Benzophenone Derivatives with Histamine H3 Receptor Affinity and Cholinesterase Inhibitory Potency as Multitarget-Directed Ligands for Possible Therapy of Alzheimer’s Disease

The multitarget-directed ligands demonstrating affinity to histamine H3 receptor and additional cholinesterase inhibitory potency represent a promising strategy for research into the effective treatment of Alzheimer’s disease. In this study, a novel series of benzophenone derivatives was designed and synthesized. Among these derivatives, we identified compound 6 with a high affinity for H3R (Ki = 8 nM) and significant inhibitory activity toward BuChE (IC50 = 172 nM and 1.16 µM for eqBuChE and hBuChE, respectively). Further in vitro studies revealed that compound 6 (4-fluorophenyl) (4-((5-(piperidin-1-yl)pentyl)oxy)phenyl)methanone) displays moderate metabolic stability in mouse liver microsomes, good permeability with a permeability coefficient value (Pe) of 6.3 × 10−6 cm/s, and its safety was confirmed in terms of hepatotoxicity in the HepG2 cell line. Therefore, we investigated the in vivo activity of compound 6 in the Passive Avoidance Test and the Formalin Test. While compound 6 did not show a statistically significant influence on memory and learning, it showed analgesic properties in both acute (ED50 = 20.9 mg/kg) and inflammatory (ED50 = 17.5 mg/kg) pain.


Introduction
Memory and other cognitive functions such as learning, perception, attention, or decision-making play a crucial role in the everyday life of humans. They are driven by the complex, synergistic interactions between diverse neurotransmitters, i.e., acetylcholine (ACh), histamine, serotonin, glutamate, noradrenaline and many others. The dysfunctions of central nervous system (CNS) mediators and lack of synergism in their cooperation are proven to be involved in the progression of cognitive decline, often resulting in degenerative disorders such as Alzheimer's disease (AD) [1]. Therefore, a number of therapeutic possibilities for AD therapy are concentrated on neurotransmission enhancement. In 2021 Following the general structural requirements for H3R, cholinesterases and affinity, the design of novel MTDLs was rationally performed. Novel derivativ based on the substituted benzophenone moiety acting as an arbitrary lipophilic which was linked by an alkoxyl chain to a (methyl)piperidine or an azepane basic The benzophenone scaffold was a structural analog of the xanthone moiety pr compound I [32] (Figure 1).  Initially, we evaluated the binding mode of the designed compounds with H 3 R, using H 3 R homology model, and with AChE and BuChE by docking to the enzyme active sites (electric ray AChE and human BuChE). Then, the biological activity of the synthesized novel MTDLs toward selected targets was assessed in vitro. The designed ligands were screened for their potency as AChE and BuChE inhibitors as well as for affinities to H 3 R. Due to the series origin, the activity toward MAO B was also tested. The most potent compounds were evaluated for their K i (H 3 R) or IC 50 values (AChE, BuChE). In further in vitro kinetic studies, selected compounds were investigated to determine the mechanism of AChE and BuChE inhibition. Additionally, regarding the MAO B inhibitory activity of the lead I and its derivatives, the binding mode of the obtained compounds within the human MAO B (hMAO B) active site has been investigated.
Moreover, for the selected lead compound, the metabolic stability, neuroprotective properties, and potential hepatotoxicity were studied; the permeability of the blood-brain barrier (BBB) in vitro was also determined. Finally, to confirm the therapeutic potential of the lead compound on memory and learning, and its potential analgesic properties, in vivo pharmacological tests were carried out.

Rational Design
Following the general structural requirements for H 3 R, cholinesterases and MAO B affinity, the design of novel MTDLs was rationally performed. Novel derivatives were based on the substituted benzophenone moiety acting as an arbitrary lipophilic region, which was linked by an alkoxyl chain to a (methyl)piperidine or an azepane basic residue. The benzophenone scaffold was a structural analog of the xanthone moiety present in compound I [32] (Figure 1).

Synthesis of Designed Benzophenones
Synthesis of final compounds  began with the O-alkylation of a suitabl phenone derivative with dibromopentane or dibromohexane in acetone (Scheme obtained bromides (1a-1j) were further used in the N-alkylation of corresponding (piperidines or azepane) in the mixture of ethanol-water (4:1) and the presence o [32]. The purity and identity of final compounds were confirmed by spectral ( 1 H NMR, LC-MS) and elemental analysis.   Xanthone, benzophenone and basic tertiary amine motifs present in compounds developed as potential AD therapeutics [13,32,36,37,40].

Synthesis of Designed Benzophenones
Synthesis of final compounds  began with the O-alkylation of a suitable benzophenone derivative with dibromopentane or dibromohexane in acetone (Scheme 1). The obtained bromides (1a-1j) were further used in the N-alkylation of corresponding amines (piperidines or azepane) in the mixture of ethanol-water (4:1) and the presence of K 2 CO 3 [32]. The purity and identity of final compounds were confirmed by spectral ( 1 H and 13 C NMR, LC-MS) and elemental analysis. Xanthone, as well as benzophenone fragments, were found to be pharmacophores for the H3R antagonism/inverse agonism [31,32]. Both elements are also important for cholinesterases and/or MAO B inhibition [36][37][38][39][40] (Figure 2). Additionally, basic cyclic tertiary amine moieties with an alkyl linker along with a lipophilic group are known structural characteristics present in MTDLs with inhibitory potency toward AChE and/or BuChE, and with H3R affinity [13,40] (Figure 2). Figure 2. Xanthone, benzophenone and basic tertiary amine motifs present in compounds developed as potential AD therapeutics [13,32,36,37,40].

Synthesis of Designed Benzophenones
Synthesis of final compounds  began with the O-alkylation of a suitable benzophenone derivative with dibromopentane or dibromohexane in acetone (Scheme 1). The obtained bromides (1a-1j) were further used in the N-alkylation of corresponding amines (piperidines or azepane) in the mixture of ethanol-water (4:1) and the presence of K2CO3 [32]. The purity and identity of final compounds were confirmed by spectral ( 1 H and 13 C NMR, LC-MS) and elemental analysis.  Final compounds were tested for their H 3 R affinity in a radioligand binding assay using [ 3 H]N α -methylhistamine as radioligand in HEK293 cells stably expressing hH 3 R [25]. Results are collected in Table 1. All the compounds revealed high affinities for H 3 R, with K i values ranging from 8 nM (lead 6) to 371 nM (2). The position of the alkoxyl chain in the benzophenone scaffold was important for the effect at the H 3 R. Para-substituted derivatives were generally the most potent. Their ortho and meta-analogs presented weaker H 3 R affinity; however, the latter were usually slightly better than the former, similarly in both subgroups of five and six carbon atoms in the alkoxyl chain (e.g., 4 vs. 3 vs. 2 and 9 vs. 8 vs. 7; 14 vs. 13 vs. 12 and 19 vs. 18 vs. 17). The number of methylene groups in the ether linker affected H 3 R affinity. Six carbon atom linker was preferable for meta and ortho-substituted benzophenones (e.g., 22 vs. 24 and 26 vs. 31). In turn, five carbon atoms in the alkoxyl part were usually better in the case of para-substituted derivatives, especially with additional halogen substituents in the second phenyl ring present in the benzophenone scaffold (e.g., 6 vs. 11, 16 vs. 21 and 29 vs. 34). Finally, it seemed that the kind of basic tertiary amine moieties (in the case of target compounds-the alicyclic ones) did not significantly influence hH 3 R affinities, with similar K i values obtained for analogs from all four subgroups (piperidine, 3-or 4-methylpiperidines and azepane). Among all the target compounds, 4-F substituted benzophenones with pentyloxyl linker presented the most promising activity, with low nanomolar K i values, i.e., 8 nM (lead 6), 12 nM (16) and 13 nM (30). The H 3 R affinity of 6 was comparable with the reference pitolisant.

