Novel Class of Chalcone Oxime Ethers as Potent Monoamine Oxidase-B and Acetylcholinesterase Inhibitors

Previously synthesized novel chalcone oxime ethers (COEs) were evaluated for inhibitory activities against monoamine oxidases (MAOs) and acetylcholinesterase (AChE). Twenty-two of the 24 COEs synthesized, except COE-17 and COE-24, had potent and/or significant selective inhibitory effects on MAO-B. COE-6 potently inhibited MAO-B with an IC50 value of 0.018 µM, which was 105, 2.3, and 1.1 times more potent than clorgyline, lazabemide, and pargyline (reference drugs), respectively. COE-7, and COE-22 were also active against MAO-B, both had an IC50 value of 0.028 µM, which was 67 and 1.5 times lower than those of clorgyline and lazabemide, respectively. Most of the COEs exhibited weak inhibitory effects on MAO-A and AChE. COE-13 most potently inhibited MAO-A (IC50 = 0.88 µM) and also significantly inhibited MAO-B (IC50 = 0.13 µM), and it could be considered as a potential nonselective MAO inhibitor. COE-19 and COE-22 inhibited AChE with IC50 values of 5.35 and 4.39 µM, respectively. The selectivity index (SI) of COE-22 for MAO-B was higher than that of COE-6 (SI = 778.6 vs. 222.2), but the IC50 value (0.028 µM) was slightly lower than that of COE-6 (0.018 µM). In reversibility experiments, inhibitions of MAO-B by COE-6 and COE-22 were recovered to the levels of reference reversible inhibitors and both competitively inhibited MAO-B, with Ki values of 0.0075 and 0.010 µM, respectively. Our results show that COE-6 and COE-22 are potent, selective MAO-B inhibitors, and COE-22 is a candidate of dual-targeting molecule for MAO-B and AChE.


Introduction
Alzheimer's disease (AD) is one of the greatest concerns confronting the medical community, and is the fourth leading cause of neurodegenerative disease-related death. Furthermore, AD has been predicted to affect 100 million patients within 30 years [1]. AD accounts for 70% of all reported cases of dementia, being characterized by cholinergic functional decline, β-amyloid oligomer formation, and the dysregulations of other cellular processes [2]. Over past years, many efforts have been made to identify the key biochemical events responsible for AD. However, AD is a multifactorial disease and, thus, its management requires the simultaneous modulations of multiple targets [3]. Based on greater understanding of the disease, recent research efforts have increasingly focused on multitarget-drugs that simultaneously bias different biological targets [4].
This novel approach is viewed optimistically, and hybridizations of the pharmacophore subunits of bioactive molecules have already resulted in the identification of multifunctional drugs [5] and, as a result, synthetic drugs, like donepezil, rivastigmine, and tacrine, have been used as structural models for molecular hybridization experiments ( Figure 1) [6]. Tacrine was the first cholinesterase (ChE) inhibitor that was approved by the FDA for the treatment of AD. However, the use of tacrine is limited by its side-effects and, thus, searches for more compatible and potent tacrine derivatives continue [7]. multitarget-drugs that simultaneously bias different biological targets [4].
This novel approach is viewed optimistically, and hybridizations of the pharmacophore subunits of bioactive molecules have already resulted in the identification of multifunctional drugs [5] and, as a result, synthetic drugs, like donepezil, rivastigmine, and tacrine, have been used as structural models for molecular hybridization experiments ( Figure 1) [6]. Tacrine was the first cholinesterase (ChE) inhibitor that was approved by the FDA for the treatment of AD. However, the use of tacrine is limited by its side-effects and, thus, searches for more compatible and potent tacrine derivatives continue [7].
On the other hand, monoamine oxidase (MAO)-A is primarily targeted for the treatment of depression and anxiety, whereas MAO-B is targeted for AD and Parkinson's disease, based on their specificity, which is, MAO-A prefers serotonin, and MAO-B prefers phenylethylamine and benzylamine [8]. Rasagiline is a MAO inhibitor, and its neuroprotective activity has been attributed to the presence of a propargyl amine moiety, which suppresses the overexpression of Bax protein in AD [9].
The complexity of AD militates against the use of consolidated mono-therapies and supports the notion that dual MAO and acetylcholinesterase (AChE)-inhibitory activities are likely to have better therapeutic effects in AD [10]. Ladostigil is an example of such multi-functional drugs, as it possesses the neuroprotective effects of rasagiline and ChE inhibitory activity ( Figure 1) [11]. Notably, most drugs used to treat AD patients in palliative care settings are ChE inhibitors with some multifunctional activity. Furthermore, many studies have shown that MAO inhibitors have attracted considerable research interest in the context of halting or retarding the progression of AD [4]. Chalcones are versatile scaffolds and they are widely distributed in edible plants. Several attempts have been made to synthesize novel biologically active chalcone derivatives due to their wide-ranging biological activities [12][13][14][15][16]. Over recent decades, the MAO-B inhibitory activities of chalcone derivatives have progressively appreciated [17], and many studies have reinforced the On the other hand, monoamine oxidase (MAO)-A is primarily targeted for the treatment of depression and anxiety, whereas MAO-B is targeted for AD and Parkinson's disease, based on their specificity, which is, MAO-A prefers serotonin, and MAO-B prefers phenylethylamine and benzylamine [8]. Rasagiline is a MAO inhibitor, and its neuroprotective activity has been attributed to the presence of a propargyl amine moiety, which suppresses the overexpression of Bax protein in AD [9].
The complexity of AD militates against the use of consolidated mono-therapies and supports the notion that dual MAO and acetylcholinesterase (AChE)-inhibitory activities are likely to have better therapeutic effects in AD [10]. Ladostigil is an example of such multi-functional drugs, as it possesses the neuroprotective effects of rasagiline and ChE inhibitory activity ( Figure 1) [11]. Notably, most drugs used to treat AD patients in palliative care settings are ChE inhibitors with some multifunctional activity. Furthermore, many studies have shown that MAO inhibitors have attracted considerable research interest in the context of halting or retarding the progression of AD [4].
The activated aryl bromides included aryl bromides bearing electron-withdrawing groups at the 4-position and bromo-chalcones. Screening phosphine ligands, Pd-catalyst, and solvents was utilized to optimize the method.

