S1, light slow wave sleep; W, wakefulness).

View larger version (19K): [in this window] [in a new window]   Fig. 13.   Representative hypnograms (4 hr) obtained in cats after oral administration of ciproxifan at different doses. Note the dose-dependent waking effect. Abscissa, Time in hours; ordinate, sleep-wake stages (PS, paradoxical sleep; S2, deep slow wave sleep; S1, light slow wave sleep; W, wakefulness).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

From our present in vitro and in vivo studies, ciproxifan appears as a pure competitive antagonist at the histamine H₃ receptor, one of the most potent so far available.

Our selection of the molecule was initially based on its ability to block the actions of histamine or imetit, a selective H₃-receptor agonist, at the autoreceptor regulating the release of neosynthesized [³H]HA from K⁺-depolarized synaptosomes, according to a previously described model (Arrang et al., 1983; Garbarg et al., 1989, 1992). On this model, ciproxifan induced a parallel rightward shift of the concentration-response curve to imetit, indicative of a competitive antagonism with an apparent dissociation constant in the low nanomolar range. Ciproxifan in increasing concentration also progressively blocked the (R)-MeHA-induced relaxation of the electrically stimulated longitudinal muscle of guinea pig ileum in a competitive fashion. At these various functional models as well as the [¹²⁵I]iodoproxyfan binding tests, the drug displayed similarly low apparent dissociation constants (K_i = 0.5-4.2 nM), an observation that does not support the hypothesis of the existence of H₃-receptor subtypes (Clapham and Kilpatrick, 1992). The hypothesis, mainly based on discrepancies in potency of some compounds in binding and functional models, is not supported either by the closely similar potencies of other antagonists in different models, examplified in table 1 (with the exception of iodoproxyfan for which the discrepancy can be fully explained by a slow equilibration rate).

The high degree of selectivity of ciproxifan toward the histamine H₃ receptor was shown by the observation that the drug displayed lower affinity by about three orders of magnitude for any other receptor subtype on which it was tested (see "Results").

Although some compounds in vitro were as potent as or even more potent than ciproxifan at the H₃ receptor, e.g., clobenpropit and iodoproxyfan, ciproxifan given orally to mice enhanced brain t-MeHA levels at much lower dosage (ED₅₀ = 0.14 mg/kg) than any of these compounds. A similar change, occurring with ED₅₀ of 0.2-0.3 mg/kg p.o., was found in various areas of rat brain. t-MeHA is the product of the major metabolic pathway for endogenous HA in brain (Schwartz et al., 1971), and its steady-state level is a reliable index of histaminergic neuron activity (Oishi et al., 1983). The increase in t-MeHA level induced by the H₃-receptor antagonists corresponds to an enhanced HA release, reflecting the tonic inhibition that the endogenous amine exerts on this process and on neuronal firing via stimulation of autoreceptors in the somatodendritic or terminal area of histaminergic neurons. In agreement, ciproxifan (1 mg/kg, p.o.) enhanced HA turnover rate in mouse brain, evaluated from the rate of t-MeHA accumulation after monoamine oxidase inhibition, from a value of 55 ng/g/hr, consistent with corresponding values obtained in the same species using either isotopic (Verdière et al., 1977) or nonisotopic methods (Oishi et al., 1989), to a value of 120 ng/g/hr.

A similar maximal effect, corresponding to a nearly doubling of t-MeHA level, was obtained with thioperamide or carboperamide whereas, at the maximal dose tested of clobenpropit (30 mg/kg), this level was not reached and no significant change occurred after administration of iodoproxyfan (fig. 8) despite the high potency of these compounds in vitro.

In the case of ciproxifan, a rather high oral bioavailability was evidenced by the ratio (>60%) of AUCs of H₃ receptor binding activity in blood serum, measured by using a novel radioreceptor assay, following drug administration (1 mg/kg) by p.o. and i.v. routes, respectively. The rather slow kinetics of ciproxifan are indicated by a serum level still about 10 times above the K_i value of the drug at the H₃ receptor 6 hr after oral administration. At this time, t-MeHA levels in brain are still enhanced by 24 ± 11% (fig. 9). Such comparison between drug levels in blood and a typical brain response in rodents might be useful to predict effective dosages in other species, particularly humans, in which only blood levels are available, assuming a similar ability of the drug to cross the blood-brain barrier. A similarly favorable bioavailability of ciproxifan on oral administration to rats and cats is suggested by measurements of t-MeHA and drug serum levels, respectively (see "Results").

