Multitargeting Histamine H3 Receptor Ligands among Acetyl- and Propionyl-Phenoxyalkyl Derivatives

Alzheimer’s disease (AD) is a neurodegenerative disorder, for which there is no effective cure. Current drugs only slow down the course of the disease, and, therefore, there is an urgent need to find effective therapies that not only treat, but also prevent it. Acetylcholinesterase inhibitors (AChEIs), among others, have been used for years to treat AD. Histamine H3 receptors (H3Rs) antagonists/inverse agonists are indicated for CNS diseases. Combining AChEIs with H3R antagonism in one structure could bring a beneficial therapeutic effect. The aim of this study was to find new multitargetting ligands. Thus, continuing our previous research, acetyl- and propionyl-phenoxy-pentyl(-hexyl) derivatives were designed. These compounds were tested for their affinity to human H3Rs, as well as their ability to inhibit cholinesterases (acetyl- and butyrylcholinesterases) and, additionally, human monoamine oxidase B (MAO B). Furthermore, for the selected active compounds, their toxicity towards HepG2 or SH-SY5Y cells was evaluated. The results showed that compounds 16 (1-(4-((5-(azepan-1-yl)pentyl)oxy)phenyl)propan-1-one) and 17 (1-(4-((6-(azepan-1-yl)hexyl)oxy)phenyl)propan-1-one) are the most promising, with a high affinity for human H3Rs (Ki: 30 nM and 42 nM, respectively), a good ability to inhibit cholinesterases (16: AChE IC50 = 3.60 µM, BuChE IC50 = 0.55 µM; 17: AChE IC50 = 1.06 µM, BuChE IC50 = 2.86 µM), and lack of cell toxicity up to 50 µM.


Introduction
The pathological cause of Alzheimer's disease (AD) is neuronal death, which leads to the disruption of connections between neurons and the inhibition of the processing and conduction of electrical impulses. This is caused by the accumulation of protein deposits in neuronal tissue: ß-amyloid or tau protein [1,2]. Recently, it has been indicated that αsynuclein may also play an important role in the pathogenesis of AD [3]. Presumably, there is a feedback loop between all these proteins. β-Amyloid increases GSK-β kinase levels, which induces tau protein phosphorylation and stimulates α-synuclein production. All of this can lead to the aggregation of both the β-amyloid and the tau protein. However, as AD is a disease with a complex a etiology, in addition to the factors discussed above, many more play a role in its development and course [1]. Among these, impaired cholinergic transmission is discussed as an important contributor.
The theory of the involvement of acetylcholine (ACh) in the development of AD has been known since the 1970s [4], and is based on the observation of reduced levels of this Figure 1. Structures of the most promising histamine H3R ligands with cholinesterase inhibitory activity from the last 3 years. hH3R-human H3R; eeAChE-AChE from electric eel; eqBuChE-BuChE from equine serum; mbAChE-AChE extracted from mouse brains; mpAChE-AChE extracted from mouse plasma. a Data from [17]; b data from [18]; c data from [19]; d data from [20]; e data from [21]; f data from [22].

Design of Compounds
Three years ago, through in silico studies (molecular modeling and docking to cholinesterases), xanthones were found to be potential ligands for H3R, cholinesterases and MAO B [20]. The designed compounds aligned with the general construction pattern of previously suggested H3R antagonists/inverse agonists ( Figure 2) [27].  Structures of the most promising histamine H 3 R ligands with cholinesterase inhibitory activity from the last 3 years. hH 3 R-human H 3 R; eeAChE-AChE from electric eel; eqBuChE-BuChE from equine serum; mbAChE-AChE extracted from mouse brains; mpAChE-AChE extracted from mouse plasma. a Data from [17]; b data from [18]; c data from [19]; d data from [20]; e data from [21]; f data from [22].