Inhibitory Activities against MAO-A, MAO-B and AChE
The MAO-A, MAO-B, and AChE inhibitory activities of 24 synthesized COEs were evaluated while using toloxatone, lazabemide, clorgyline, pargyline, and tacrine as reference molecules ( Table  1). The synthesized COEs were of two structural categories, that is, chalcone ketoxime or chalconechalcone oxime hybrids. Nineteen of the 24 COEs showed residual MAO-B activities of <50% at 1. The activated aryl bromides included aryl bromides bearing electron-withdrawing groups at the 4-position and bromo-chalcones. Screening phosphine ligands, Pd-catalyst, and solvents was utilized to optimize the method.
Twenty three of the 24 COEs showed residual MAO-A activities of >60% at 1.0 µM, but only COE-13 had a residual activity of <50% at 1.0 µM (Table 1) and also inhibited MAO-B well and it had the highest SI value, which suggest its possible use for the dual-targeting of MAO-B and AChE.

SARs for Inhibition Studies
Twenty two of 24 COEs included in the present study were selective MAO-B inhibitors ( Table 1)

Kinetics of MAO-B Inhibitions
Kinetic studies were performed on MAO-B inhibition by COE-6 and COE-22. Lineweaver-Burk plots and secondary plots showed that COE-6 and COE-22 competitively inhibited MAO-B ( Figure 3A,C) with K i values of 0.0075 ± 0.00067 and 0.010 ± 0.0035 µM, respectively ( Figure 3B,D). These results suggest that COE-6 and COE-22 are potent, selective, and competitive inhibitors of MAO-B.

Reversibility Studies
Reversibility studies were conducted on MAO-B inhibition by COE-6 and COE-22. In these experiments, inhibitions of MAO-B by COE-6 and COE-22 were recovered from 19.7 (A U ) to 81.1% (A D ) and from 22.6 (A U ) to 86.8% (A D ), respectively (Figure 4), and these values were similar to those of the reversible reference inhibitor lazabemide (from 2.4 to 76.4%). However, inhibition by the irreversible reference inhibitor pargyline was only slightly recovered (from 3.7 to 10.4%). These experiments showed that inhibitions of MAO-B by COE-6 and COE-22 were recovered to the reversible reference level, which suggested that both are reversible inhibitors.

