Introduction

Top Abstract Introduction Methods Results Discussion References

The HA H₃ receptor mediates presynaptic inhibition of neurotransmitter release on a variety of neuronal systems in the central and peripheral nervous system (Schwartz et al., 1990a, b, 1991). Initially characterized as an autoreceptor controlling HA synthesis in and release from endings of tuberomammillary neurons (Arrang et al., 1983, 1987) it turned out afterward to mediate heterosynaptic inhibition of release of several aminergic transmitters in brain (Clapham and Kilpatrick, 1992; Schlicker et al., 1988, 1989, 1993).

In addition, H₃ receptor stimulation was shown to inhibit vagal cholinergic transmission in the ileum (Trzeciakowski, 1987; Hew et al., 1990) and the airways (Ichinose and Barnes, 1989a; Ichinose et al., 1989) and to reduce plasma protein extravasation induced in the airways (Ichinose et al., 1990; Barnes, 1992) or meninges (Matsubara et al., 1992) by sensory C-fibers stimulation either electrically or by capsaicin. Again the latter effects result from a presynaptic inhibition, in this case of substance P antidromic release (Ichinose and Barnes, 1989b; Ichinose et al., 1990; Matsubara et al., 1992). In addition, H₃ receptor-mediated inhibitions of gastric acid secretion induced either by gastrin or vagal stimulations (Bado et al., 1991; Coruzzi et al., 1991; Soldani et al., 1993, 1996) and inhibition of mast-cell activity (Dimitriadou et al., 1994, 1997) were identified.

Selective stimulation of the H₃ receptor, which has been instrumental in the discovery of these various responses, was made possible by the identification of (R)-MeHA, a prototypic full agonist, displaying 15-fold higher potency than HA at the H₃ receptor and negligible potency at other receptor subtypes (Arrang et al., 1987). Thereafter, other potent and selective H₃ receptor agonists, i.e., (R, S), -dimethylhistamine (Lipp et al., 1992), imetit (Garbarg et al., 1992; Howson et al., 1992; van der Goot et al., 1992), imepip (Vollinga et al., 1994) and immepyr (Shih et al., 1995) were designed.

Taking into account some of these actions, it was considered that H₃ receptor agonists display a potential therapeutic value in several fields. (R)-MeHA was the first agent to be introduced in phase I clinical trials during which, however, its low plasma level after oral administration, unexpected from animal studies, was discovered (A. Rouleau, M. Garbarg, J.-M. Lecomte, and J.-C. Schwartz, unpublished observation).

In the present set of studies we have identified important methylation of (R)-MeHA as the cause for this important drawback. Therefore, we designed a series of azomethine prodrugs of (R)-MeHA (Krause et al., 1995) able to release slowly the bioactive compound, thus protecting its early metabolic degradation. Among these prodrugs, we have selected BP 2-94, i.e., (R)-(-)-2-[[N-[1-(1H-imidazol-4-yl)-2-propyl]imino]phenylmethyl] phenol (fig.1) for further development. We show here that, when administered orally to rodents and human volunteers, this compound is readily absorbed, releases (R)-MeHA in plasma and tissues, and displays antiinflammatory and antinociceptive properties in rodents.

	Methods

Top Abstract Introduction Methods Results Discussion References

Radioenzymatic assay of (R)-MeHA. Before developing a RIA (R)-MeHA was determined in plasma using a REA developed by Garbarg et al. (1989b) based on the observation that (R)-MeHA is a substrate for HMT and its methylated derivative is readily extractible into chloroform (Hough et al., 1981). Briefly, (R)-MeHA was mixed with [³H]S-adenosyl-methionine and a preparation of rat kidney HMT purified according to the method of Bowsher et al. (1983) slightly modified by Garbarg et al. (1989b). The reaction was stopped after 1 hr incubation at 25°C by addition of perchloric acid and the methylated derivative extracted into chloroform and quantified by liquid scintillation spectrometry.

RIA of (R)-MeHA and BP 2-94. The immunogen was prepared according to Ternynck and Avrameas (1977). (R)-MeHA was cross-linked to BSA using BZQ in a two-step reaction as described for HA and t-MeHA (Garbarg et al., 1989a) with slight modifications. Briefly, BSA (10 mg) was treated with 45 mg of BZQ. After 1-hr reaction the BZQ-BSA formed was purified by passing it through a Sephadex G-25 column and then mixed with 10 mg (R)-MeHA for a 24-hr incubation. After dialysis against water using [³H](R)-MeHA as a tracer in the coupling reaction, it was estimated that approximately 13 molecules of (R)-MeHA were conjugated to each BSA molecule (i.e., 45 µg antigen per mg of BSA). (R)-MeHA-BZQ-BSA conjugate (0.15 mg/animal) was injected i.d. to three female rabbits (New Zealand, Iffa-Credo, France) according to Garbarg et al. (1989a). Antibodies with sufficient titer and affinity were obtained after approximately 6 mo.

