Synthesis and Characterization of New Bivalent Agents as Melatonin- and Histamine H3-Ligands

Melatonin is an endogenous molecule involved in many pathophysiological processes. In addition to the control of circadian rhythms, its antioxidant and neuroprotective properties have been widely described. Thus far, different bivalent compounds composed by a melatonin molecule linked to another neuroprotective agent were synthesized and tested for their ability to block neurodegenerative processes in vitro and in vivo. To identify a novel class of potential neuroprotective compounds, we prepared a series of bivalent ligands, in which a prototypic melatonergic ligand is connected to an imidazole-based H3 receptor antagonist through a flexible linker. Four imidazolyl-alkyloxy-anilinoethylamide derivatives, characterized by linkers of different length, were synthesized and their binding affinity for human MT1, MT2 and H3 receptor subtypes was evaluated. Among the tested compounds, 14c and 14d, bearing a pentyl and a hexyl linker, respectively, were able to bind to all receptor subtypes at micromolar concentrations and represent the first bivalent melatonergic/histaminergic ligands reported so far. These preliminary results, based on binding affinity evaluation, pave the way for the future development of new dual-acting compounds targeting both melatonin and histamine receptors, which could represent promising therapeutic agents for the treatment of neurodegenerative pathologies.


, calmodulin,
calreticulin, and tubulin [4], which could have a role in those phenotypic effects of melatonin that are independent of the activation of membrane-bound receptors.Additionally, melatonin effects could be also sustained by its metabolites, since it is rapidly transformed in the peripheral sites [5].In addition to its well-established function in the regulation of the sleep-wake cycle and in the entrainment of circadian rhythms [6], melatonin is involved in a variety of other pathophysiological processes, including radical scavenging, attenuation of oxidative damage and neuroprotection [7][8][9][10].Experimental evidence highlights the important role played by the activation of MT1 and MT2 receptors in sustaining its antioxidant and neuroprotective actions [11][12][13][14].Even if the mechanism at the basis of this neuroprotective effect has not been completely elucidated, it is probably related to a multifactorial action exerted within the cell.For example, in motoneurons melatonin attenuated the production of reactive oxygen species, modulated Ca 2+ levels and inhibited proapoptotic signaling [11,12].In this context, not only melatonin but also synthetic compounds, like 2-iodomelatonin or other non-indole melatonergic ligands, have shown to efficiently produce antioxidant and cytoprotective effects and to block neurodegenerative processes both in vitro and in vivo [15,16].Recently, different series of dual-acting compounds, constituted by melatonin linked to another known neuroprotective agent, have been reported as novel potential therapeutic agents for the treatment of neurodegenerative disorders.Indeed, hybrid melatonin-tacrine compounds showed potent anticholinesterase and antioxidant activity and melatonin-N,N-dibenzyl(N-methyl)amine hybrids also showed neuroprotective effects and were proposed as new potential therapeutic agents for neurodegenerative pathologies such as Parkinson's and Alzheimer's diseases.[17,18] Curcumin-melatonin hybrids showed antioxidant and neuroprotective actions as well, as demonstrated in in vitro studies on a cellular model of Alzheimer's disease [19].

Histamine H3 receptors are mainly expressed in the CNS, with a predominantly presynaptic localization.They have been characterized as both autoreceptors and heteroreceptors, exerting a negative feedback mechanism on the synthesis and release of histamine and on the release of other neurotransmitters, such as acetylcholine, noradrenaline, dopamine, serotonin, etc. [20].Histamine H3 receptor antagonists, which increase the release of such neurotransmitters, have been extensively investigated for the treatment of different CNS pathologies, such as narcolepsy, attention deficit hy eractivity disorder, obesity, and Alzheimer's disease [21,22] and a number of potent and selective compounds are currently undergoing clinical trials.H3-antagonists have been also investigated for their neuroprotective potential and the improvement of cognitive disorders [23].Several experimental studies evaluated the neuroprotective effects of different classes of H3 receptor antagonists in vitro and in vivo [24][25][26], highlighting their potential usefulness in the treatment of cognitive pathologies.Indeed, pretreatment with the H3 receptor antagonist ABT-239 was able to significantly attenuate kainic acid-mediated behavioral and excitotoxic effects [27].Recently, a new class of compounds has been reported in which the pharmacophore for H3 receptor antagonists was combined with a 3-indolyl-alkyl portion.These compounds exhibited potent H3 receptor antagonist activity and free radical scavenging properties and were hypothesized to be superior agents for Alzheimer's disease therapy by acting in a complementary manner [28].