Computational Studies
Computational analyses were performed using QM-polarized docking and MM-GBSA calculations in order to investigate the binding modes of COE-6 and COE-22 towards MAO-A and MAO-B and with the purpose of clarifying the MAO-B selectivity of the two compounds. Table 2  Compound COE-22 interacts with MAO-A and MAO-B with different binding modes ( Figure 5). The residues of MAO-A involved in COE-22 binding are Tyr62 and Lys218, which establish π and cation-π interactions, respectively, with the para-fluorine phenyl ring, Lys341, which engages a hydrogen bond with the carbonyl oxygen atom of the chalcone portion of COE-22, and Lys316, which forms a cation-π interaction with the para-methoxy phenyl ring, as shown on panel (a) of Figure 5. Notably, the distance between the para-fluorine phenyl ring and aromatic rings of the flavin adenine dinucleotide (FAD) molecule is~11 Å. On the other hand, as shown on panel (b) of Figure 5, the residues of MAO-B involved in COE-22 binding are similar to those hypothesized in previous studies [25], whereby the para-fluorine phenyl ring of COE-22 is trapped within an aromatic cage made up of FAD, Tyr398, and Tyr435. Furthermore, the para-fluorine phenyl ring, the para-methoxy styryl, and the chalcone aromatic ring establish π−π interactions with Tyr398, Trp119, and Tyr236 (MAO-B selective residue), respectively. In addition, the carbonyl oxygen of the chalcone scaffold of COE-22 forms a hydrogen bond with the thiol group of Cys172.

Kinetics of MAO-B Inhibitions
Kinetic studies were performed on MAO-B inhibition by COE-6 and COE-22. Lineweaver-Burk plots and secondary plots showed that COE-6 and COE

Reversibility Studies
Reversibility studies were conducted on MAO-B inhibition by COE-6 and COE-22. In these experiments, inhibitions of MAO-B by COE-6 and COE-22 were recovered from 19.7 (AU) to 81.1% (AD) and from 22.6 (AU) to 86.8% (AD), respectively (Figure 4), and these values were similar to those of the reversible reference inhibitor lazabemide (from 2.4 to 76.4%). However, inhibition by the irreversible reference inhibitor pargyline was only slightly recovered (from 3.7 to 10.4%). These experiments showed that inhibitions of MAO-B by COE-6 and COE-22 were recovered to the reversible reference level, which suggested that both are reversible inhibitors.