RIA of N-tele-(R)-dimethylhistamine. Based on its 100% cross-reactivity with t-MeHA, t-(R)-diMeHA was radioimmunoassayed using antibodies raised against t-MeHA (Garbarg et al., 1989a). Briefly, after treatment of plasma or tissues with perchloric acid (0.4 N final concentration) and derivatization with BZQ, t-(R)-diMeHA was mixed with a [¹²⁵I]iodinated tracer (t-MeHA-BZQ-Leu-[¹²⁵I]Tyr) and the antibodies. After 15 to 18 hr at 4°C, the bound radioactivity was precipitated and counted.

BP 2-94 hydrolysis in vitro. A 10 mM BP 2-94 solution prepared extemporaneously in DMSO was diluted to a final concentration of 4 µM in 0.4 N perchloric acid, 0.05 M potassium phosphate buffer, pH 7.4, or rat liver homogenates (prepared in 5 volumes, w/v, of 0.05 M Tris buffer pH 7.5) and incubated at 20, 37 or 95°C. At various time intervals, an aliquot was withdrawn, diluted and brought up to a final concentration of 0.4 N perchloric acid. Samples were immediately derivatized with BZQ and then radioimmunoassayed for (R)-MeHA. Hydrolysis was evaluated as the amount of (R)-MeHA formed during incubation. A blank value corresponding to the (R)-MeHA level at zero time incubation (and representing hydrolysis occurring during the RIA) was determined and subtracted.

HMT activity. HMT activity was quantified by measuring the conversion of HA, (R)-MeHA and BP 2-94 into corresponding tritiated derivatives, methylated in N-tele-position of the imidazole ring by using [³H]S-adenosyl-methionine as [³H]methyl donor. HMT was purified from rat kidney according to the method of Bowsher et al. (1983) slightly modified by Garbarg et al. (1989b). HA, (R)-MeHA or BP 2-94 were incubated at increasing concentrations with HMT and a mixture of unlabeled and [³H]labeled S-adenosylmethionine (5 µM final concentration) for 1 hr at 25°C. The reaction was stopped by addition of perchloric acid (0.4 N final concentration). Tritiated methylated derivatives were extracted into toluene-isoamyl alcohol (3:2) and quantified by liquid scintillation spectrometry.

[³H]HA release from synaptosomes of rat cerebral cortex. Release experiments with synaptosomes were performed according to Garbarg et al. (1992). Briefly, a crude synaptosomal fraction 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 BP 2-94 alone or together with HA (1 µM) was added. After 5 min the synaptosomes were depolarized by bringing the K⁺-concentration to 30 mM for 2 min. Incubations were ended by a rapid centrifugation and [³H]HA levels determined in the supernatant according to Garbarg et al. (1983).

Pharmacokinetic studies in healthy human volunteers. The design of these phase I clinical studies was approved by a local Ethical Committee and the studies authorized by the Agence Française du Médicament. They were performed in a specialized Phase I Clinical Study Center.

Capsaicin-induced plasma extravasation. Male Wistar rats (100-150 g, Iffa-Credo) were administered (R)-MeHA or BP 2-94 orally in increasing doses or their vehicle. After 90 min they were anesthetized with pentobarbital (6 mg/kg, i.p.). In one group of rats the following tissues were examined: skin (ears), eye conjunctiva, nasal mucosa, trachea, main bronchi, esophagus and urinary bladder (average weights: 548, 24, 56, 25, 25, 46 and 56 mg, respectively); they received capsaicin (90 µg/kg, i.v.) together with Evans blue dye (30 mg/kg) or vehicles 2 hr after oral treatment. Five min after capsaicin treatment animals were perfused with saline via the left cardiac ventricule for 2 min (constant flow: 24 ml/min) to remove intravascular dye. Tissues were dissected out and analyzed for extravasated Evans blue. Where specified trachea and main bronchi were analyzed together and designated "airways." Dye extraction was carried out by the method of Gamse et al. (1980). Tissues were immersed in formamide and maintained at 45°C for 18 hr. Extracted dye was measured by its absorption at 630 nm using a spectrophotometer (Dynatech, MR5000). Another group of animals were used to examine extravasation in dura mater (average weight: 17 mg) using a more sensitive method. Rats were administered FITC-labeled bovine serum albumin (FITC-albumin) (50 mg/kg, i.v.) or its vehicle 115 min after BP 2-94 and 5 min before capsaicin (300 µg/kg, i.v.) or its vehicle. Five min later, animals were perfused as described above, dura mater was carefully dissected out and homogenized in saline phosphate buffer. After centrifugation (15,000 g × 1 min) the fluorescence intensity (excitation wavelength 492 nm, emission wavelength 520 nm) (Kurose et al., 1994) was evaluated in the supernatant using a fluospectrometer (Jobin Yvon, JY3D). A control value corresponding to the basal dye level in the absence of capsaicin injection was determined for each tissue and subtracted. The inhibitory effect of H₃ receptor agonists on capsaicin-induced extravasation was calculated for each dose used as the percent ratio of: [dye (or FITC-albumin) concentration after administration of capsaicin plus BP 2-94 (or (R)-MeHA) minus concentration after capsaicin alone] over concentration after capsaicin alone. These calculated values were analyzed with an iterative computer least-squares method derived from that of Parker and Waud (1971) as described below (see Analysis of data) and ED₅₀ values and maximal inhibitory effects were deduced.