Design of Compounds
Three years ago, through in silico studies (molecular modeling and docking to cholinesterases), xanthones were found to be potential ligands for H 3 R, cholinesterases and MAO B [20]. The designed compounds aligned with the general construction pattern of previously suggested H 3 R antagonists/inverse agonists ( Figure 2) [27].
Then, xanthones were synthesized, and tested in vitro, in appropriate assays that confirmed their biological activity against these targets, which have been previously described by Łażewska et al. [20]. The most promising structure found in that work, compound IV (Figure 1), was selected as the lead 1 for further modifications, in order to check which fragments of the xanthone moiety affect their pharmacological activity. As a first step, oxygen was removed from this molecule to obtain benzophenone derivatives that have recently been described [26]. The direction of the modifications is shown in Figure 3. Among the benzophenone derivatives, there were structures that showed good activity as multitarget ligands (e.g., the structures in Figure 3). The aim of the current work was to test whether further structural modifications (reduction of substituent size), i.e., replacing the phenyl moiety with an ethyl or a methyl group, would improve or worsen interactions with selected biological targets. In the search for CNS-active compounds (but not exclusively), it is important that compounds have moderate lipophilicity (log P in the range of one to three), as this favorably affects ADMET parameters and increases the compound's chances of being a drug [28]. Replacing the phenyl ring with an alkyl (methyl, ethyl) substituent provides a reduction in the lipophilicity of the compounds, which was confirmed by preliminary calculations performed using the SwissADME server [29] (data in Supplementary Materials).

Design of Compounds
Three years ago, through in silico studies (molecular modeling and docking to cholinesterases), xanthones were found to be potential ligands for H3R, cholinesterases and MAO B [20]. The designed compounds aligned with the general construction pattern of previously suggested H3R antagonists/inverse agonists ( Figure 2) [27].  Then, xanthones were synthesized, and tested in vitro, in appropriate assays that confirmed their biological activity against these targets, which have been previously described by Łażewska et al. [20]. The most promising structure found in that work, compound IV (Figure 1), was selected as the lead 1 for further modifications, in order to check which fragments of the xanthone moiety affect their pharmacological activity. As a first step, oxygen was removed from this molecule to obtain benzophenone derivatives that have recently been described [26]. The direction of the modifications is shown in Figure 3. Among the benzophenone derivatives, there were structures that showed good activity as multitarget ligands (e.g., the structures in Figure 3). The aim of the current work was to test whether further structural modifications (reduction of substituent size), i.e., replacing the phenyl moiety with an ethyl or a methyl group, would improve or worsen interactions with selected biological targets. In the search for CNS-active compounds (but not exclusively), it is important that compounds have moderate lipophilicity (log P in the range of one to three), as this favorably affects ADMET parameters and increases the compound's chances of being a drug [28]. Replacing the phenyl ring with an alkyl (methyl, ethyl) substituent provides a reduction in the lipophilicity of the compounds, which was confirmed by preliminary calculations performed using the SwissADME server [29] (data in Supplementary Materials).  [20]; b data from [26].

Synthesis of Compounds 2-17
In the first stage, it was necessary to obtain phenoxyalkylbromides (1a-1d; Scheme 1) by O-alkylation of the corresponding phenols with 1,5-dibromopentane or 1,6dibromohexane. The reactions were performed in freshly prepared sodium propanolate. The resulting bromides were then subjected to N-alkylation with the corresponding amines (piperidines or azepane; Scheme 1). The reactions were carried out in the mixture  [20]; b data from [26].

Synthesis of Compounds 2-17
In the first stage, it was necessary to obtain phenoxyalkylbromides (1a-1d; Scheme 1) by O-alkylation of the corresponding phenols with 1,5-dibromopentane or 1,6-dibromohexane. The reactions were performed in freshly prepared sodium propanolate. The resulting bromides were then subjected to N-alkylation with the corresponding amines (piperidines or azepane; Scheme 1). The reactions were carried out in the mixture of ethanol and water (21:4), in the presence of potassium carbonate and catalytic amounts of potassium iodide. The final products were purified by extraction. The oily free bases were transformed into oxalic acid salts. The purity and identity of the products were confirmed by spectroscopic methods ( 1 H and 13 C NMR, LC-MS) and elemental analysis. Spectra data are presented in the Supplementary Materials.