Characteristic behavioral responses were found in rats and cats receiving ciproxifan in low dosage. In water-deprived rats, ciproxifan blocked the enhancement of drinking elicited by (R)-MeHA, an H₃-receptor agonist (Clapham and Kilpatrick, 1993), with an ID₅₀ value of ~0.1 mg/kg. The exact site (central or peripheral) and mechanism of action of H₃-receptor ligands in this test has not been clearly established. Thus, whereas the involvement of the renin-angiotensin system in HA-induced drinking was postulated by Kraly and Miller (1982), an AT₁ antagonist did not affect the (R)-MeHA-induced drinking (Clapham and Kilpatrick, 1993). The observation that (R)-MeHA-induced drinking is blocked at doses of ciproxifan (this study), thioperamide and particularly clobenpropit (Barnes et al., 1993), close to those enhancing endogenous HA release in brain (table 1), suggests that H₃ receptors in brain rather than in periphery are involved. The observation that ciproxifan or thioperamide given alone do not affect drinking suggests that the effect of (R)-MeHA is mediated by H₃ hetero- rather than autoreceptors. In agreement, H₃ receptors on noradrenergic, serotoninergic, cholinergic, dopaminergic or peptidergic neurons do not appear to be, as with those on histaminergic neurons, tonically modulated by endogenous HA, because they respond to agonists but not to antagonists given alone (Schwartz et al., 1991, 1995; Schlicker et al., 1994).

The marked dose-dependent waking effect of ciproxifan in cats is consistent with a large variety of experimental evidence showing that histaminergic neurons play a prominent role in cortical activation and arousal in cats and rats (reviewed by Lin et al., 1996; Schwartz et al., 1991, 1995). The arousing effects of H₃-receptor antagonists, characterized by an enhancement of wakefulness at the expense of slow wave and paradoxical sleep, was previously shown in both animal species using thioperamide (Lin et al., 1990; Monti et al., 1991) and carboperamide (Monti et al., 1996). The effect of ciproxifan was, as in the case of thioperamide (Lin et al., 1990), prevented by administration of mepyramine, an H₁-receptor antagonist, suggesting that it resulted from a H₃-receptor mediated enhancement of endogenous HA release. The brain site(s) at which endogenous HA promote(s) cortical EEG desynchronization via activation of the H₁ receptor could be one of the brain areas to which ascending or descending histaminergic pathways project known to express the H₁ receptor and to control sleep/wakefulness states. These potential targets comprise cortical neurons receiving direct histaminergic projections from the tuberomammillary nucleus, preoptic anterior hypothalamic neurons, thalamic relay neurons and basal forebrain or mesopontine tegmentum neurons (Lin et al., 1996).

Because the effect of ciproxifan in cats was to enhance fast cortical rhythms, known to occur during increased vigilance, and to cause a quiet waking state, a positive outcome in attentional tests could be anticipated. In confirmation the drug significantly enhanced choice accuracy in the five-choice serial reaction-time task when a visual stimulus of short duration (0.25 sec) was used. Such reduction of the stimulus duration increases the attentional load placed on the task, reduces choice accuracy and has been used to observe the effects of cholinergic agents on attentional function (Muir et al., 1994, 1995). "Pro-cognitive" effects of the H₃-receptor antagonist thioperamide have been reported in other behavioural tasks, e.g., step-through passive avoidance response in senescence-accelerated mice (Meguro et al., 1995); elevated plus-maze performance in mice with scopolamine-induced learning deficits (Miyazaki et al., 1995) and in a test of social memory in rats (Prast et al., 1996). However, it has also been reported that thioperamide failed to improve scopolamine-induced attentional dysfunction in the same 5-choice task used in the present study (Kirkby et al., 1996).

Taken together these various observations suggest that ciproxifan is a potent, orally active H₃-receptor antagonist and it seems of interest to assess its potential therapeutic applications, namely in aging or degenerative disorders in which vigilance, attention and memory are impaired.