Histamine H3 Receptor Affinity
Final compounds were tested for their H3R affinity in a radioligand binding assay using [ 3 H]N α -methylhistamine as radioligand in HEK293 cells stably expressing hH3R [25]. Results are collected in Table 1. All the compounds revealed high affinities for H3R, with Ki values ranging from 8 nM (lead 6) to 371 nM (2). The position of the alkoxyl chain in the benzophenone scaffold was important for the effect at the H3R. Para-substituted derivatives were generally the most potent. Their ortho and meta-analogs presented weaker H3R affinity; however, the latter were usually slightly better than the former, similarly in both subgroups of five and six carbon atoms in the alkoxyl chain (e.g., 4 vs. 3 vs. 2 and 9 vs. 8 vs. 7; 14 vs. 13 vs. 12 and 19 vs. 18 vs. 17). The number of methylene groups in the ether linker affected H3R affinity. Six carbon atom linker was preferable for meta and orthosubstituted benzophenones (e.g., 22 vs. 24 and 26 vs. 31). In turn, five carbon atoms in the alkoxyl part were usually better in the case of para-substituted derivatives, especially with additional halogen substituents in the second phenyl ring present in the benzophenone scaffold (e.g., 6 vs. 11, 16 vs. 21 and 29 vs. 34). Finally, it seemed that the kind of basic tertiary amine moieties (in the case of target compounds-the alicyclic ones) did not significantly influence hH3R affinities, with similar Ki values obtained for analogs from all four subgroups (piperidine, 3-or 4-methylpiperidines and azepane). Among all the target compounds, 4-F substituted benzophenones with pentyloxyl linker presented the most promising activity, with low nanomolar Ki values, i.e., 8 nM (lead 6), 12 nM (16) and 13 nM (30). The H3R affinity of 6 was comparable with the reference pitolisant. Final compounds were tested for their H3R affinity in a radioligand binding assay using [ 3 H]N α -methylhistamine as radioligand in HEK293 cells stably expressing hH3R [25]. Results are collected in Table 1. All the compounds revealed high affinities for H3R, with Ki values ranging from 8 nM (lead 6) to 371 nM (2). The position of the alkoxyl chain in the benzophenone scaffold was important for the effect at the H3R. Para-substituted derivatives were generally the most potent. Their ortho and meta-analogs presented weaker H3R affinity; however, the latter were usually slightly better than the former, similarly in both subgroups of five and six carbon atoms in the alkoxyl chain (e.g., 4 vs. 3 vs. 2 and 9 vs. 8 vs. 7; 14 vs. 13 vs. 12 and 19 vs. 18 vs. 17). The number of methylene groups in the ether linker affected H3R affinity. Six carbon atom linker was preferable for meta and orthosubstituted benzophenones (e.g., 22 vs. 24 and 26 vs. 31). In turn, five carbon atoms in the alkoxyl part were usually better in the case of para-substituted derivatives, especially with additional halogen substituents in the second phenyl ring present in the benzophenone scaffold (e.g., 6 vs. 11, 16 vs. 21 and 29 vs. 34). Finally, it seemed that the kind of basic tertiary amine moieties (in the case of target compounds-the alicyclic ones) did not significantly influence hH3R affinities, with similar Ki values obtained for analogs from all four subgroups (piperidine, 3-or 4-methylpiperidines and azepane). Among all the target compounds, 4-F substituted benzophenones with pentyloxyl linker presented the most promising activity, with low nanomolar Ki values, i.e., 8 nM (lead 6), 12 nM (16) and 13 nM (30). The H3R affinity of 6 was comparable with the reference pitolisant. Final compounds were tested for their H3R affinity in a radioligand binding assay using [ 3 H]N α -methylhistamine as radioligand in HEK293 cells stably expressing hH3R [25]. Results are collected in Table 1. All the compounds revealed high affinities for H3R, with Ki values ranging from 8 nM (lead 6) to 371 nM (2). The position of the alkoxyl chain in the benzophenone scaffold was important for the effect at the H3R. Para-substituted derivatives were generally the most potent. Their ortho and meta-analogs presented weaker H3R affinity; however, the latter were usually slightly better than the former, similarly in both subgroups of five and six carbon atoms in the alkoxyl chain (e.g., 4 vs. 3 vs. 2 and 9 vs. 8 vs. 7; 14 vs. 13 vs. 12 and 19 vs. 18 vs. 17). The number of methylene groups in the ether linker affected H3R affinity. Six carbon atom linker was preferable for meta and orthosubstituted benzophenones (e.g., 22 vs. 24 and 26 vs. 31). In turn, five carbon atoms in the alkoxyl part were usually better in the case of para-substituted derivatives, especially with additional halogen substituents in the second phenyl ring present in the benzophenone scaffold (e.g., 6 vs. 11, 16 vs. 21 and 29 vs. 34). Finally, it seemed that the kind of basic tertiary amine moieties (in the case of target compounds-the alicyclic ones) did not significantly influence hH3R affinities, with similar Ki values obtained for analogs from all four subgroups (piperidine, 3-or 4-methylpiperidines and azepane). Among all the target compounds, 4-F substituted benzophenones with pentyloxyl linker presented the most promising activity, with low nanomolar Ki values, i.e., 8 nM (lead 6), 12 nM (16) and 13 nM (30). The H3R affinity of 6 was comparable with the reference pitolisant. scaffold (e.g., 6 vs. 11, 16 vs. 21 and 29 vs. 34). Finally, it seemed that the kind of basic tertiary amine moieties (in the case of target compounds-the alicyclic ones) did not significantly influence hH3R affinities, with similar Ki values obtained for analogs from all four subgroups (piperidine, 3-or 4-methylpiperidines and azepane). Among all the target compounds, 4-F substituted benzophenones with pentyloxyl linker presented the most promising activity, with low nanomolar Ki values, i.e., 8 nM (lead 6), 12 nM (16) and 13 nM (30). The H3R affinity of 6 was comparable with the reference pitolisant.  [32]; * data from [31]; ** data from [33], nt, not tested.

Inhibition of eeAChE, eqBuChE, hAChE and hBuChE
All the target compounds were screened for their AChE and BuChE inhibitory activities, using Ellman's colorimetric assay [41] and electric eel AChE (eeAChE) or equine serum BuChE (eqBuChE) enzymes. The screening concentration was 10 μM. For most active compounds, with at least 50% inhibitory potency at 10 μM, IC50 values were determined. Lead compound 6 was also tested using human AChE and human BuChE (hAChE and hBuChE) and its IC50 values for the inhibition of the human enzymes were determined. Tacrine was used as the reference. The results of the experiments are collected in Table 1. The inhibitory activity of most target structures toward eeAChE was good, with a micromolar range of the obtained IC50 values. However, all ortho-substituted benzophenones (2, 7, 12, 17, 26, 31) and para-substituted 4 displayed lower inhibitory potency, with 3-41% of enzyme inhibition at screening concentration. Based on the performed assays, it was indicated that a six carbon atom linker was preferred for the meta-substituted benzophenones (e.g., 3 vs. 8, or 13 vs. 18, and 22 vs. 24). Furthermore, the eeAChE inhibitory activity was more dependent on the presence of halogen substituents in the benzophenone scaffold, the position of the alkoxyl chain and the type of amine moieties than on small changes in the linker 22 3  [32]; * data from [31]; ** data from [33], nt, not tested.

Inhibition of eeAChE, eqBuChE, hAChE and hBuChE
All the target compounds were screened for their AChE and BuChE inhibitory activities, using Ellman's colorimetric assay [41] and electric eel AChE (eeAChE) or equine serum BuChE (eqBuChE) enzymes. The screening concentration was 10 μM. For most active compounds, with at least 50% inhibitory potency at 10 μM, IC50 values were determined. Lead compound 6 was also tested using human AChE and human BuChE (hAChE and hBuChE) and its IC50 values for the inhibition of the human enzymes were determined. Tacrine was used as the reference. The results of the experiments are collected in Table 1. The inhibitory activity of most target structures toward eeAChE was good, with a micromolar range of the obtained IC50 values. However, all ortho-substituted benzophenones (2, 7, 12, 17, 26, 31) and para-substituted 4 displayed lower inhibitory potency, with 3-41% of enzyme inhibition at screening concentration. Based on the performed assays, it was indicated that a six carbon atom linker was preferred for the meta-substituted benzophenones (e.g., 3 vs. 8, or 13 vs. 18, and 22 vs. 24). Furthermore, the eeAChE inhibitory activity was more dependent on the presence of halogen substituents in the benzophenone scaffold, the position of the alkoxyl chain and the type of amine moieties than on small changes in the linker 26 3  [32]; * data from [31]; ** data from [33], nt, not tested.