OH


O

Given the promising neuroprotective effects shown by melatonin receptor agonists and H3 receptor antagonists, we combined the pharmacophore elements of these classes into a single molecule, with the aim to retain the ability to bind both receptor subtypes.Structure-activity relationships (SARs) for melatonin receptor ligands showed that the introduction of an aryl-alkyloxy chain of suitable length in a position corresponding to that of the 5-methoxy group of melatonin is tolerated and leads to potent compounds selective for the MT1 subtype (e.g., compounds 2-4, Figure 1) [29][30][31].According to the results of docking studies on a homology model of the MT1 receptor, it has been hypothesized that the substituent conferring subtype selectivity can be accommodated within a lipophilic channel, exposed to the solvent at its cytosolic terminus, which is available in the MT1 receptor.In the MT2 receptor this channel is much more crowded, given the presence of bulkier amino acids hampering the accommodation of the aryl-alkyloxy chain in the same manner as in the MT1 receptor [31].

The classical pharmacophore model for H3 receptor antagonists is composed by three main portions, i.e., a basic group, a central lipophilic core usually connected through an alkyl spacer and a terminal group, which displays high chemical diversity as it could be a polar group, a lipophilic group or another basic center [32].The first H3 receptor antagonists were imidazole derivatives, while the second generation of compounds carried a different basic group, usually a piperidine or a pyrrolidine.Several series of potent imidazole-based H3 antagonists lacking a second basic center have been reported [33].A polar atom/group is often present in these compounds, such as a thiourea (compound 5, Figure 2), a sulfonamide, or an oxygen atom (compounds 6 and 7).Good antagonists were also obtained with the insertion of a lipophilic chain only (compound 8), indicating that the presence of the imidazole ring is sufficient to preserve binding to the H3 receptor [34].In addition, SARs for these non-basic imidazole-based H3 antagonists showed that the length of the alkyl linker can be significantly increased, sometimes reaching 9-10 methylene units, while maintaining significant H3 binding affinity [35][36][37][38].The possibility of obtaining high H3 receptor binding affinity for compounds lacking a strongly basic center is an important point in light of the merging of the pharmacophore elements of H3-antagonists with those of melatonergic agonists.In fact, introduction of basic groups on the scaffold of melatonergic ligands invariably brought to a loss of binding affinity [39,40].On the other hand, the tolerance shown by the H3 receptor for the length of the alkyl spacer supports the possibility of combining H3 receptor and melatonin receptor pharmacophore elements through a suitable alkyl spacer.In fact, good MT1 binding affinities could be obtained with aryl-alkyloxy substituents having alkyl chains of four methylene units or longer.

The melatonergic pharmacophore was provided by an anilinoethylamide fragment which already proved to bioisosterically reproduce the indol-3-ylethylamide portion of melatonin, affording compounds with similar binding affinity and intrinsic activity [41].Indeed, being the anilinoethylamide smaller than the indol-3-ylethylamide of melatonin, it could be better tolerated at the H3 receptor binding site.The anilinoethylamide fragment was decorated with an aryl-alkyloxy substituent containing the elements necessary for molecular recognition at the H3 receptor.The aryl-alkyloxy chain was a 4(5)-i idazolyl-alkyloxy one, seen in ciproxifan (7) and its analogs [42] (Figure 3).Docking studies suggested that these compounds could be accommodated within the MT1 receptor binding site.The 4(5)-imidazolyl-alkyl portion could occupy the lipophilic pocket delimited by transmembrane (TM) helices 3, 4, and 5, with the terminal imidazole ring positioned at the rim of the TM portion of the receptor, where it could undertake polar interactions with amino acids in helices or extracellular loops (Figure 4, left).