Antinociceptive activity. Male Swiss mice (20-25 g, Iffa-Credo) were used in three tests. The phenylbenzoquinone-induced writhing test was performed as described by Chaillet et al. (1983). BP 2-94, aspirin (Siegmund et al., 1957), (R)-MeHA or vehicle were given orally before injection of phenylbenzoquinone (PQ) (2 mg/kg, i.p.), and 10 min later, the number of writhing episodes evaluated during a 10-min observation period. When required, thioperamide (51 µmol/kg) or naloxone (30 µmol/kg) were administered i.p. 1 hr before BP 2-94 (0.16 µmol/kg, p.o.). The rather high doses of thioperamide and naloxone was selected to ensure H₃ or opiate receptor blockade for over 1 hr (von Voigtlander and Lewis, 1988; Garbarg et al., 1989b; Fujibayashi and Iizuka, 1995). The inhibitory effect of increasing doses of BP 2-94 on PQ-induced writhing was determined as the percent ratio of the difference of writhing score after vehicle and BP 2-94 treatment over score after vehicle alone, and used to calculate the ED₅₀ value and maximal effect of BP 2-94 as mentioned below (see "Analysis of data").

Zymosan-induced edema in mice. Edema was induced in male Swiss mice (20-25 g) by injecting 25 µl of 0.05% zymosan suspension in saline into the left hind paw (Stefanova et al., 1995). The right hind paw was used as a control and was injected with 25 µl of 0.9% saline. The animals were killed 4 hr later, both hind paws were cut off at the ankle, and the difference between their weights was calculated. BP 2-94 was orally administered 1 hr before, simultaneously with or 1 hr after zymosan injection. The effect of BP 2-94 was also studied in the presence of thioperamide (51 µmol/kg) administered i.p.1 hr before BP 2-94.

Analysis of data. For determination of ED₅₀ values and maximal effects, inhibitory effects of drugs were analyzed with an iterative computer least-squares method derived from that of Parker and Waud (1971), using the following non linear regression: Half-lives have been determined by a computerized program using the following nonlinear regression: y = A exp^Bt + C exp^Dt + E exp^Ft. Statistical analysis were by one-way analysis of variance and Dunnett's or Tukey-Kramer's post tests with a preset probability level of P < .05.

Radiochemicals, drugs and drug solutions. S-adenosyl-L-[³H]methyl methionine (70 Ci/mmol), and [¹²⁵I]NaI (2,000 Ci/mmol) were from Amersham (Amersham, U.K.). Tritiated (R)-MeHA (10 Ci/mmol) was prepared as described by Arrang et al. (1990). t-MeHA was from Sigma Chemical Co. (St. Louis, MO), and t-(R)-diMeHA, , -diMeHA were synthesized by one of us (W.S.). N-MeHA, 2-MeHA, 4-MeHA and N, N-diMeHA were from Smith Kline Beecham (London, UK). 2-0H-BZP was from Acros (Pittsburgh, PA). BP 2-94 and (R)-MeHA were from Laboratoire Bioprojet (Paris, France). For administration to animals, BP 2-94, (R)-MeHA or morphine and aspirin (Sigma) were introduced into 1% methylcellulose plus 5% DMSO; naloxone (Sigma) and thioperamide (a kind gift of Pr. Robba, Caen) were solubilized in DMSO, PQ (Sigma) was dissolved in 0.9% NaCl, 3% ethanolic. Capsaicin (Sigma) was solubilized in saline:DMSO:distilled water (36:3:1). Evans blue dye was prepared in 0.9% NaCl and filtered (Millipore, 0.45-µm pore diameter) before use. FITC-albumin (Sigma) was diluted in 0.9% NaCl. Zymosan (Sigma) was suspended in 0.9% saline by sonication and heated for 30 min at 100°C before use. All chemicals were from Sigma except BZQ (reagent grade) which was from Fluka AG (Buchs, Switzerland), synthetic peptides which were from Bachem (Bubendorf, Switzerland), and S-adenosyl-L-methionine which was from Boehringer (Mannheim, FRG). All other reagents (analytical grade) were from commercial sources and were of the highest purity available.