Human Histamine H3 Receptor Affinity
The affinity for hH3R was assessed in vitro in radioligand binding assays, using membrane preparations of HEK293 cells stably expressing hH3R, while [ 3 H]-N αmethylhistamine was used as a radioligand. The exact procedure was described previously by Kottke et al. [30]. Novel compounds were tested as oxalic acid salts. All compounds (except for 9 and 10) showed very good affinities for these receptors, with Ki values below 100 nM ( Table 1). The activity depended on all three variables, i.e., the presence of an amine group, the chain length (five or six carbons), and the type of substituent in the phenyl ring. With regard to the change in the amine moiety, it can be observed that, regardless of the length of the chain and the type of R substituent, the activity of the derivatives was arranged in the following order: 3-methylpiperidine ≥ piperidine ≥ azepane >> 4-metylpiperidine derivatives. Regarding the length of the carbon chain, compounds bearing the pentylene chain are characterized by a higher binding affinity to the receptor than hexylene derivatives (exception: 10). The change of the acyl group to propionyl one resulted in an almost twofold increase in affinity to hH3R receptors (exception 4). The highest affinity for hH3R among all compounds was observed in compound 8, with a Ki of 12 nM. Seven other compounds (2, 4, 5, 6, 9, 16 and 17) had also very high affinities, with Ki values < 50 nM.

Human Histamine H 3 Receptor Affinity
The affinity for hH 3 R was assessed in vitro in radioligand binding assays, using membrane preparations of HEK293 cells stably expressing hH 3 R, while [ 3 H]-N α -methylhistamine was used as a radioligand. The exact procedure was described previously by Kottke et al. [30]. Novel compounds were tested as oxalic acid salts. All compounds (except for 9 and 10) showed very good affinities for these receptors, with K i values below 100 nM ( Table 1). The activity depended on all three variables, i.e., the presence of an amine group, the chain length (five or six carbons), and the type of substituent in the phenyl ring. With regard to the change in the amine moiety, it can be observed that, regardless of the length of the chain and the type of R substituent, the activity of the derivatives was arranged in the following order: 3-methylpiperidine ≥ piperidine ≥ azepane >> 4-metylpiperidine derivatives. Regarding the length of the carbon chain, compounds bearing the pentylene chain are characterized by a higher binding affinity to the receptor than hexylene derivatives (exception: 10). The change of the acyl group to propionyl one resulted in an almost twofold increase in affinity to hH 3 R receptors (exception 4). The highest affinity for hH 3 R among all compounds was observed in compound 8, with a K i of 12 nM. Seven other compounds (2, 4, 5, 6, 9, 16 and 17) had also very high affinities, with K i values < 50 nM. potency of one micromole (IC50 = 1.06 µM).

Kinetic Studies of the Inhibition of eeAChE and eqBuChE
For the most active compounds, 17 (AChE inhibitor) and 16 (BuChE inhibitor), tests were carried out to determine the type of inhibition exhibited for the appropriate enzymes. The Michaelis-Menten equation was used to calculate the maximum velocity (Vmax) and the Michaelis constant (Km). For the most active compounds, 17 (AChE inhibitor) and 16 (BuChE inhibitor), tests were carried out to determine the type of inhibition exhibited for the appropriate enzymes. The Michaelis-Menten equation was used to calculate the maximum velocity (Vmax) and the Michaelis constant (Km). Kinetic Studies of the Inhibition of eeAChE and eqBuChE For the most active compounds, 17 (AChE inhibitor) and 16 (BuChE inhibitor), tests were carried out to determine the type of inhibition exhibited for the appropriate enzymes. The Michaelis-Menten equation was used to calculate the maximum velocity (Vmax) and the Michaelis constant (Km). Kinetic Studies of the Inhibition of eeAChE and eqBuChE For the most active compounds, 17 (AChE inhibitor) and 16 (BuChE inhibitor), tests were carried out to determine the type of inhibition exhibited for the appropriate enzymes. The Michaelis-Menten equation was used to calculate the maximum velocity (Vmax) and the Michaelis constant (Km).