The H3 receptor should be able to interact with these compounds by binding their imidazole ring through E206 on TM5, as supported by mutagenesis experiments on histamine and other imidazole-based ligands [43].The anilinoethylamide fragment could be docked within a lipophilic cavity identified in H3 receptor models [44,45], mainly delimited by TM2, 6 and 7, and roughly perpendicular to the region where E206 is located (Figure 4, right).The existence of this lipophilic cavity is supported by the high binding affinity shown by H3 receptor antagonists with two basic centers and a lipophilic substituent that could extend over both binding site cavities (Figure 5) [45][46][47].We report here the synthesis and evaluation of the binding affinity for histamine H3 and melatonin MT1 and MT2 receptors of four 3-(4(5)-imidazolyl-alkyloxy)-anilinoethylacetamides in which alkyl spacers of different length were introduced at position 3 of the aniline core, looking for an optimal linker to combine the H3 receptor-and the melatonin receptor-binding moieties.


Results and Discussion


Chemistry

The synthesis of the dual melatonergic/histaminergic ligands is described in Schemes 1 and 2.

The key starting alcohols 11a, 11c-d were prepared from the suitable aldehydes 8a, 8c-d following the previously reported three-step sequence (Wittig reaction, hydrogenation, LiAlH4 ester reduction) [48], whereas the alcohol 11b could be obtained by direct reduction of the aldehyde 8d [48] (Scheme 1).

The N-protected compounds 13a-d were obtained by mesylation of the suitable (1-trityl-imidazol-4-yl)alkan-1-ol 11a-d, followed b

mediate me
hansulfonates 12a-d with N-{2-[(3-hydroxyphenyl)methylamino]ethyl}acetamide [31] in the presen e of NaH.Deprotection of the imidazole ring under acidic conditions finally yielded the target compounds 14a-d (Scheme 2).


Binding Affinities of Compounds 14a-d for Melatonin MT1, MT2 and Histamine H3 Receptors

Binding affinities at human MT1, MT2 and H3 receptors of the newly synthesized N-(4(5)-im dazolyl-alkyloxy-anilinoethyl)acetamides 14a-d were assessed as described in the Experimental Section and are reported in Table 1.The bivalent melatonergic/histaminergic ligands differ for the length of the alkyl chain connecting the imidazole ring to the anilinoethylamide portion, spanning from three to six methylene units.Ind

an ethyl or a propyl chain is usually preferred, for melatonergic ligands the optimal c
ain length is four methylene units or longer.Compound 14a with the shortest spacer was totally inactive towards both MT1 and MT2 melatonin receptors, whereas it showed micromolar affinity for the H3 receptor.Elongation of the alkyl chain with a fourth methylene unit (14b) was not sufficient to achieve binding affinity towards melatonin receptors The first active bivalent ligand was obtained with a pentyl linker connecting the imidazole ring to the melatonergic fragment (14c).Indeed, although compound 14c is significantly less potent than melatonin towards MT1 and MT2 receptors, it is able to bind to both melatonin and H3 receptor subtypes with micromolar affinity.A further elongation of the linker yielded the second active derivative 14d, with improved binding affinity for melatonin receptors compared to 14c.When evaluated in the GTPγS assay, 14c behaved as an antagonist at both MT1 and MT2 receptors.

The low affinities of these compounds for melatonin receptors are likely to be ascribed to the presence of the imidazole ring.Indeed, compound 3 (Figure 1) displayed pKi = 8.93 for MT1 receptors and, more generally, in the series of phenyl-alkyloxy-anilinoethylamides alkyl chains longer than four methylene units provided compounds with good binding affinities [31].According to our docking hypothesis, the partial recovery of binding affinity observed with longer spacers could be due to the accommodation of the imidazole ring into a more peripheral, solvent-exposed regi n of the receptor.The limited binding affinity might be related to the presence, in this area, of some positively-charged amino acids belonging to extracellular loop 2 (e.g., R164 and R173 in the MT1 receptor) which could negatively interact with the imidazole ring (Figure 4).The presence of the imidazole ring is likely related to the antagonist behavior shown by 14d.It remains to be evaluated if replacement of this ring with other heterocycles, tolerated by the H3 receptor, could restore the agonist activity.