Inhibition of eeAChE, eqBuChE, hAChE and hBuChE
All the target compounds were screened for their AChE and BuChE inhibitory activities, using Ellman's colorimetric assay [41] and electric eel AChE (eeAChE) or equine serum BuChE (eqBuChE) enzymes. The screening concentration was 10 µM. For most active compounds, with at least 50% inhibitory potency at 10 µM, IC 50 values were determined. Lead compound 6 was also tested using human AChE and human BuChE (hAChE and hBuChE) and its IC 50 values for the inhibition of the human enzymes were determined. Tacrine was used as the reference. The results of the experiments are collected in Table 1. The inhibitory activity of most target structures toward eeAChE was good, with a micromolar range of the obtained IC 50 values. However, all ortho-substituted benzophenones (2, 7, 12, 17, 26, 31) and para-substituted 4 displayed lower inhibitory potency, with 3-41% of enzyme inhibition at screening concentration. Based on the performed assays, it was indicated that a six carbon atom linker was preferred for the meta-substituted benzophenones (e.g., 3 vs. 8, or 13 vs. 18, and 22 vs. 24). Furthermore, the eeAChE inhibitory activity was more dependent on the presence of halogen substituents in the benzophenone scaffold, the position of the alkoxyl chain and the type of amine moieties than on small changes in the linker length. The most potent eeAChE inhibitor was found among azepane derivatives (30; IC 50 = 1.110 µM). The piperidine-substituted lead 6 presented slightly weaker activity with IC 50 = 2.303 µM and 9.59 µM for eeAChE and hAChE, respectively.
All the target compounds revealed significant activity toward eqBuChE, with IC 50 values ranging from 161 nM (35) to 4.006 µM (3). In general, para-substituted benzophenones with a fluorine atom (4-F) were preferable. Their 4-Cl and unsubstituted analogs, as well as metaand ortho-substituted derivatives, were less potent (6 vs. 2-5; 11 vs. 7-10; 16 vs. 12-15; 30 vs. 26-29 and 35 vs. 31-34). The number of methylene groups present in the ether linkers was of less importance. Interestingly, some of the best eqBuChE inhibitors possessed five carbon atom linkers (e.g., 6, 16, 26 and 30) and others, six carbon atom linkers (e.g., 31, and 35). Regarding the alicyclic amine moiety, compounds belonging to the azepane subgroup were the most promising eqBuChE inhibitors; almost all with nanomolar IC 50 values. Compound 6 bearing the piperidine moiety, chosen to be the lead structure in terms of all selected AD targets, presented IC 50 values of 172 nM and 1.16 µM for eqBuChE and hBuChE, respectively.
Compounds 2, 7, 12, 17, 26, and 31 were found to be selective eqBuChE inhibitors. As both AChE and BuChE seem to be involved in acetylcholine hydrolysis, the non-selective cholinesterase inhibitory properties of most synthesized compounds are very valuable. However, those selective toward BuChE may be attractive because the BuChE activity increases continuously during AD progression at the expense of the declining activity of AChE [9].
Considering the aims of our research, all the synthesized molecules revealed good multidirectional activity toward H 3 R and cholinesterases, with the most interesting lead 6.

Kinetic Studies of eeAChE and eqBuChE Inhibition
For the most potent compound 6, kinetic studies were performed to determine the mechanism of cholinesterase (eeAChE and eqBuChE) inhibition.
The Lineweaver-Burk plots obtained for 6 displayed a series of converging lines on the same point on the x-axis (1/[ATC] ( Figure 3A)) or not much below the x-axis (1/[BTC] ( Figure 4A)), profiling a non-competitive type mechanism of eeAChE and mix type mechanism of eqBuChE inhibition. For both eeAChE and eqBChE the plots showed increased slopes (decreased V max ) at increasing concentrations of the inhibitor. Additionally, the preserved intercepts at the x-axis were observed for eeAChE (unchanged K m ), whereas for eqBuChE different intercepts at the x-axis are presented in Figure 4A (increased K m ). Noncompetitive types of eeAChE inhibition indicate an equal affinity of the inhibitor for the free enzyme and for the enzyme−substrate complex. On the other hand, mixed types of eqBuChE inhibition with increasing K m at increasing concentrations of 6 indicate a higher affinity of the inhibitor to the free enzyme than to the enzyme−substrate complex.
All the target compounds revealed significant activity toward eqBuChE, with IC50 values ranging from 161 nM (35) to 4.006 μM (3). In general, para-substituted benzophenones with a fluorine atom (4-F) were preferable. Their 4-Cl and unsubstituted analogs, as well as meta-and ortho-substituted derivatives, were less potent (6 vs. 2-5; 11 vs. 7-10; 16 vs. 12-15; 30 vs. 26-29 and 35 vs. 31-34). The number of methylene groups present in the ether linkers was of less importance. Interestingly, some of the best eqBuChE inhibitors possessed five carbon atom linkers (e.g., 6, 16, 26 and 30) and others, six carbon atom linkers (e.g., 31, and 35). Regarding the alicyclic amine moiety, compounds belonging to the azepane subgroup were the most promising eqBuChE inhibitors; almost all with nanomolar IC50 values. Compound 6 bearing the piperidine moiety, chosen to be the lead structure in terms of all selected AD targets, presented IC50 values of 172 nM and 1.16 μM for eqBuChE and hBuChE, respectively.
Compounds 2, 7, 12, 17, 26, and 31 were found to be selective eqBuChE inhibitors. As both AChE and BuChE seem to be involved in acetylcholine hydrolysis, the non-selective cholinesterase inhibitory properties of most synthesized compounds are very valuable. However, those selective toward BuChE may be attractive because the BuChE activity increases continuously during AD progression at the expense of the declining activity of AChE [9].
Considering the aims of our research, all the synthesized molecules revealed good multidirectional activity toward H3R and cholinesterases, with the most interesting lead 6.

Kinetic Studies of eeAChE and eqBuChE Inhibition
For the most potent compound 6, kinetic studies were performed to determine the mechanism of cholinesterase (eeAChE and eqBuChE) inhibition.
The Lineweaver-Burk plots obtained for 6 displayed a series of converging lines on the same point on the x-axis (1/[ATC] ( Figure 3A)) or not much below the x-axis (1/[BTC] ( Figure 4A)), profiling a non-competitive type mechanism of eeAChE and mix type mechanism of eqBuChE inhibition. For both eeAChE and eqBChE the plots showed increased slopes (decreased Vmax) at increasing concentrations of the inhibitor. Additionally, the preserved intercepts at the x-axis were observed for eeAChE (unchanged Km), whereas for eqBuChE different intercepts at the x-axis are presented in Figure 4A (increased Km). Noncompetitive types of eeAChE inhibition indicate an equal affinity of the inhibitor for the free enzyme and for the enzyme−substrate complex. On the other hand, mixed types of eqBuChE inhibition with increasing Km at increasing concentrations of 6 indicate a higher affinity of the inhibitor to the free enzyme than to the enzyme−substrate complex.

Human MAO B Inhibitory Activity
Inhibitory activity for hMAO B was evaluated in vitro in the fluorescence-based assay as described previously [32]. All compounds were tested at the concentration of 1 μM. Rasagiline and safinamide were used as the reference compounds. The results are collected in Table 1. None of the tested compounds showed inhibition for hMAO B higher than 50%, thus they were not further evaluated for both MAO B and MAO A inhibition and the IC50 values were not determined.

Molecular Modeling Studies to Histamine H3 Receptor
For this study, we used the previously described H3R homology model, constructed on the template of the crystal structure of the M2 muscarinic acetylcholine receptor (PDB ID: 3UON) [42].
All the compounds were docked and characterized by relatively high docking score values and fit the H3R binding pocket. Independently of the alkyl chain length, west-and east-end variations, the docked ligands occupied the binding pocket in a similar mode preserving a crucial H3R antagonist/inverse agonist interactions, namely salt bridge and/or hydrogen bond formation between protonated amine nitrogen and Glu206 5.46 (superscripts denote Ballesteros-Weinstein numbering)( Figure 5) [43]. The east-end benzophenone fragments occupied the space fenced by the aromatic features of Phe193, Tyr189 (ECL2) and Tyr 394 7.35 on the sides and Tyr91 2.61 and Tyr94 2.64 on top. In the case of orthoderivatives, the terminal benzene ring was placed perpendicular to Tyr189, stabilized by π-π stacking interactions and, due to its twisted position, a single hydrogen bond with one of the Arg381 6.58 nitrogens. In the two remaining subgroups (meta-and para-) these rings were stabilized either by Tyr189 or/and Tyr394 7.35 , complimented by a double Hbond with the aforementioned arginine 6.58 . Additional stabilization through halogen bond formation with Tyr91 2.61 for 4-Cl substituted derivatives was also present. The stability of the calculated poses for selected compounds (6, 16 and 30) was further evaluated by means of short molecular dynamics (MD) simulations. From each simulation, seven poses were selected (starting pose, and after every 100 ps up to 600 ps). In most cases complexes

Human MAO B Inhibitory Activity
Inhibitory activity for hMAO B was evaluated in vitro in the fluorescence-based assay as described previously [32]. All compounds were tested at the concentration of 1 µM. Rasagiline and safinamide were used as the reference compounds. The results are collected in Table 1. None of the tested compounds showed inhibition for hMAO B higher than 50%, thus they were not further evaluated for both MAO B and MAO A inhibition and the IC 50 values were not determined.