	Results

Top Abstract Introduction Methods Results Discussion References

Characteristics of the RIA for (R)-MeHA. The RIA was designed using a pool of bleedings exhibiting the best binding parameters. The final antiserum dilution used to obtain 10 to 20% binding of the tracer, in the absence of competing derivatized amine (B₀) was 1:32,000. The pH of the reaction and the BZQ concentration were selected to minimize the interference of BZQ with the antibodies and to obtain the maximal derivatization yield. Addition of (R)-MeHA derivatized with BZQ progressively inhibited the binding of the [¹²⁵I]tracer to antibodies (fig. 2) with an IC₅₀ of 0.5 ± 0.1 nM. The detection limit, defined as the concentration corresponding to 20% inhibition, was 10 pg/well (2 ng/ml). The specificity of the antibodies was tested by measuring the cross-reactivities of various BZQ-derivatized compounds (fig. 2). The cross-reactivity of HA was 7% and that of histidine, 2-0H-BZP and various methylated derivatives of HA was less than 0.05%. The apparent cross-reactivity of BP 2-94 (1-2%) reflects hydrolysis of BP 2-94 into (R)-MeHA taking place during the 30-min derivatization step. Dilutions of derivatized plasma from rats pretreated with (R)-MeHA (0.24 mmol/kg, p.o.) inhibited the binding of the [¹²⁵I]tracer to antibodies with an inhibition curve paralleling that of the standard (fig. 2). The recovery of (R)-MeHA added to plasma or tissue extracts was about 90%, and all results were corrected accordingly.

View larger version (25K): [in this window] [in a new window]   Fig. 2.   RIA of (R)-MeHA: Inhibition of [125I]tracer binding to antibodies by various compounds and plasma extracts from rats pretreated with (R)-MeHA. Rats received (R)-MeHA (0.24 mmol/kg, p.o.) 2 hr before being killed and their plasma was radioimmunoassayed for (R)-MeHA after dilution and derivatization with BZQ. All compounds were pretreated with BZQ as described in "Methods," preincubated in the presence of antiserum and incubated with a 10 pM concentration of [125I]tracer.

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Chemical structures of (R)-MeHA and BP 2-94.

In vitro hydrolysis of BP 2-94. The compound BP 2-94 was incubated to estimate its ability to release (R)-MeHA under various conditions. The hydrolysis rates of BP 2-94 at room temperature were similar in neutral and strongly acidic media, representing about 4% per hour. To investigate a possible enzymatic hydrolysis of BP 2-94 incubations were also performed in the presence of rat liver homogenates at 37°C. Under such conditions the hydrolysis rate was in the same range as that obtained in the absence of tissue. Conditions ensuring a complete hydrolysis of BP 2-94 into (R)-MeHA were also investigated for measuring BP 2-94 levels in biological samples. Thus, incubation of BP 2-94 in an acidic medium at 95°C for 30 min, leading to a complete hydrolysis, was selected. The specificity of the assay was assessed in mouse plasma samples having received 0.3 mmol/kg of BP 2-94 p.o. and killed 1 hr later: HPLC analysis on a C₁₈ µBondapak column showed only two immunoreactive peaks corresponding to (R)-MeHA and BP 2-94, the latter after hydrolysis (retention times 6 and 28 min, respectively; linear gradient of 10-50% AcN in 10 mM Ac0NH₄ over 30 min).

Methylation of (R)-MeHA and BP 2-94 by HMT. BP 2-94 was tested as a possible substrate of HMT in comparison with HA and (R)-MeHA. The K_M values of HMT were 3.0 and 2.2 µM and the V_max values 2.2 and 1.7 nmol/mg/hr for HA and (R)-MeHA, respectively, but methylation was not detectable with BP 2-94 as a substrate using the standard assay.

Effects of BP 2-94 and t-(R)-diMeHA on [³H]HA release from rat brain synaptosomes. BP 2-94 (in concentrations from 1 µM to 1 mM) tested as an H₃ receptor antagonist was ineffective in preventing the inhibition of depolarization-induced [³H]HA release from brain synaptosomes elicited by 1 µM HA. When tested as an agonist in concentrations from 1 to 100 µM, BP 2-94 inhibited [³H]HA release by up to 72% with an EC₅₀ of about 10 µM to be compared with 4.0 ± 0.9 nM for (R)-MeHA (Garbarg et al., 1992). t-(R)-diMeHA inhibited [³H]HA release with an EC₅₀ about 50 µM.

Distribution of BP 2-94 and (R)-MeHA in mouse tissues after oral administration of BP 2-94. After oral administration of 24 µmol/kg of BP 2-94 to groups of four mice, both the prodrug and the active drug (R)-MeHA-ir were detected in plasma and various tissues as early as 30 min later. The levels of both compounds peaked at 1 hr and then declined with a half-life of about 1 hr. In figure 3 A.U.C.s derived from these data are shown which indicate that similar levels were reached in lung and plasma, whereas levels in liver and kidney were twice as high and hardly detectable in cerebral cortex. C_max values (in nmol/g or nmol/ml) were 1.1, 1.7, 4.4 and 1.5 for BP 2-94 in lung, kidney, liver and plasma, respectively, whereas corresponding values for (R)-MeHA were 0.8, 2.1, 2.0 and 0.6. 