Kinetic Studies of the Inhibition of eeAChE and eqBuChE
For the most active compounds, 17 (AChE inhibitor) and 16 (BuChE inhibitor), tests were carried out to determine the type of inhibition exhibited for the appropriate enzymes. The Michaelis-Menten equation was used to calculate the maximum velocity (Vmax) and the Michaelis constant (Km).

Cholinesterase Inhibitory Activity
All compounds were tested at a concentration of 10 µM in a modified colorimetric method first described by Ellman et al. in the 1960s [19,31], using AChE from electrophorus electricus (eeAChE) and BuChE from equine serum (eqBuChE). The compounds (5-8, 11-13,  16 and 17) that showed inhibitions of greater than 70% were selected for further studies, in order to obtain IC 50 values. All results are collected in Table 1. The IC 50 values are in the low micromolar range (<5 µM). In general, compounds that showed activity against both cholinoesterases inhibited BuChE to a greater extent than AChE (with the exception of 17). However, we did not observe any difference in inhibition potency. No correlation was observed between inhibitory activity and hH 3 R affinity. The most interesting compounds were among the azepane derivatives. Compound 16 is the most potent BuChE inhibitor in the whole series, with inhibitory activity in the sub-micromolar range (IC 50 = 0.55 µM). In contrast, compound 17 shows the highest inhibition of AChE in this series, with a potency of one micromole (IC 50 = 1.06 µM).

Kinetic Studies of the Inhibition of eeAChE and eqBuChE
For the most active compounds, 17 (AChE inhibitor) and 16 (BuChE inhibitor), tests were carried out to determine the type of inhibition exhibited for the appropriate enzymes. The Michaelis-Menten equation was used to calculate the maximum velocity (V max ) and the Michaelis constant (K m ).
Lineweaver-Burk plots obtained for 17 ( Figure 4A) showed a series of lines converging at the same point near the x-axis (1/[ATC], which confirmed the non-competitive (mixed) type of eeAChE inhibition, as V max decreased with increasing concentrations of inhibitor. mean values of two independent experiments. g Data from Ref. [19]. h Not tested. i IC50 in nM; data from ref. [20]. Lineweaver-Burk plots obtained for 17 ( Figure 4A) showed a series of lines converging at the same point near the x-axis (1/[ATC], which confirmed the noncompetitive (mixed) type of eeAChE inhibition, as Vmax decreased with increasing concentrations of inhibitor.
Furthermore, Cornish-Bowden plots (S/V vs. I) obtained for 17 ( Figures 4B) showed a series of lines converging at the same point over the x-axis, thus confirming a mixed mechanism of cholinesterase inhibition. The same situation was observed for compound 16, for which the type of BuChE inhibition was investigated and evaluated in the Lineweaver-Burk and Cornish-Bowden plots ( Figure 5). The type of line intersection and the location of this intersection confirms the mixed type of inhibition.

Monoamine Oxidase B Inhibitory Activity
The ability to inhibit MAO B was estimated by a fluorescence assay, as previously described [32]. The screening assay was carried out at a concentration of 1 µM to determine the percentage inhibition of the enzyme, as compared to reference inhibitors (i.e., pargiline, rasagiline and safinamide). All compounds showed weak percentage of inhibition (<50%) and were not selected for further testing (Table 1). Furthermore, Cornish-Bowden plots (S/V vs. I) obtained for 17 ( Figure 4B) showed a series of lines converging at the same point over the x-axis, thus confirming a mixed mechanism of cholinesterase inhibition. The same situation was observed for compound 16, for which the type of BuChE inhibition was investigated and evaluated in the Lineweaver-Burk and Cornish-Bowden plots ( Figure 5). The type of line intersection and the location of this intersection confirms the mixed type of inhibition.
Molecules 2023, 28, x FOR PEER REVIEW 7 of 16 mean values of two independent experiments. g Data from Ref. [19]. h Not tested. i IC50 in nM; data from ref. [20].
Lineweaver-Burk plots obtained for 17 ( Figure 4A) showed a series of lines converging at the same point near the x-axis (1/[ATC], which confirmed the noncompetitive (mixed) type of eeAChE inhibition, as Vmax decreased with increasing concentrations of inhibitor.
Furthermore, Cornish-Bowden plots (S/V vs. I) obtained for 17 ( Figures 4B) showed a series of lines converging at the same point over the x-axis, thus confirming a mixed mechanism of cholinesterase inhibition. The same situation was observed for compound 16, for which the type of BuChE inhibition was investigated and evaluated in the Lineweaver-Burk and Cornish-Bowden plots ( Figure 5). The type of line intersection and the location of this intersection confirms the mixed type of inhibition.