Molecular Modeling Studies to Histamine H 3 Receptor
For this study, we used the previously described H 3 R homology model, constructed on the template of the crystal structure of the M 2 muscarinic acetylcholine receptor (PDB ID: 3UON) [42].
All the compounds were docked and characterized by relatively high docking score values and fit the H 3 R binding pocket. Independently of the alkyl chain length, west-and east-end variations, the docked ligands occupied the binding pocket in a similar mode preserving a crucial H 3 R antagonist/inverse agonist interactions, namely salt bridge and/or hydrogen bond formation between protonated amine nitrogen and Glu206 5.46 (superscripts denote Ballesteros-Weinstein numbering)( Figure 5) [43]. The east-end benzophenone fragments occupied the space fenced by the aromatic features of Phe193, Tyr189 (ECL2) and Tyr 394 7.35 on the sides and Tyr91 2.61 and Tyr94 2.64 on top. In the case of ortho-derivatives, the terminal benzene ring was placed perpendicular to Tyr189, stabilized by π-π stacking interactions and, due to its twisted position, a single hydrogen bond with one of the Arg381 6.58 nitrogens. In the two remaining subgroups (meta-and para-) these rings were stabilized either by Tyr189 or/and Tyr394 7.35 , complimented by a double H-bond with the aforementioned arginine 6.58 . Additional stabilization through halogen bond formation with Tyr91 2.61 for 4-Cl substituted derivatives was also present. The stability of the calculated poses for selected compounds (6, 16 and 30) was further evaluated by means of short molecular dynamics (MD) simulations. From each simulation, seven poses were selected (starting pose, and after every 100 ps up to 600 ps). In most cases complexes appeared stable through the whole 600 ps simulation, retaining the key interactions, and the potential energy (U) of the atomic system at the level of~1000 kcal/mol, with starting poses (marked in grey, Figure 6) similar to its orientation at the end of the simulation (marked yellow). However, a shift in the binding site with retained conformation appeared in the first 100 ps of the simulation. In the case of ortho-derivatives, a + 90 • bend of benzophenone fragment occurred, which resulted in the change of hydrogen bond formation between the carbonyl group oxygen from Arg381 6.58 to Tyr374 7.35 and Tyr115 3.33 . Moreover, for the most potent 4-fluoro derivatives, a contact with highly conserved W371 6.48 can be seen for six out of seven frames, as well as additional π-π stabilization for both benzene rings. Detailed analysis of the compound behavior during MD simulations is provided by examining the changes in their interactions with H 3 R (Figure 7). The results indicate a relatively consistent set of ligand-protein interactions that occur during the whole MD simulation. Most of the interactions occur within the TM3, TM5, TM6 and TM7 and ECL2, with consistent interactions with Tyr115 3.33 , Tyr189 45.51 , Glu206 5.46 , Tyr 374 6.51 , and Tyr394 7.35 . appeared stable through the whole 600 ps simulation, retaining the key interactions, and the potential energy (U) of the atomic system at the level of ~1000 kcal/mol, with starting poses (marked in grey, Figure 6) similar to its orientation at the end of the simulation (marked yellow). However, a shift in the binding site with retained conformation appeared in the first 100 ps of the simulation. In the case of ortho-derivatives, a + 90° bend of benzophenone fragment occurred, which resulted in the change of hydrogen bond formation between the carbonyl group oxygen from Arg381 6.58 to Tyr374 7.35 and Tyr115 3.33 . Moreover, for the most potent 4-fluoro derivatives, a contact with highly conserved W371 6.48 can be seen for six out of seven frames, as well as additional π-π stabilization for both benzene rings. Detailed analysis of the compound behavior during MD simulations is provided by examining the changes in their interactions with H3R ( Figure 7). The results indicate a relatively consistent set of ligand-protein interactions that occur during the whole MD simulation. Most of the interactions occur within the TM3, TM5, TM6 and TM7 and ECL2, with consistent interactions with Tyr115 3.33 , Tyr189 45.51 , Glu206 5.46 , Tyr 374 6.51 , and Tyr394 7.35 .   appeared stable through the whole 600 ps simulation, retaining the key interactions, and the potential energy (U) of the atomic system at the level of ~1000 kcal/mol, with starting poses (marked in grey, Figure 6) similar to its orientation at the end of the simulation (marked yellow). However, a shift in the binding site with retained conformation appeared in the first 100 ps of the simulation. In the case of ortho-derivatives, a + 90° bend of benzophenone fragment occurred, which resulted in the change of hydrogen bond formation between the carbonyl group oxygen from Arg381 6.58 to Tyr374 7.35 and Tyr115 3.33 . Moreover, for the most potent 4-fluoro derivatives, a contact with highly conserved W371 6.48 can be seen for six out of seven frames, as well as additional π-π stabilization for both benzene rings. Detailed analysis of the compound behavior during MD simulations is provided by examining the changes in their interactions with H3R ( Figure 7). The results indicate a relatively consistent set of ligand-protein interactions that occur during the whole MD simulation. Most of the interactions occur within the TM3, TM5, TM6 and TM7 and ECL2, with consistent interactions with Tyr115 3.33 , Tyr189 45.51 , Glu206 5.46 , Tyr 374 6.51 , and Tyr394 7.35 .

Analysis of Binding Mode within AChE
Docking studies of benzophenone derivatives allowed us to describe their binding modes in the AChE active site. The most active compounds were bound similarly: the heterocyclic rings interacted with the catalytic triad and anionic site residues. Next, linkers with the ether group were located near amino acids from the anionic site and acyl pocket, and the benzophenone aromatic fragment was close to the peripheral anionic site (PAS) residues [44].

Analysis of Binding Mode within AChE
Docking studies of benzophenone derivatives allowed us to describe their binding modes in the AChE active site. The most active compounds were bound similarly: the heterocyclic rings interacted with the catalytic triad and anionic site residues. Next, linkers with the ether group were located near amino acids from the anionic site and acyl pocket, and the benzophenone aromatic fragment was close to the peripheral anionic site (PAS) residues [44].
The heterocyclic moieties of the above-mentioned compounds were close to His440 from the catalytic triad and Trp84, and Phe330 from the anionic site. The ionized amine groups from piperidine or azepane moieties interacted with aromatic residues through cation-π interactions. Protonated nitrogen atoms created a hydrogen bond with a water molecule (1159). Alkyl, flexible chains were responsible for hydrophobic interactions with Phe330, Phe331 from the anionic site and Phe290 from the acyl pocket of the enzyme. Additionally, ligands interacted with residues from the peripheral anionic site-Tyr334 and Tyr121. Oxygen atoms from alkyl linkers were able to create hydrogen bonds with a hydroxyl group of Tyr121 which next interacted with water 1159. This hydrogen bond network was described in our previous work [26]. Large, benzophenone fragments of the analyzed compounds were located between residues from PAS, such as Tyr121, Tyr70, Tyr334, and Trp279. Additionally, the oxygen atom from the carbonyl group interacted with Tyr70 through a hydrogen bond. To check the stability of the pose obtained from docking studies, we decided to perform a molecular dynamics simulation. For calculations, we selected compound 6. In the course of the MD simulation, we observed changes in the conformation of the ligand. During the last three ns of the simulation, the ligand revealed an RMSD value of 4.5 Å ( Figure S1). Docking studies showed the presence of hydrogen bonds with hydroxy The heterocyclic moieties of the above-mentioned compounds were close to His440 from the catalytic triad and Trp84, and Phe330 from the anionic site. The ionized amine groups from piperidine or azepane moieties interacted with aromatic residues through cation-π interactions. Protonated nitrogen atoms created a hydrogen bond with a water molecule (1159). Alkyl, flexible chains were responsible for hydrophobic interactions with Phe330, Phe331 from the anionic site and Phe290 from the acyl pocket of the enzyme. Additionally, ligands interacted with residues from the peripheral anionic site-Tyr334 and Tyr121. Oxygen atoms from alkyl linkers were able to create hydrogen bonds with a hydroxyl group of Tyr121 which next interacted with water 1159. This hydrogen bond network was described in our previous work [26]. Large, benzophenone fragments of the analyzed compounds were located between residues from PAS, such as Tyr121, Tyr70, Tyr334, and Trp279. Additionally, the oxygen atom from the carbonyl group interacted with Tyr70 through a hydrogen bond.
To check the stability of the pose obtained from docking studies, we decided to perform a molecular dynamics simulation. For calculations, we selected compound 6. In the course of the MD simulation, we observed changes in the conformation of the ligand. During the last three ns of the simulation, the ligand revealed an RMSD value of 4.5 Å ( Figure  S1). Docking studies showed the presence of hydrogen bonds with hydroxy groups from Tyr121 and Tyr70. Interaction with Tyr70 was not found to be stable in the MD course; however, interaction with Tyr121 appeared again at the end of the simulation ( Figure S2). Possibly, such behavior in the course of MD might be explained by the level of ligand activity (IC 50 = 2.306 µM).