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Tissue and plasma levels of BP 2-94 and (R)-MeHA immunoreactivities in mice receiving only BP 2-94: compared areas under the curves. Groups of four mice were killed at various times (1-16 hr) after oral administration of BP 2-94 (24 µmol/kg). (R)-MeHA was radioimmunoassayed in tissue and plasma extracts before and after hydrolysis of BP 2-94 into (R)-MeHA. BP 2-94 level was calculated as the difference between these two determinations. Results expressed in nmol.hr.g1 for tissues and in nmol.hr.ml1 for plasma.

Pharmacokinetics of (R)-MeHA and BP 2-94 in human volunteers. In human volunteers receiving 1.4 mmol of (R)-MeHA orally the plasma level of (R)-MeHA (determined by a REA) was maximal 1.7 hr after the administration and decayed with an apparent half-life of about 1.6 hr. The levels of t-(R)-diMeHA immunoreactivity displayed similar changes. C_max values were 40.2 pmol/ml and 3.0 nmol/ml for (R)-MeHA and t-(R)-diMeHA, respectively. In this study, the A.U.C. of plasma t-(R)-diMeHA levels represented 141-fold that of the plasma (R)-MeHA level (table 1). For the purpose of comparison, plasma levels of (R)-MeHA and its methylated derivative were also evaluated in mice receiving 24 µmol/kg of (R)-MeHA orally and A.U.C.s calculated (table 1). In contrast to the data in human volunteers the ratio of the two A.U.C.s was only about 1.5. 

View this table: [in this window] [in a new window]   TABLE 1 Plasma levels of (R)-MeHA and its methylated metabolite after oral administration of (R)-MeHA to mice and healthy human volunteers

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Compared plasma levels of (R)-MeHA-ir (logarithmic scale) after administration of either (R)-MeHA or BP 2-94 to healthy human volunteers. A, Levels of (R)-MeHA and BP 2-94 after administration of BP 2-94 (0.1 mmol, p.o.) to 12 human volunteers. Means ± S.E.M. B, Levels of (R)-MeHA after administration of either (R)-MeHA (1.4 mmol, p.o.) to 6 human volunteers or BP 2-94 (0.1 mmol, p.o.) to 12 human volunteers. Means ± S.E.M.

Effects of BP 2-94 on capsaicin-induced plasma protein extravasation in rat tissues. After capsaicin administration the amount of Evans blue dye (µg/g) raised from 16.3 ± 1.9 to 159.5 ± 12.5, from 13.9 ± 1.8 to 107.9 ± 10.2 and from 6.1 ± 0.5 to 25.5 ± 1.9 µg/g, respectively, within bronchi, esophagus and bladder, representing the percent increases (table 2). Corresponding values were 2.3 ± 0.3, 5.3 ± 0.6, 17.5 ± 1.3 and 12.7 ± 1.2 µg/g in skin, conjunctiva, nasal mucosa and trachea of controls and increased significantly (3- to 7-fold) after capsaicin (table 2). In dura mater, plasma extravasation was evaluated in animals having received FITC-albumin whose levels were 21.2 ± 2.0 µg/g in controls and enhanced by 86% after capsaicin (table 2). Pretreatment with BP 2-94 in increasing doses (0.8-240 µmol/kg) significantly reduced the response to capsaicin in all tissues studied. This inhibitory effect took place dose-dependently with an ED₅₀ of 0.8 ± 0.4, 3.0 ± 1.2 and 0.6 ± 0.1 µmol/kg and a maximal reduction of capsaicin-induced extravasation of 54, 65 and 63% in bronchi, esophagus and bladder respectively (fig. 5). The ED₅₀ values and maximal effects in the other tissues were derived from similar dose-response curves and are reported in table 2. Administration of thioperamide (51 µmol/kg) 1 hr before BP 2-94 (40 µmol/kg) reversed completely its inhibitory effect in all tissues but thioperamide alone did not significantly affect the capsaicin-induced extravasation. This is shown on figure 6 in bronchi plus trachea (airways). In addition, administration of (R)-MeHA (80 µmol/kg) also inhibited significantly (by 40%) the response to capsaicin in airways, and thioperamide reversed this effect (fig. 6).

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Effect of BP 2-94 on capsaicin-induced plasma extravasation in rat bronchi, esophagus and urinary bladder. The animals were given BP 2-94 orally at increasing dosages (logarithmic scale) 2 hr before capsaicin (90 µg/kg, i.v.) and Evans blue dye (30 mg/kg, i.v.) and were perfused 5 min after capsaicin. *P < .01, **P < .001. Means ± S.E.M. of 6 to 37 values.

View larger version (36K): [in this window] [in a new window]   Fig. 6.   Thioperamide reversal of the inhibition by (R)-MeHA or BP 2-94 of capsaicin-induced plasma extravasation in rat airways. Thioperamide (51 µmol/kg, i.p.) was administered 1 hr before (R)-MeHA (80 µmol/kg, p.o.) or BP 2-94 (40 µmol/kg, p.o.) and 2.5 hr before capsaicin (90 µg/kg, i.v.) and Evans blue dye (30 mg/kg, i.v.). Animals were perfused 5 min after capsaicin treatment. a, P < .01 as compared to capsaicin alone; b, P < .01 as compared to capsaicin and (R)-MeHA or BP 2-94. Means ± S.E.M. of 8 to 14 values.