Alkyl chain length seems not to significantly influence binding affinity at the H3 receptor.The longest derivatives 14c-d are those with highest binding affinities, about 10 times lower than the reference imidazole-based H3 antagonist thioperamide (5).We may speculate that the anilinoethylamide portion undertakes some unfavorable interaction with the binding site which could contribute to the limited binding affinity observed for these compounds.


Experimental Section


General Experimental Procedures

1 H NMR (200 MHz) and 13 C NMR (50 MHz) spectra were recorded on a Bruke (Billerica, MA, USA) AVANCE 200 spectrometer, using CDCl3 as solvent.Chemical shifts (δ scale) are reported in parts per million (ppm) relative to the central peak of the solvent.Coupling constants (J values) are given in hertz (Hz).ESI-MS spectra were taken on a Waters (Milford, MA, USA) Micromass Zq instrument.Elemental analyses for C, H and N were performed on a Carlo Erba (Milan, Italy) analyzer, and the results are within 0.4% of the calculated

Synthesis of Unsaturated Esters 9a, 9c-d

Ge
eral Procedure: A solution of triethyl phosphonoacetate (0.2 mL, 1 mMol) in THF (6 mL) was added to an ice-cooled suspension of NaH (80% dispersed in mineral oil, 0.030 g, 1 mMol) in THF (6 mL) and the resulting mixture was stirred at room temperature for 30 min.A solution of the opportune aldehyde 8a, 8c-d [48] (0.9 mMol) in THF (6 mL) was added dropwise and the mixture was stirred for 16 h at room temperature.The reaction mixture was then poured into iced water and extracted 3× with diethyl ether; the organic phases were combined, dried (Na2SO4) and concentrated to give a crude residue which was purified by flash chromatography (silica gel, EtOAc as eluent).


(E)-Ethyl 3-(1-trityl-1H-imidazol-4-yl)acrylate

f Ester Derivatives 10a, 10c-d

General Proced
re: A solution of the suitable ethyl ester 9a, 9c-d (1 mMol) in

quenched by slow addition of a saturated aqueous solution of NaHCO3 and EtOAc.The mixtu
e was filtered on Celite, the filtrate was concentrated under re

ced pressure to afford a residue which was taken up in CH2Cl2 and washed with a saturat
d aqueous solution of NaHCO3.The organic phase was dried (Na2SO4) and concentrated in vacuo to give a crude product which was purified by flash chrom
3 (t, 1H, J = 7.0), 2.94 (s, 3H), 3.45 (m, 4H), 4.03 (t, 1H, J = 6.5), 6.10 (brs, 1H), 6.26-6.37 (m, 3H), 6.83 (s, 1H), 7.13 (dd, 1H, J1 = J2 = 8.0), 7.60 (s, 1H). 13 Prism (San Diego, CA, USA) GraphPad 2005, and converted into pKi values according to the Cheng-Prusoff equation [51].Data are expressed as mean ± SEM of three independent experiments.


Human Melatonin Receptors Binding Assay

Binding affinities were determined using 2-[ 125 I]iodomelatonin as the labeled ligand in competition experiments with cloned human MT1 and MT2 receptors expressed in NIH3T3 rat fibroblast cells.The characterization of NIH3T3 MT1 and MT2 cells was already described in detail [52,53].Membranes were incubated for 90 min at 37 °C in binding buffer (50 mM Tris/HCl, pH 7.4).The final membrane concentration was 5-10 μg of protein per tube.The membrane protein level was determined in accordance with a previously reported method [54].2-[ 125 I]Iodomelatonin (100 pM) and different concentrations of the tested compounds were incubated with the receptor preparation for 90 min at 37 °C.Nonspecific binding was assessed with 10 μM melatonin; IC50 values were determined by nonlinear fitting strategies with Prism (GraphPad SoftWare Inc., San Diego, CA, USA).The pKi values were calculated from the IC50 values according to the Cheng-Prusoff equation [51].The pKi values are the mean of at least three independent determinations performed in duplicate.