Analysis of Binding Mode within BuChE
The results of docking to the butyrylcholinesterase active site were analyzed, and it was noticed that the most active ligands obtained similar binding modes to one another. The aliphatic rings with tertiary amine were found near the anionic site residues. Alkyl linkers interacted with residues from the peripheral anionic site. Aromatic rings were located between residues from the acyl pocket and the anionic site [44].
Compounds 35 (IC 50 = 0.161 µM) and 6 (IC 50 = 0.172 µM) revealed binding modes as follows: azepane and piperidine fragments interacted with Tyr128 and Trp82 through cation-π interactions. Additionally, ionic interactions between the tertiary amine groups of the described compounds and the carboxyl group from Glu197 were observed. Alkyl chains were located near Tyr332 from the peripheral anionic site. Large, aromatic benzophenone fragments interacted with hydrophobic residues, such as Leu286 and Val288 from the acyl pocket and Phe329 counted to the anionic site. Figure 9 presents compounds 35 (upper panel) and 6 (lower panel).
The binding mode of compound 6 within the butyrylcholinesterase active site was assessed by molecular dynamics simulation. Ligand presented a stable binding mode in the simulation and the RMSD value was at the level of 2 Å ( Figure S3). These results are promising in comparison to the MD results for the same compound in a complex with AChE. Such behavior could be associated with higher biological activity at the nanomolar level (IC 50 = 0.172 µM). Contrary to AChE, in the docking to BuChE, it was observed interaction between tertiary basic amine and the carboxy group of Glu197 which seems to be crucial for the ligand binding ( Figure 10).
follows: azepane and piperidine fragments interacted with Tyr128 and Trp82 through cation-π interactions. Additionally, ionic interactions between the tertiary amine groups of the described compounds and the carboxyl group from Glu197 were observed. Alkyl chains were located near Tyr332 from the peripheral anionic site. Large, aromatic benzophenone fragments interacted with hydrophobic residues, such as Leu286 and Val288 from the acyl pocket and Phe329 counted to the anionic site. Figure 9 presents compounds 35 (upper panel) and 6 (lower panel).  Molecules 2023, 27, x FOR PEER REVIEW 13 The binding mode of compound 6 within the butyrylcholinesterase active site assessed by molecular dynamics simulation. Ligand presented a stable binding mod the simulation and the RMSD value was at the level of 2 Å ( Figure S3). These results promising in comparison to the MD results for the same compound in a complex w AChE. Such behavior could be associated with higher biological activity at the nanom level (IC50 = 0.172 μM). Contrary to AChE, in the docking to BuChE, it was observed teraction between tertiary basic amine and the carboxy group of Glu197 which seem be crucial for the ligand binding ( Figure 10).

Analysis of Binding Mode with Monoamine Oxidase B
The most active compounds showed convergent binding modes within the monoamine oxidase B active site. Heterocyclic rings with protonated nitrogen atoms were located in the aromatic pocket of the enzyme. Alkyl chains were able to create hydrophobic interactions with the protein. Large, aromatic benzophenone fragments were found near the cofactor (FAD).
Compound 4 was the most active in a screening concentration (% inh. at 1 µM = 44%). Figure 11 presents its binding mode (upper panel) and the binding mode of ligand 6 (lower panel; % inh at 1 µM = 25%). Piperidine moieties interacted with the following amino acids: Trp119, Phe103, Phe168 and Ile199. Tertiary amines from the piperidine rings created cationπ interactions with the Trp119 aromatic residue. Carbon linkers with ether groups were positioned near Leu167, Ile316, and Tyr326. The benzophenone fragments were located near FAD and interacted with Tyr398, Phe343, Tyr60 and Tyr435. It is worth noticing that the oxygen atoms from the carbonyl group could create a hydrogen bond with Tyr435. Next, we performed an MD simulation for a complex of ligand 6 with the monoamine oxidase B. Analysis showed only small changes in the binding mode. The RMSD value of the ligand in the last two ns of the MD simulation was about 3 Å ( Figure S4). A hydrogen bond with the hydroxy group of Tyr435 was present and it was noticed that additional interactions with the nitrogen atom of the amide group from Gln206 or weak interaction with the hydroxy group of Tyr398 appeared due to the residues and ligand movement ( Figure S5).

Selected ADMET Properties
The chemical stability of the drug candidate affects its pharmacokinetic profile. Therefore, ADMET studies should be performed at the early stage of the drug development process. In vitro pharmacological profiling includes, among others, metabolic stability and hepatotoxicity. A metabolically unstable molecule may undergo rapid metabolism leading to the formation of metabolites that may not only be toxic but also devoid of biological activity. One of the possible harmful effects of compounds is liver damage. It is Next, we performed an MD simulation for a complex of ligand 6 with the monoamine oxidase B. Analysis showed only small changes in the binding mode. The RMSD value of the ligand in the last two ns of the MD simulation was about 3 Å ( Figure S4). A hydrogen bond with the hydroxy group of Tyr435 was present and it was noticed that additional interactions with the nitrogen atom of the amide group from Gln206 or weak interaction with the hydroxy group of Tyr398 appeared due to the residues and ligand movement ( Figure S5).

Selected ADMET Properties
The chemical stability of the drug candidate affects its pharmacokinetic profile. Therefore, ADMET studies should be performed at the early stage of the drug development process. In vitro pharmacological profiling includes, among others, metabolic stability and hepatotoxicity. A metabolically unstable molecule may undergo rapid metabolism leading to the formation of metabolites that may not only be toxic but also devoid of biological activity. One of the possible harmful effects of compounds is liver damage. It is advisable to eliminate this risk at an early stage of preclinical development and to test the ability of the most promising compounds to cause hepatotoxicity. Another important factor for compounds expected to induce CNS effects is their ability to penetrate the BBB. On the one hand, the BBB plays an important role in protecting the brain, especially against harmful agents, but on the other hand, it can hinder the penetration of therapeutically important compounds. The evaluation of this ability of compounds at an initial stage may be a clue to the selection of suitable ligands for further pharmacological studies, especially in vivo studies. In neurodegenerative diseases, such as AD, the relevant neurons degenerate. If we can find drugs that can protect the nerves from damage, we have a chance to stop the progression of the disease. Therefore, it justifies testing the ability of compounds to protect neurons at an early stage of research.
In the presented study, in silico and in vitro methods were used to evaluate the selected properties of lead compound 6. The metabolic stability was evaluated using in silico predictions (MetaSite 6.0.1) and in vitro experiments with mouse liver microsomes (MLMs). Additionally, the neuroprotection studies as well as the hepatotoxic effect of the most active compound 6 were evaluated with the use of the SH-SY5Y neuroblastoma cell line and hepatoma HepG2 cell line, respectively. Finally, the passive permeability through biological membranes for the lead 6 was also estimated.

In Vitro Metabolic Stability of Compound 6
The metabolic stability of 6 was investigated first in silico with the use of MetaSite 6.0.1 software [46]. The results showed that the piperidine moiety and the aliphatic linker were the most susceptible sites for metabolism ( Figure 12). However, the in silico results were not completely confirmed in vitro. After incubation with MLMs for 120 min, the formation of three metabolites was observed ( Figure 13) and none of them was obtained as modifications at the piperidine moiety. The molecular mass of the metabolite M1 (m/z = 372.39) and MS/MS ion fragmentation analysis suggested the reduction of the carbonyl group as the main metabolic pathway (Supporting information Figure S6; Table 2). In addition to the carbonyl group reduction, hydroxylation at two different sites of 6 was also identified (Supporting information Figure S6; Table 2). Moreover, the UPLC chromatograms showed that compound 6 was biotransformed at around 34%, whereas the metabolically unstable reference verapamil was at 76%. The obtained results illustrated the rather moderate metabolic stability of 6. advisable to eliminate this risk at an early stage of preclinical development and to test the ability of the most promising compounds to cause hepatotoxicity. Another important factor for compounds expected to induce CNS effects is their ability to penetrate the BBB. On the one hand, the BBB plays an important role in protecting the brain, especially against harmful agents, but on the other hand, it can hinder the penetration of therapeutically important compounds. The evaluation of this ability of compounds at an initial stage may be a clue to the selection of suitable ligands for further pharmacological studies, especially in vivo studies. In neurodegenerative diseases, such as AD, the relevant neurons degenerate. If we can find drugs that can protect the nerves from damage, we have a chance to stop the progression of the disease. Therefore, it justifies testing the ability of compounds to protect neurons at an early stage of research.
In the presented study, in silico and in vitro methods were used to evaluate the selected properties of lead compound 6. The metabolic stability was evaluated using in silico predictions (MetaSite 6.0.1) and in vitro experiments with mouse liver microsomes (MLMs). Additionally, the neuroprotection studies as well as the hepatotoxic effect of the most active compound 6 were evaluated with the use of the SH-SY5Y neuroblastoma cell line and hepatoma HepG2 cell line, respectively. Finally, the passive permeability through biological membranes for the lead 6 was also estimated.