Monoamine Oxidase B Inhibitory Activity
The ability to inhibit MAO B was estimated by a fluorescence assay, as previously described [32]. The screening assay was carried out at a concentration of 1 µM to determine the percentage inhibition of the enzyme, as compared to reference inhibitors (i.e., pargiline, rasagiline and safinamide). All compounds showed weak percentage of inhibition (<50%) and were not selected for further testing (Table 1).

Monoamine Oxidase B Inhibitory Activity
The ability to inhibit MAO B was estimated by a fluorescence assay, as previously described [32]. The screening assay was carried out at a concentration of 1 µM to determine the percentage inhibition of the enzyme, as compared to reference inhibitors (i.e., pargiline, rasagiline and safinamide). All compounds showed weak percentage of inhibition (<50%) and were not selected for further testing (Table 1).

Cytotoxicity Studies of Selected Compounds
Compounds showing activity against cholinesterases (AChE and/or BuChE) were selected for preliminary evaluation of their toxic effects (i.e., compounds 5-8, 11-13, 16 and 17) in the MTS assay. The studies were conducted on two cell lines: HepG2 and SH-SY5Y. To assess viability, increasing concentrations of the compounds tested (from 0.78 µM to 50 µM) were incubated with their respective cell lines (HepG2 or SH-SY5Y) for 48 h and 24 h, respectively. After the indicated time, the MTS reagent was added directly to the cultured cells for 1 h, and then the absorbance at 490 nm was read. The study showed that both the tested lines (HepG2 and SH-SY5Y) remained more than 50% viable the highest doses tested. The results are presented in Table 2 and Figures S1 and S2 (Supplementary Materials).