Antinociceptive activity of BP 2-94 in mouse. Phenylbenzoquinone-induced writhing was significantly reduced in mice receiving BP 2-94 orally. The antinociceptive activity of the prodrug was dose-related and occurred with an ED₅₀ of 0.03 ± 0.01 µmol/kg and a maximal reduction of the writhing score of 70%, a value close to that obtained with aspirin in maximal dosage (556 µmol/kg) (fig. 7). Combination of BP 2-94 and aspirin at moderate dosages elicited an additive reduction of the nociceptive response to PQ (-32 and -39% for BP 2-94 and aspirin, respectively vs. -56% when given together). The antinociceptive activity of BP 2-94 (16 µmol/kg) was maximal after 1 hr, still significant after 3 hr and, although not significant, represented a 43% decrease after 6 hr (fig. 8). (R)-MeHA (160 µmol/kg) given orally 1 hr before PQ induced an effect similar to that elicited by a maximal dose of BP 2-94. DMSO itself, the vehicle used for thioperamide and naloxone, reduced the nociceptive effect of PQ by about 50% (writhing score: 11.2 ± 1.9 vs. 21.0 ± 2.3 for DMSO and control treated mice, respectively). However, BP 2-94 (0.16 µmol/kg) and aspirin (556 µmol/kg) administered after DMSO were still effective in inhibiting the nociceptive effect of PQ. The antinociceptive effect of BP 2-94 (0.16 µmol/kg) was significantly abolished by previous administration of thioperamide (51 µmol/kg), but not by naloxone (30 µmol/kg). Thioperamide alone did not change the writhing score, and a slight but not significant enhancement of the writhing score could be noticed after naloxone.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Antinociceptive activity of BP 2-94 in the phenylbenzoquinone (PQ) writhing test in mouse. PQ (2 mg/kg, i.p.) was administered i.p. 1 hr after BP 2-94 or aspirin (p.o.), and 10 min later the number of writhing episodes were counted for 10 min. *P < .05, **P < .001 as compared to control mice. Means ± S.E.M. of 6 to 45 values.

View larger version (13K): [in this window] [in a new window]   Fig. 8.   Kinetics of the antinociceptive effect of BP 2-94 in the phenyl-benzoquinone writhing test in mouse. BP 2-94 (16 µmol/kg, p.o.) was given at various times before administration of PQ (2 mg/kg, i.p.). Starting from 10 min after PQ administration, the number of writhing episodes displayed by the animals was counted for 10 min. *P < .05, **P < .01 as compared to control mice. Means ± S.E.M. of 9 to 33 values.

View larger version (13K): [in this window] [in a new window]   Fig. 9.   Antinociceptive activity of BP 2-94 in the mouse formalin test. BP 2-94 (16 µmol/kg, p.o.) or morphine (35 µmol/kg, p.o.) were administered 1 hr before formalin injection into the hind paw. The duration of licking and biting was measured for 5 min after the injection of 20 µl of 0.2% formalin (early phase) or 20 min after 20 µl 1% formalin (late phase). *P < .01, **P < .001. Means ± S.E.M. of 6 to 12 values.

View larger version (32K): [in this window] [in a new window]   Fig. 10.   Effect of BP 2-94 on jump latency in the hot plate test in mouse. Mice were given BP 2-94 (16 µmol/kg, p.o.) or morphine (35 µmol/kg, p.o.) 1 hr before evaluation of the jump latency time. *P < .001 as compared to control or BP 2-94. Means ± S.E.M. of 9 to 10 values.

Antiinflammatory activity of BP 2-94 in mouse. After preliminary trials the inflammatory effect of zymosan was studied 4 hr after administration of 25 µl of a 0.05% suspension. The inflammatory response to zymosan, i.e., edema, was significantly reduced by about 50% in mouse receiving BP 2-94 (66 µmol/kg, p.o.) at the same time as, before or after zymosan administration. Administered 1 hr before zymosan BP 2-94 decreased in a dose-dependent manner the zymosan-induced edema with an ED₅₀ of 1.0 ± 0.4 µmol/kg and a maximal effect of 60%. The antiinflammatory effect of BP 2-94 was significantly abolished by prior administration of thioperamide (51 µmol/kg, i.p.) that alone did not induce significant change (fig. 11).

View larger version (30K): [in this window] [in a new window]   Fig. 11.   Thioperamide reversal of the effect of BP 2-94 on zymosan-induced paw edema in mouse. Thioperamide (51 µmol/kg, i.p.) was injected 1 hr before BP 2-94 (1.6 µmol/kg, p.o.) and 2 hr before zymosan injection. a, P < .01 as compared to zymosan, b, P < .001 as compared to BP 2-94. Means ± S.E.M. of 10 values.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Our study identifies BP 2-94 as an optimal prodrug of (R)MeHA, enhancing markedly the plasma level of the latter in healthy human volunteers and exerting potent antiinflammatory activity mediated by H₃ receptors on capsaicin-sensitive fibers in rodents.