In Vitro Metabolic Stability of Compound 6
The metabolic stability of 6 was investigated first in silico with the use of MetaSite 6.0.1 software [46]. The results showed that the piperidine moiety and the aliphatic linker were the most susceptible sites for metabolism ( Figure 12). However, the in silico results were not completely confirmed in vitro. After incubation with MLMs for 120 min, the formation of three metabolites was observed ( Figure 13) and none of them was obtained as modifications at the piperidine moiety. The molecular mass of the metabolite M1 (m/z = 372.39) and MS/MS ion fragmentation analysis suggested the reduction of the carbonyl group as the main metabolic pathway (Supporting information Figure S6; Table 2). In addition to the carbonyl group reduction, hydroxylation at two different sites of 6 was also identified (Supporting information Figure S6; Table 2). Moreover, the UPLC chromatograms showed that compound 6 was biotransformed at around 34%, whereas the metabolically unstable reference verapamil was at 76%. The obtained results illustrated the rather moderate metabolic stability of 6.     [47] obtained according to the same procedure as used for 6. Figure 13. UPLC after 120 min incubation of compound 6 with MLMs.

Neuroprotection Studies of Compound 6
The neuroprotection studies were performed with the use of the SH-SY5Y neuroblastoma cell line and H2O2 at a toxic concentration of 300 μM co-incubated with 10 μM (the highest non-toxic concentration, estimated before the assay) of 6 for 24 h. The cell necrosis and viability were next detected by LDH and MTS assays, respectively. In the LDH test slight inhibition of the necrosis process by 6 at a concentration of 10 μM was observed; however, it was statistically insignificant. At the same time salsolinol (SAL), the reference compound with previously confirmed neuroprotective activity [48], lowered the number of necrotic cells to the control level ( Figure 14A). Statistical significance was calculated against the number of necrotic cells in the H2O2-treated population. Additionally, the MTS test was performed in which the viability of cells was measured. In this case, there was no increase in the viability of the SH-SY5Y cells population treated with H2O2 in the presence of 6 at 10 μM, whereas SAL also showed neuroprotective potential and statistically significantly increased the number of living cells compared to H2O2 alone ( Figure 14B).

Neuroprotection Studies of Compound 6
The neuroprotection studies were performed with the use of the SH-SY5Y neuroblastoma cell line and H 2 O 2 at a toxic concentration of 300 µM co-incubated with 10 µM (the highest non-toxic concentration, estimated before the assay) of 6 for 24 h. The cell necrosis and viability were next detected by LDH and MTS assays, respectively. In the LDH test slight inhibition of the necrosis process by 6 at a concentration of 10 µM was observed; however, it was statistically insignificant. At the same time salsolinol (SAL), the reference compound with previously confirmed neuroprotective activity [48], lowered the number of necrotic cells to the control level ( Figure 14A). Statistical significance was calculated against the number of necrotic cells in the H 2 O 2 -treated population. Additionally, the MTS test was performed in which the viability of cells was measured. In this case, there was no increase in the viability of the SH-SY5Y cells population treated with H 2 O 2 in the presence of 6 at 10 µM, whereas SAL also showed neuroprotective potential and statistically significantly increased the number of living cells compared to H 2 O 2 alone ( Figure 14B).

The Permeability of Compound 6
The PAMPA method was used to estimate the compoundʹs passive permeability as described previously [32]. Caffeine was used as a well-permeable drug. Results presented in Table 3 show good permeability of compound 6 with a permeability coefficient value (Pe) of 6.3 × 10 −6 cm/s. Caffeine (Pe = 15.1 × 10 −6 cm/s) showed higher Pe than 6 [49].

Hepatotoxicity Studies of 6
There are many factors that can lead to liver damage, e.g., the simultaneous use of other drugs, genetic risk (polymorphism in CYP450), or personal factors (age, gender) [50]. The hepatotoxicity of a drug candidate is the main cause of the interruption of its further development, and in vitro toxicity testing is an essential part of drug safety testing at an early stage of its development. Thus, the hepatoma HepG2 cell line was used here to evaluate the hepatotoxicity of compound 6. Doxorubicin (DX) at the concentration of 1 μM was used as the reference cytostatic drug. Compound 6 decreased cell viability starting from 10 μM concentration and above, while DX showed the toxic effect already at 1 μM concentration ( Figure 15). Regardless, the obtained results indicate that the risk of hepatotoxicity cannot be excluded.

The Permeability of Compound 6
The PAMPA method was used to estimate the compound's passive permeability as described previously [32]. Caffeine was used as a well-permeable drug. Results presented in Table 3 show good permeability of compound 6 with a permeability coefficient value (P e ) of 6.3 × 10 −6 cm/s. Caffeine (P e = 15.1 × 10 −6 cm/s) showed higher P e than 6 [49].

Hepatotoxicity Studies of 6
There are many factors that can lead to liver damage, e.g., the simultaneous use of other drugs, genetic risk (polymorphism in CYP450), or personal factors (age, gender) [50]. The hepatotoxicity of a drug candidate is the main cause of the interruption of its further development, and in vitro toxicity testing is an essential part of drug safety testing at an early stage of its development. Thus, the hepatoma HepG2 cell line was used here to evaluate the hepatotoxicity of compound 6. Doxorubicin (DX) at the concentration of 1 µM was used as the reference cytostatic drug. Compound 6 decreased cell viability starting from 10 µM concentration and above, while DX showed the toxic effect already at 1 µM concentration ( Figure 15). Regardless, the obtained results indicate that the risk of hepatotoxicity cannot be excluded.

In Silico Prediction of Mutagenic Effect of Selected Compounds
The Ames test is used worldwide as a first-order filter to assess the mutagenic potential of new chemicals and drugs. The calculations carried out for compound 6 and its analogs (i.e., compounds 16: 3-methylpiperidine and 30: azepane) were performed using the free online programme pkSCM (https://biosig.lab.uq.edu.au/pkcsm/prediction, accessed on 5 December 2022). In addition, the metabolites predicted by the MetaSite program for compound 6 and its analogs (i.e., compounds with a reduced carbonyl group; compounds with a hydroxyl group in the benzene and compounds with a reduced carbonyl and hydroxyl group in the benzene ring) were also added to this study. Calculations did not predict the Ames test activity for all compounds evaluated.

In Silico Prediction of Mutagenic Effect of Selected Compounds
The Ames test is used worldwide as a first-order filter to assess the mutageni potential of new chemicals and drugs. The calculations carried out for compound 6 and its analogs (i.e., compounds 16: 3-methylpiperidine and 30: azepane) were performed using the free online programme pkSCM (https://biosig.lab.uq.edu.au/pkcsm/prediction accessed on 5 December 2022). In addition, the metabolites predicted by the MetaSite program for compound 6 and its analogs (i.e., compounds with a reduced carbonyl group compounds with a hydroxyl group in the benzene and compounds with a reduced carbonyl and hydroxyl group in the benzene ring) were also added to this study Calculations did not predict the Ames test activity for all compounds evaluated.

In Vivo Pharmacological Studies
Encouraged by both the BChE inhibitory activity and high H3R affinity, low hepato toxicity, and proven passive permeability, we decided to test the effects of compound 6 on behavioral activities. Therefore, we investigated the in vivo activity of compound 6 in the Passive Avoidance Test (PAT) to evaluate its influence on memory and learning in mice with scopolamine-induced amnesia, which is widely referred to as a model simulat ing dementia in AD. Considering that the procognitive and analgesic effects are among many possible activities described for H3Rs ligands, we set out to investigate compound 6 in the formalin test to assess its analgesic potential.
In the PAT, mice with correct memory and learning processes behave contrary to their innate tendencies for preference of dark areas. The test consists of two trials, the acquisition and retention trials. In the first one, the mouse placed in the white compart ment receives a mild foot shock when it innately crosses to the black compartment. During this phase, the animal learns that crossing into the dark compartment has negative conse quences. In the next phase, the animal is again placed in the white compartment and the latency to escape from the white compartment is evaluated. The longer the latency, the better the memory performance. Thus, substances that impair memory, such as scopola mine, shorten the latency time in the retention trial, while substances improving memory reverse the scopolamine-induced impairment [51]. Our study showed (Figure 16) tha