Discussion
Continuing our previous works on searching for multi-targeted H 3 R ligands, we obtained a series of sixteen compounds designed on the basis of our previous studies [20,26]. We conducted pharmacological tests of acetyl-and propionyl-phenoxy-pentyl (-hexyl) amines, and confirmed the high affinity of these compounds for hH 3 R. The determined K i values for all compounds are in the nanomolar range (K i < 120 nM). This is in line with our expectations, as these structures fit the proposed pharmacophore model for H 3 R antagonists/reverse agonists [27]. This pharmacophore has the following elements: a basic center, a linker, a central core, and an arbitrary region with high variability, which are required in a molecule for activity. The investigated compounds have all these elements. Among all the used amines (azepane, 3-methylpiperidine, 4-methylpiperidine and piperidine), the least active compounds were the 4-methylpiperidines, and such a correlation was also observed in previous works, e.g., [33]. In regard to the most active compounds, it is difficult to say, without a doubt, as to which amine's presence is most beneficial. The activities of the derivatives with the other three amines are usually comparable, although compounds with a 3-methylpiperidine moiety are often the most active.
As for the results obtained for cholinesterase inhibition, they are also unclear. No clear correlation was observed between the structure of the compounds and the observed inhibitory potency. Only nine of sixteen compounds showed enough activity in the prescreening (>70% at 10 µM) to qualify them for testing at other concentrations to determine the IC 50 values. All these compounds showed inhibitory potencies in the low micromolar concentration range (0.55 µM ≤ IC 50 ≤ 4.79 µM). Some of them inhibited, at micromolar concentrations, both cholinesterases (four compounds); others only inhibited one cholinesterase (five compounds), whereas a low percentage of inhibition (<70%) was observed for seven compounds. These results showed, in comparison with recently described benzophenone derivatives [26], that the exchange of the phenyl substituent for an ethyl substituent and, in particular, a methyl substituent, did not favorably influence cholinesterase inhibition. Almost all of the thirty-four previously described benzophenones (with the exception of three compounds) inhibited the cholinesterases in the micromolar range (0.17 µM ≤ IC 50 ≤ 7.75 µM) [26], while now only nine compounds showed activity against cholinesterases. Comparing the results for the lead 2 ( Figure 3) directly with analogous compounds 14 (the methyl analogue) and 16 (the ethyl analogue), it was seen that the exchange of the phenyl for the ethyl substituent, in this case, did not affect the activity against all three tested biological targets, i.e., (hH 3 R, AChE and BuChE), very significantly. The opposite result was observed for the methyl substituent (14). Shortening the chain to the methyl substituent caused a decrease in activity against both cholinesterases (only 63% of inhibition). Among all nine compounds, compound 16 had the highest BuChE inhibitory activity (IC 50 = 0.55 µM), whereas compound 17 showed the highest inhibition of AChE (IC 50 = 1.06 µM). Both these compounds had a hexyl carbon linker.
None of the compounds tested showed a promising inhibition of hMAO B. All compounds that showed inhibitory activity against cholinesterases were selected for further in vitro studies, in order to evaluate potential toxic effects on HepG2 cells and SH-SY5Y cells in the MTS assay. Human liver carcinoma HepG2 cell lines are the most popular lines to assess the risk of hepatotoxicity [34]. The human neuroblastoma SH-SY5Y lines are used in experimental models of AD to assess intracellular factors that lead to AD (e.g., tau-related pathology [35]) or the neuroprotection ability of ligands. The results of these studies showed that the compounds had very low toxicity. Specific IC 50 values calculated with non-linear regression, and fitted to a sigmoidal dose-response curve, were higher than 90 µM. The only exception was compound 17, which exhibited some toxicity against HepG2 cells, with an IC 50 of 60 µM.
In summary, only four multitarget compounds (5, 8, 16 and 17) with simultaneous activity towards H 3 R and cholinesterases were obtained in the present work. All of them are propionyl phenoxy derivatives. Unfortunately, pharmacological studies have shown that these compounds have a much greater affinity for hH 3 R (effects in the nanomolar range) than against cholinesterases (effects in the micromolar range).

Synthesis of Compounds
Reagents and solvents were purchased from Merck and Alfa Aesar. The course of the reaction was controlled using TLC (Merck silica 60F aluminum sheets; solvent: methylene chloride or methylene chloride: methanol 9:1). Spots were visualized under a UV lamp and/or stained with Dragendorf's reagent. NMR spectra ( 1 H and 13 C) were obtained in DMSO-d 6 on the following instruments: ( 1 H) -Mercury 300 MHz PFG spectrometer (Varian, Palo Alto, CA, USA) and ( 13 C)-FTNMR 500 MHz spectrometer (Joel Ltd., Akishima, Tokyo, Japan). Chemical shifts (δ) are given in ppm with respect to the solvent signal. Data are reported as follows: multiplicity (br s, broad singlet; d, dublet; m, multiplet; q, quartet; s, singlet; t, triplet), coupling constants (J) in Hz, and the number of protons. The mass spectrum (LC-MS) was recorded on a Waters TQ detector mass spectrometer (Water Corporation, Milford, CT, USA). Retention times (t R ) are given in minutes. All compounds (except 6, 11 and 17) showed a purity of >95%. Elemental analysis was performed on an Elemental Analyser Vario El-III (Hanau, Germany). The results are in agreement with the theoretical values, within ±0.4%. Melting points (Mp) were determined on a MEL-TEMP melting point apparatus II (LD Inc., Long Beach, CA, USA), and are uncorrected.