Initial studies showed that (R)-MeHA displays significantly lower oral bioavailability in humans compared to rodents, as judged from the radioenzymatic assay of plasma levels (table 1). Since much higher plasma levels of t-(R)-diMeHA were detected in humans receiving (R)-MeHA orally, it was hypothesized that this was due to an extensive imidazole ring methylation by the enzyme histamine-N-methyltransferase (EC2.1.1.8) during the first pass in the liver and leading to an inactive metabolite. In support of this hypothesis (R)-MeHA is readily methylated by the enzyme, and the product is, as with any other ring-substituted compound, lacking any agonist activity at the H₃ receptor (see Arrang et al., 1983; Ganellin et al., 1995; Stark et al., 1994, 1995) (see "Results"). In addition, a much lower hepatic HMT activity is found in the rat (which inactivates HA in peripheral tissues mainly by oxidative deamination) than in many other species including humans (Brown et al., 1959; Hesterberg et al., 1984). To circumvent this difficulty we have explored a strategy based on the design of a series of azomethine derivatives of (R)-MeHA in which the ammonium group, being essential for recognition by HMT (Barth and Lorenz, 1978; Barth et al., 1980) (as well as by the H₃ receptor), is reversibly engaged in a Schiff base, the latter being stabilized by hydrogen bonding with the hydroxyl group of the hydroxybenzophenone moiety (Garbarg et al., 1994; Krause et al., 1995). Such azomethines were previously used in halogenated form as lipophilic prodrugs to promote the brain penetration of GABA (Kaplan et al., 1980; Bergmann, 1985), but apparently were never used before with the aim of protecting a bioactive compound from metabolic degradation.

To identify an optimal prodrug among a large series of azomethine derivatives, i.e., with adequate hydrolysis rate in vitro, high oral bioavailability and adequate generation of (R)-MeHA in tissues, we developed a sensitive and specific RIA for this amine after its derivatization according to a principle previously applied to RIAs for HA and t-MeHA (Garbarg et al., 1989a). This assay was also applied to the determination of the level of the prodrug in tissues after its total hydrolysis into (R)-MeHA by heating the tissue extracts in acidic medium.

In agreement with our expectations BP 2-94, a compound selected among a series of related azomethine derivatives of (R)-MeHA, was no longer a substrate for the HA-methylating enzyme. On its unchanged form it did not display any significant H₃ receptor agonist activity, but it was slowly hydrolyzed in vitro into the potent agonist (R)-MeHA. This hydrolysis appeared to be essentially of chemical nature because it was not accelerated in the presence of tissue extracts and also occurred at a slow rate in vivo, resulting in long-lasting (over 24 hr) plasma levels of the prodrug and (R)-MeHA in human volunteers.

In the latter, 24 hr after an oral dose of 0.1 mmol (about 0.5 mg/kg) of BP 2-94, plasma levels of (R)-MeHA-ir were ~30 nM, i.e., one order of magnitude higher than the EC₅₀ of the drug at the H₃ receptor. The success of the prodrug strategy was also shown by the markedly improved human bioavailability of (R)-MeHA when administered orally in form of BP 2-94. The A.U.C. of (R)-MeHA-ir in plasma was approximately 100 times higher in this case.

In addition, in mice receiving BP 2-94 orally, high levels of both the prodrug and (R)-MeHA-ir were found in most tissues, except in the brain where they remained almost undetectable at any time. This was rather unexpected if one takes into account the central effects of GABA prodrugs, which are also azomethine derivatives of 4-hydroxybenzophenone (Kaplan et al., 1980; Jilek et al., 1990) but in halogenated form. In fact, substitution of the hydroxybenzophenone moiety by halogens leads also to prodrugs of (R)-MeHA with markedly enhanced brain penetration as compared with BP 2-94, and which therefore could, in contrast to BP 2-94, be targeted to therapeutic applications resulting from H₃ receptor stimulation in brain (Garbarg et al., 1994; Krause et al., 1995).

In rodents, oral administration of BP 2-94 in low dosage resulted in a series of characteristic and long-lasting responses, all attributable to H₃ receptor activation by slowly released (R)-MeHA. In all tissues tested there was particularly an inhibition of capsaicin-induced plasma protein extravasation, generally by up to 60% (but up to 87% in dura mater), which occurred with widely varying ED₅₀ values (table 2). The effect of capsaicin is known to result indirectly from release of proinflammatory neuropeptides, such as tachykinins or CGRP from perivascular peripheral endings of unmyelinated sensory C-fibers (Holzer, 1991; Maggi, 1995). The inhibitory modulatory role of the H₃ receptor on these sensory fibers was first demonstrated on vagal nerve endings in airways (Ichinose and Barnes, 1989a; Ichinose et al., 1989) and then on trigeminal nerve endings in dura mater (Matsubara et al., 1992), in both cases evidenced by plasma protein extravasation. In addition, H₃ receptor mediated inhibition of the immunoreactive substance P release elicited by antidromic stimulation of rat sciatic nerve (Ohkubo et al., 1995) and of CGRP release in heart (Imamura et al., 1996) were demonstrated directly.