In Vivo Pharmacological Studies
Encouraged by both the BChE inhibitory activity and high H 3 R affinity, low hepatotoxicity, and proven passive permeability, we decided to test the effects of compound 6 on behavioral activities. Therefore, we investigated the in vivo activity of compound 6 in the Passive Avoidance Test (PAT) to evaluate its influence on memory and learning in mice with scopolamine-induced amnesia, which is widely referred to as a model simulating dementia in AD. Considering that the procognitive and analgesic effects are among many possible activities described for H 3 Rs ligands, we set out to investigate compound 6 in the formalin test to assess its analgesic potential.
In the PAT, mice with correct memory and learning processes behave contrary to their innate tendencies for preference of dark areas. The test consists of two trials, the acquisition and retention trials. In the first one, the mouse placed in the white compartment receives a mild foot shock when it innately crosses to the black compartment. During this phase, the animal learns that crossing into the dark compartment has negative consequences. In the next phase, the animal is again placed in the white compartment and the latency to escape from the white compartment is evaluated. The longer the latency, the better the memory performance. Thus, substances that impair memory, such as scopolamine, shorten the latency time in the retention trial, while substances improving memory reverse the scopolamine-induced impairment [51]. Our study showed ( Figure 16) that compound 6 administrated at the dose of 15 mg/kg did not significantly reduce the latency time in the retention trial (169.80 ± 2.18 for vehicle-treated mice vs. 151.89 ± 5.46 for compound 6 treated mice). That proved that the tested compound itself did not induce memory impairment. It is worth emphasizing that in the retention trial, a dose-dependent slight increase in prolonged step-through latency was observed for compound 6 (15 and 30 mg/kg) in mice treated with scopolamine. Nevertheless, this compound did not statistically reverse the harmful effects of scopolamine (22.71 ± 7.56) at either the 15 mg/kg dose (47.07 ± 17.56) or the 30 mg/kg dose (51.56 ± 21.17). In summary, based on the above screening data in the PAT, it can be concluded that compound 6 has no significant harmful or beneficial effect on memory and learning. flammation. As it is presented in Figure 17, the i.p. administration of compound 6 prio the subcutaneous injection of formalin significantly attenuated the nociceptive respo in mice in both phases of the test. Its ED50 value in phase I was found to be 20.9 mg whereas the ED50 value in phase II was found to be 17.5 mg/kg. These results sug significant analgesic properties of compound 6 in both acute (phase I) and inflamma (phase II) pains. Figure 16. Activity of compound 6 in the Passive Avoidance Test. Compound 6 (15, or 30 mg i.p.) at the same volume of vehicle (1% Tween 80) was administered 30 min before the acquisi trial. Memory impairment was induced by scopolamine (1 mg/kg, s.c.) 20 min before the acquisi trial. 24 h after the acquisition trial, a retention trial was conducted for 180 s. Data represent m ± SEM (n = 8-10 per group); **** p < 0.0001 versus the scopolamine-treated group (0 on the x-a Statistical analysis: two-way ANOVA followed by Tukey's Multiple Comparison Test. Figure 16. Activity of compound 6 in the Passive Avoidance Test. Compound 6 (15, or 30 mg/kg, i.p.) at the same volume of vehicle (1% Tween 80) was administered 30 min before the acquisition trial. Memory impairment was induced by scopolamine (1 mg/kg, s.c.) 20 min before the acquisition trial. 24 h after the acquisition trial, a retention trial was conducted for 180 s. Data represent means ± SEM (n = 8-10 per group); **** p < 0.0001 versus the scopolamine-treated group (0 on the x-axis). Statistical analysis: two-way ANOVA followed by Tukey's Multiple Comparison Test.
Continuing the preclinical characterization of compound 6, we assessed its analgesic activity in the formalin test. Subcutaneous injection of formalin stimulates and then impairs sensory endings, which results in two distinct phases that can be observed during the test. Phase I mainly results from the immediate activation of nociceptors, while phase II results from the sensitization of spinal reflex circuits and peripherally developing inflammation. As it is presented in Figure 17, the i.p. administration of compound 6 prior to the subcutaneous injection of formalin significantly attenuated the nociceptive response in mice in both phases of the test. Its ED 50 value in phase I was found to be 20.9 mg/kg, whereas the ED 50 value in phase II was found to be 17.5 mg/kg. These results suggest significant analgesic properties of compound 6 in both acute (phase I) and inflammatory (phase II) pains. memory impairment. It is worth emphasizing that in the retention trial, a dose-dependent slight increase in prolonged step-through latency was observed for compound 6 (15 and 30 mg/kg) in mice treated with scopolamine. Nevertheless, this compound did not statistically reverse the harmful effects of scopolamine (22.71 ± 7.56) at either the 15 mg/kg dose (47.07 ± 17.56) or the 30 mg/kg dose (51.56 ± 21.17). In summary, based on the above screening data in the PAT, it can be concluded that compound 6 has no significant harmful or beneficial effect on memory and learning.
Continuing the preclinical characterization of compound 6, we assessed its analgesic activity in the formalin test. Subcutaneous injection of formalin stimulates and then impairs sensory endings, which results in two distinct phases that can be observed during the test. Phase I mainly results from the immediate activation of nociceptors, while phase II results from the sensitization of spinal reflex circuits and peripherally developing inflammation. As it is presented in Figure 17, the i.p. administration of compound 6 prior to the subcutaneous injection of formalin significantly attenuated the nociceptive response in mice in both phases of the test. Its ED50 value in phase I was found to be 20.9 mg/kg, whereas the ED50 value in phase II was found to be 17.5 mg/kg. These results suggest significant analgesic properties of compound 6 in both acute (phase I) and inflammatory (phase II) pains. Figure 16. Activity of compound 6 in the Passive Avoidance Test. Compound 6 (15, or 30 mg/kg, i.p.) at the same volume of vehicle (1% Tween 80) was administered 30 min before the acquisition trial. Memory impairment was induced by scopolamine (1 mg/kg, s.c.) 20 min before the acquisition trial. 24 h after the acquisition trial, a retention trial was conducted for 180 s. Data represent means ± SEM (n = 8-10 per group); **** p < 0.0001 versus the scopolamine-treated group (0 on the x-axis). Statistical analysis: two-way ANOVA followed by Tukey's Multiple Comparison Test.

Concluding Remarks
A series of 34 benzophenone derivatives was designed, synthesized, and evaluated in vitro as potential future AD multitarget therapeutics affecting the CNS neurotransmission. Compounds were divided into four subgroups, bearing piperidine, 3methylpiperidine, 4-methylpiperidine, or the azepane moiety, linked by five to six carbon atom alkoxyl chains to the benzophenone scaffolds. Most of the compounds revealed significant biological activity toward the selected targets, i.e., human histamine H 3 receptors and cholinesterase enzymes (AChE and BuChE). Compounds were also screened for human MAO B inhibitory potency, but none of them showed promising results (no more than 44% of enzyme inhibition at 1 µM). 4-F substituted benzophenones with pentyloxyl linkers were found to be the most potent hH 3 R antagonists/inverse agonists. Among them piperidine derivative 6 presented the highest hH 3 R affinity, being also a nonselective cholinesterase inhibitor (hH 3 R K i = 8 nM; eeAChE IC 50 = 2.31 µM; hAChE IC 50 = 9.59 µM; eqBuChE IC 50 = 0.17 µM; hBuChE IC 50 = 1.16 µM; 25% of MAO B inhibition at 1 µM). In vitro studies of benzophenone derivatives were supplemented by molecular docking that allowed us to describe the binding modes of the compounds in the H 3 R, AChE, BuChE and MAO B active sites. The obtained binding modes of the lead compound 6 within H 3 R, AChE, BuChE and MAO B active sites were tested during molecular dynamics simulations. In further in vitro studies, compound 6 showed moderate metabolic stability in mouse liver microsomes (66% of the parent compound remained after 2 h of incubation), good permeability with a permeability coefficient value (P e ) of 6.3 × 10 −6 cm/s, and a lack of both hepatotoxicity effect in the 0.1-1 µM range in HepG2 cells, and unfortunately, neuroprotective activity in the LDH and MTS assays. Given the in vitro results, we selected compound 6 for in vivo evaluation of its memory-enhancing potency and its analgesic potential in mice. In the passive avoidance task, compound 6 used at a dose of 15 mg/kg and 30 mg/kg did not statistically prolong step-through latency in the model of scopolamine-induced cognitive dysfunction. On the other hand, compound 6 showed significant analgesic properties in both acute (ED 50 = 20.9 mg/kg) and inflammatory (ED 50 = 17.5 mg/kg) pain in the Formalin Test. The obtained results indicate a promising direction for further structural modifications of the herein-described compound 6, considered a lead structure for further development.