To determine the functional activity of compound 14d at MT1 and MT2 receptor subtypes, [ 35 S]GTPγS binding assays in NIH3T3 cells expressing human-cloned MT1 or MT2 receptors were performed.The amount of bound [ 35 S]GTPγS is proportional to the level of the analog-induced G protein activation and is related to the intrinsic activity of the compound under study.The detailed description and validation of this method were reported elsewhere [52,55].Membranes (15-25 μg of protein

00 mM), and M
Cl2 (3 mM).Nonspecific binding was defined using GTPγS (10 μM).In cell lines expressing human MT1 or MT2 receptors, melatonin produced concentration-dependent stimulation of basal [ 35 S]GTPγS binding with a maximal stimulation above basal levels of 370% and 250% in MT1 and MT2 receptors, respectively.Basal stimulation is the amount of [ 35 S]GTPγS specifically bound in the absence of compounds and was taken as 100%.The maximal G protein activation was measured in each experiment using melatonin (100 nM).Compounds were added at three different concentrations (one concentration equivalent to 100 nM melatonin, a second one, 10-fold smaller, and a third one, 10-fold larger), and the percent stimulation above basal was determined.The equivalent concentration was estimated on the basis of the ratio of the affinity of the test compound to that of melatonin.It was assumed that, at the equivalent concentration, the test compound occupies the same number of receptors as 100 nM melatonin.All of the measurements were performed in triplicate.The relative intrinsic activity (IAR) values were obtained by dividing the maximum ligand induced stimulation of [ 35 S]GTPγS binding by that of melatonin, as measured in the same experiment.


Docking Studies

All docking simulations were performed with Glide 5.7 [56]; Maestro 9.2 [57] was applied to prepare ligand structures and to refine protein-ligand complexes.


MT1 Receptor

A previously-reported MT1 receptor model [31] was taken as starting point for induced-fit docking (IFD) of compounds 14a-d.An initial softened-potential docking run was performed applying van der Waals radii scaling of 0.7 and 0.5 on protein and ligand non-polar atoms, respectively.Amino acids hampering the accommodation of the imidazole ring of 14a-d, i.e., Q101 and Q169, were temporarily mutated to alanines.Energy grids generated for the initial softened-potential docking were centered in the putative binding site of the receptor, setting enclosing and bounding boxes to default dimensions.During flexible dock

g runs, two
ydrogen bond constraints were applied between Y187 hydroxyl group and the phenolic oxygen of the ligand, and between Y285 hydroxyl group and the amide oxygen of the ligand, to reproduce the main polar interactions previously proposed for melatonergic ligands.Ligand docking was performed in standard precision (SP) mode, collecting fifty poses.The resulting ligand-receptor complexes were then submitted to a protein structure refinement stage; once amino acid side chains that had previously been removed were re-introduced, residues within a shell of 5 Å around any ligand pose were refined by a side chain conformational search, followed by energy minimization of the residues and the ligand molecule.In the final docking stage, each ligand structure obtained at the end of the protein structure refinement was energetically optimized (refined) in the field of the receptor and subsequently scored using default Glide settings.The final ligand-protein complexes were ranked according to their IFD score, a composite score that

ccounts for
igand-receptor interaction energy, receptor strain and solvation terms.The best-ranked MT1-14a-d complexes were minimized applying the OPLS2005 force field [58] to a convergence threshold of 0.05 kJ•mol −1 •Å −1 .During this minimization procedure, the ligand and residues within 8 Å from the ligand were free to move, while all other atoms were retained fixed.


H3 Receptor

A previously-reported model of the histamine H3 receptor [45] was used as reference structure for docking studies of compounds 14a-d.Glide grids were centered in the putative binding site of the receptor, located between D114 and E206, setting the dimension of enclosing and bounding boxes to default values.Compounds 14a-d were flexibly docked within the H3 receptor binding cavity in SP mode, applying a van der Waals radii scaling of 0.7 on ligand non-polar atoms.During docking ru

, a hydrogen bond con
traint was applied between E206 and the imidazole ring of the ligands.Fifty poses were collected for each ligand and ranked according to their Emodel value.The best-scored ligand conformations were merged into the H3 receptor model and the resulting complexes were minimized applying the OPLS2005 force field to an energy gradient of 0.05 kJ•mol −1 Å −1 .During this minimization phase,