These data, together with those of our study, suggest that the H₃ receptor is expressed by C-fibers ending in a large variety of tissues, i.e., not only by sensory cranial nerves but also sensory spinal nerves like those ending in the urinary bladder. The mechanism responsible for the presynaptic inhibition of neuropeptide release could be the same as for the other presynaptic effects of H₃ receptor agonists, i.e., inhibition of calcium ion influx (Arrang et al., 1990) or activation of calcium- or ATP-sensitive potassium channels (Stretton et al., 1992; Ohkubo and Shibata, 1995). These mechanisms are shared by other receptors displaying similar presynaptic localizations and functions, i.e., somatostatin, ₂ adrenergic or opiate receptors (Maggi, 1995). From a physiological point of view the presence of the H₃ receptor on C-fibers has been proposed to reflect its participation in a negative feedback loop controlling the release of mast-cell mediators (including HA) by these closely apposed fibers (Dimitriadou et al., 1994). Via this loop HA release triggered by the secretion of endogenous substance P and/or CGRP is limited by activation of H₃ receptors inhibiting substance P and/or CGRP release (Foreman, 1987; Imamura et al., 1996; Okhubo et al., 1994).

The alleged role of tachykinins retrogradely or anterogradely released from sensory fibers in the mediation of inflammatory and nociceptive responses, respectively, prompted us to test the effect of BP 2-94 in classical rodent models of inflammation and nociception.

Zymosan-induced paw swelling was inhibited via H₃ receptor stimulation after BP 2-94 administration, as shown by the blockade of this effect elicited by thioperamide. The multiple inflammatory responses to zymosan, an insoluble fraction of yeast cell wall, are known to include 1) generation of anaphylatoxins that induce HA release from mast cells, 2) biosynthesis of eicosanoids by neutrophils or macrophages, 3) generation and release of PAF, oxygen free radicals and lysosomal enzymes (Doherty et al., 1985; Rao et al., 1994). Additional studies are required to establish which of these various inflammatory pathways are modified by H₃ receptor agonists, although it seems likely that the reductions in proinflammatory substance P (Payan, 1989) release and impairment of mast cell reactivity induced by these agents might contribute.

Finally, a marked and long-lasting antinociceptive activity of BP 2-94 in low dosage (ED₅₀ = 0.03 ± 0.01 µmol/kg, p.o.), also clearly mediated by the H₃ receptor and independent from endogenous opioids, was evidenced in the PQ-induced writhing test. The maximal effect of BP 2-94, i.e., a 70% reduction in the number of abdominal torsions, was similar to that of aspirin, and the effects of the two agents in moderate dosages were apparently additive.

In the formalin test the antinociceptive effect of BP 2-94 in maximal dosage was less marked than that of morphine during the early phase (which is thought to correspond to direct activation of sensory fibers by formalin) but equivalent to the latter in the late phase, which may correspond to a secondary inflammatory reaction (Kayser and Guilbaud, 1994). In contrast, no significant antinociceptive activity was detected in the hot plate jump test. Taken together these observations suggest that the antinociceptive activity of the H₃ receptor agonist results from an inhibition of nociceptive messages transmission by sensory C-fibers via an action at peripheral sites.

The pattern of peripheral antinociceptive and antiinflammatory actions of the H₃ receptor agonist evidenced here and largely attributable to the widespread inhibition of tachykinin release, suggests novel therapeutic applications for this class of drugs. Tachykinins modulate the activity of a number of different leukocytes involved in both acute and delayed inflammatory responses and may play a role in the pathogenesis of such diverse diseases as arthritis, asthma and inflammatory bowel diseases (Payan, 1989). The applications of H₃ receptor agonists might be more general than those of tachykinin receptor antagonists that are generally specific for a single receptor subtype. In addition, the application of these drugs in asthma is further supported by their inhibitory effect on vagally induced bronchoconstriction (Barnes, 1992). In migraine and related disorders, their efficacy might derive from the inhibition of release of both proinflammatory tachykinins and vasodilatory CGRP from trigeminal nerve endings, a mechanism proposed to account for the therapeutic efficacy of sumatriptan (Buzzi et al., 1991; Moskowitz, 1991). These various hypotheses are currently tested in ongoing clinical trials with BP 2-94 due to the good plasma level of this compound. Finally, it should be stressed that among antiinflammatory agents H₃ receptor agonists display the interesting and unique property of decreasing gastric acid secretion (Bertaccini and Coruzzi, 1995) and exerting antiulcer activity (Morini et al., 1995) in relation with the inhibition of HA release from enterochromaffin-like cells (Prinz et al., 1993; Soldani et al., 1996).