Synthesis
Reagents and solvents were obtained from commercial suppliers and used without further purification. Reactions were monitored by thin-layer chromatography using commercially available plates (Merck silica gel 60 F254 plates). The spots were visualized by UV lamp and Dragendorff's reagent (solvent system: methylene chloride or methylene chloride: methanol 9:1). The structures and purity of compounds were confirmed by nuclear magnetic resonance ( 1 H NMR and 13 C NMR), mass spectra (LC/MS) and the elemental analysis (C, H, N). DMSO-d 6 was used as a solvent for NMR spectra. Data were measured by Mercury 300 MHz PFG spectrometer (Varian, Palo Alto, CA, USA) or FT-NMR 500 MHz spectrometer (Joel Ltd., Akishima, Tokyo, Japan). The chemical shifts (δ) are reported in relation to tetramethylsilane (TMS) and the coupling constants (J) are expressed in Hz. The multiplicity of each peak is reported as: s, singlet; d, doublet; dd, doublet of doublets; t, triplet; quin, quintet; m, multiplet; br, broad. LC/MS spectra were performed on Waters TQ Detector Mass Spectrometer (Water Corporation, Milford, CT, USA). Retention times (t R ) are given in minutes. The purity of compounds confirmed by UPLC/MS analysis was ≥95% (except 21: 93.94% and 35: 94.57%). The elemental analysis (C, H, N) for compounds was performed on Vario EL III Elemental Analyser (Hanau, Germany) and results agreed within 0.5% of the theoretical value (except 3: 0.62% for H; 21: 0.58% for C; 35: 0.52% for C).

General Procedure for the Synthesis of Final Compounds
Compounds were synthesized and purified as described previously [29]. Briefly, a proper bromoalkoxybenzophenone (1 eq) was added to the solution of a proper amine (2 eq) in a mixture of ethanol and water (21:4, v/v) in the presence of K 2 CO 3 (15 eq) and KI (the catalytic amount) and refluxed for 10-20 h. They were purified by extraction between CH 2 Cl 2 and 1% HCl. The raw product was transformed into an oxalic acid salt in absolute ethanol and precipitated (if necessary) by adding diethyl-ether.

Inhibition of Human MAO B
The precise method was described in [32]. Briefly, tested compounds were screened for hMAO B inhibitory activity at 1 µM concentration by the fluorometric method. Paratyramine (200 µM) was used as a substrate for the enzyme while safinamide (1 µM) and rasagiline (1 µM) were used as reference inhibitors (reversible and irreversible, respectively). Each experiment was performed twice in duplicate. For docking purposes, MOE v 2019.1 [53] was used. Ligands were built in their ionized forms (protonated N4 piperazine nitrogen, structure charge +1). Bioactive conformations were generated using the OMEGA package [54] (maximum number of conformers = 100; RMS = 0.5). For all compounds, the five lowest energy conformers were selected for docking studies. Possible binding pocket adaptation in presence of certain ligands was examined using the induced fit refinement protocol with 6. To validate the methods used, the ligand was then redocked with high confidence. The site was then centered on the ligand. Docking to the rigid form of the receptor was performed using standard docking protocol. Ligands were rated according to their position in the binding pocket, the docking score value and interactions with binding pocket amino acids. Ligand-receptor visualizations were generated using Schrödinger Maestro free academic. Dynamics simulations (in time of 600 ps, T = 300 K) were performed using the Nosé-Poincaré-Andersen equations of motion, forcefield AMBER10:EHT; R-Field 1:80, Cutoff (8,10) and performed in MOE v 2019.1 [53].

Molecular Modeling Studies to AChE, BuChE and MAO B
The tested compounds were drawn in Marvin and saved in smiles format. Subsequently, they were prepared with a LigPrep module in pH 7.4 ± 0.2 with an ionizer function, and available isomers were generated. Docking studies were carried out according to the earlier validated procedures [44]. Ligands were docked into the: acetylcholinesterase (PDB code: 1EVE), butyrylcholinesterase (PDB code: 1P0I) and human monoamine oxidase B (PDB code: 4A79). The crystal structures of proteins were prepared with Hermes: the protonation of histidines was set at Nε atoms, and hydrogens were added. Ligand and water molecules were removed from the complexes, except for water molecules with numbers: 1159, 1249 and 1254 in AChE. Binding sites for docking were marked within 10 Å, 20 Å and 12 Å from the reference compounds: donepezil-AChE, butyrate-BuChE and pioglitazone for MAO B. For calculations, standard parameters of the genetic algorithm were used. The population size was defined as 100 and the operation number was 100,000. For each compound 10 poses were obtained. Results were analyzed in PyMOL software. For molecular dynamics simulation, all-atom systems and input files were prepared with the CHARMM-GUI online server. The complexes of protein with ligands were prepared as follows: solvation was performed with TIP3P water molecules, and Na + and Cl − ions were added (0.15 M NaCl). For system equilibration, the one-step protocol implemented in CHARMM-GUI was used. Molecular dynamics simulations were performed with NAMD version 2.13 at a temperature of 303.15 K. MD timesteps were selected as 2fs, and the total time was set as 10 ns. The calculation was performed with a CHARMM36m force field. Results analysis was performed with the VMD program.

Neuroprotection
Neuroprotection studies were performed according to the procedures described previously [46]. In brief, SH-SY5Y CRL-2266 ™ (American Type Culture Collection, Manassas, VG, USA) were seeded in a microplate at a concentration of 2.5 × 10 4 cells/well in 100 µL culture medium and cultured for 24 h at 37 • C and 5% CO 2 . The cells were preincubated first for 1h with compound 6 (10 µM) or with the reference neuroprotectant salsolinol (SAL). Next, H 2 O 2 was added at a final concentration of 300 µM and the cells were placed into the incubator. After 24 h of the compound co-incubation with H 2 O 2 , the CellTiter 96 ® AQueous Non-Radioactive Cell Proliferation Assay (MTS) and CytoTox-ONE™ Homogeneous Membrane Integrity Assay (second independent experiment) were performed. Both tests were purchased from Promega (Madison, WI, USA). The absorbance and fluorescence were measured using a microplate reader EnSpire (PerkinElmer, Waltham, MA USA) at 490 nm. All results were shown as mean ± SD. The statistical significances were calculated by GraphPad Prism 6.0 software.

Permeability
The passive transport of compound 6 through the cell membranes was estimated by Pre-coated PAMPA Plate System Gentest ™ provided by Corning (Tewksbury, MA, USA) and was carried out according to the procedure previously described [32]. The assay was carried out according to the manufacturer's instruction and the permeability coefficient Pe was counted on the basis of the provided equations. The concentrations of a tested compound in the apical and basolateral wells were determined using the LC/MS analysis with an internal standard. The results were compared to the high permeable reference caffeine (CFN).

Hepatotoxicity
Hepatoma HepG2 HB-8065 ™ (American Type Culture Collection, Manassas, VG, USA) cells were seeded in a microplate at a concentration of 1.0 × 10 4 cells/well in 100 µL and incubated for 24 h. Compound 6 was added next to the cells in the six concentrations: 0.1, 1, 10, 25, 50 and 100 µM and incubated for 72 h. The cell viability was defined by an MTS assay provided by Promega (Madison, WI, USA). The results were compared to the reference toxin DX which was implemented at 1 µM. A multifunctional reader EnSpire (PerkinElmer, Waltham, MA, USA) was used to measure the absorbance.

Passive Avoidance Test
The experimental procedure was performed based on the previously described study [30]. The main equipment used in the test was the step-through passive avoidance task apparatus (LE872, Panlab Harvard Apparatus, Spain) containing the dark and the light compartment. In the acquisition session on the first day of the test, each mouse was placed in the light compartment and punished by an electric foot shock (0.2 mA for 2 s, Shocker LE100-26) when it entered into the dark compartment. In the retention session (24 h later) each animal was observed up to 180 s and the latency to the entrance to the dark compartment was measured. The latency time to the entrance to the dark compartment in the retention session close to the maximum possible value was considered proper memory performance.

Formalin Test in Mice
The mice were administered (ip) the tested compound or vehicle, and after 30 min, 20 µL of a 2.5% formalin solution was injected intraplantarly into the right hind paw. Immediately after the formalin injection, the animals were placed individually into glass beakers and were observed over the next 30 min. The time that the animal spent licking or biting the injected paw was measured during selected intervals: 0-5 min (I phase), and 15-30 min (II phase) post-formalin injection [55,56].
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules28010238/s1, Figure S1: RMSD changes for ligand 6 within AChE active site during molecular dynamics simulation; Figure S2: Distances between ligand 6 (ether O atom) and Tyr121 (OH group) from AChE during molecular dynamics simulation; Figure S3: RMSD changes for ligand 6 within BuChE active site during molecular dynamics simulation; Figure S4: Changes of ligand 6 positions within MAO B active site during molecular dynamics simulation; Figure S5: Distances between ligand 6 (CO group) and selected residues: Gln206(NH), Tyr398(OH), Tyr435(OH) from MAO B during molecular dynamics simulation; Figure S6: MS/MS ion fragmentation analyses of 6 and its metabolites.