Computational Studies
Computational analyses were performed using QM-polarized docking and MM-GBSA calculations in order to investigate the binding modes of COE-6 and COE-22 towards MAO-A and MAO-B and with the purpose of clarifying the MAO-B selectivity of the two compounds. Table 2   Compound COE-22 interacts with MAO-A and MAO-B with different binding modes ( Figure  5). The residues of MAO-A involved in COE-22 binding are Tyr62 and Lys218, which establish  and cation- interactions, respectively, with the para-fluorine phenyl ring, Lys341, which engages a hydrogen bond with the carbonyl oxygen atom of the chalcone portion of COE-22, and Lys316, which forms a cation- interaction with the para-methoxy phenyl ring, as shown on panel (a) of Figure 5. Notably, the distance between the para-fluorine phenyl ring and aromatic rings of the flavin adenine dinucleotide (FAD) molecule is ~11 Å . On the other hand, as shown on panel (b) of Figure 5, the residues of MAO-B involved in COE-22 binding are similar to those hypothesized in previous studies [25], whereby the para-fluorine phenyl ring of COE-22 is trapped within an aromatic cage made up of FAD, Tyr398, and Tyr435. Furthermore, the para-fluorine phenyl ring, the para-methoxy styryl, and the chalcone aromatic ring establish − interactions with Tyr398, Trp119, and Tyr236 (MAO-B selective residue), respectively. In addition, the carbonyl oxygen of the chalcone scaffold of COE-22 forms a hydrogen bond with the thiol group of Cys172.   Proteins are rendered as white cartoons, while ligands are rendered as yellow sticks. Green, blue and red arrows indicate π−π interactions, cation-π interactions and hydrogen bonds, respectively. Y326, responsible for MAO-B selectivity, is labeled red.
Docking analysis did not report meaningful differences for interactions of COE-6 towards MAO-A and MAO-B. The para-fluorine phenyl ring of COE-6 is involved in π−π interaction with Tyr407 of MAO-A, and it is trapped within an aromatic cage delimited by Tyr407, Tyr444, and FAD, unlike COE-22, as shown on panel (a) of Figure 6. When considering MAO-B, the para-fluorine phenyl ring of COE-6 engages π−π interaction with Tyr435 and faces the aromatic cage that is formed by Tyr435, Tyr398, and FAD. Notably, the selective MAO-B residue Tyr326 establishes π−π interaction with the chalcone aromatic ring, and Cys172 makes a hydrogen bond with the carbonyl oxygen of the COE-6 chalcone scaffold.
Docking studies carried out on COE-6 and COE-22 have proved that compound COE-22 had the highest MAO-B affinity and appreciable selectivity. More specifically, COE-22 interacts with MAO-A and MAO-B, but with different binding modes. In particular, in agreement with previous findings [24][25][26][27][28][29], the chalcone head of COE-22 faces the FAD of MAO-B, whereas COE-6 adopts similar poses for MAO-A and MAO-B, probably because of its smaller size. Interestingly, docking studies successfully explained at the molecular level the different experimental affinities of COE-6 and COE-22 for the two MAO isoforms. In particular, the gain in binding for MAO-B was mostly supported by the chance of forming π−π hydrophobic interaction with Tyr326. This is a key residue, which changed to I335 in MAO-A [49], capable of giving access to the binding pocket (for COE-22 compound) and stabilizing the chalcone aromatic ring.
Docking analysis did not report meaningful differences for interactions of COE-6 towards MAO-A and MAO-B. The para-fluorine phenyl ring of COE-6 is involved in − interaction with Tyr407 of MAO-A, and it is trapped within an aromatic cage delimited by Tyr407, Tyr444, and FAD, unlike COE-22, as shown on panel (a) of Figure 5. When considering MAO-B, the para-fluorine phenyl ring of COE-6 engages − interaction with Tyr435 and faces the aromatic cage that is formed by Tyr435, Tyr398, and FAD. Notably, the selective MAO-B residue Tyr326 establishes − interaction with the chalcone aromatic ring, and Cys172 makes a hydrogen bond with the carbonyl oxygen of the COE-6 chalcone scaffold. Docking studies carried out on COE-6 and COE-22 have proved that compound COE-22 had the highest MAO-B affinity and appreciable selectivity. More specifically, COE-22 interacts with MAO-A and MAO-B, but with different binding modes. In particular, in agreement with previous findings [24][25][26][27][28][29], the chalcone head of COE-22 faces the FAD of MAO-B, whereas COE-6 adopts similar poses for MAO-A and MAO-B, probably because of its smaller size. Interestingly, docking studies successfully explained at the molecular level the different experimental affinities of COE-6 and COE-22 for the two MAO isoforms. In particular, the gain in binding for MAO-B was mostly supported by the chance of forming − hydrophobic interaction with Tyr326. This is a key residue, which changed to I335 in MAO-A [49], capable of giving access to the binding pocket (for COE-22 compound) and stabilizing the chalcone aromatic ring.

Analysis of Enzyme Inhibitions and Kinetics
The inhibitory activities of the 24 COEs synthesized against MAO-A and MAO-B were first investigated at a concentration of 1.0 µM, and IC 50 values were then determined. AChE inhibitory activities were also determined, except at a concentration of 10 µM. Time-dependent inhibitions and reversibilities were measured, and kinetic studies were performed on the most potent MAO-B inhibitors, i.e., COE-6, and COE-22, as previously described [53]. Kinetic experiments were carried out at five substrate and three inhibitor concentrations.

Computational Studies
The three-dimensional (3D) structures of MAO-A (PDB ID: 2Z5X) and MAO-B (PDB ID: 2V5Z) were obtained from the Protein Data Bank. The protein preparation wizard available in the Schrödinger suite was used to optimize X-ray crystal structures [55,56]. MAO-A and MAO-B active sites contained nine and eight water molecules, respectively. The LigPrep tool was used to optimize ligand structures and generate possible tautomers and ionization states at physiological pH. Docking simulations were carried out using the QM polarized ligand docking protocol available from Schrödinger Suite. While retaining the rigidities of protein structures, QM polarized ligand docking allows for ligands with a certain degree of conformational flexibility. Centers of mass of X-ray cognate ligands of MAO-A and MAO-B structures were used as references for the cubic grid center.
The QM-polarized ligand docking protocol that was implemented in Glide was used with default options. This protocol uses three computational steps, that is: a) a standard precision (SP) initial docking using Glide; b) calculation of QM partial charges of the docked ligand based; and, c) a SP re-docking phase for each ligand pose when considering computed QM based charges.
A Molecular Mechanics/Generalized Born Surface Area (MM-GBSA) method was added to the workflow for the calculation of the binding free energies (∆G) between protein and ligands in order to estimate ligand-binding affinities. Such a method is implemented in Prime available in the Schrodinger software 2018-2 (New York, NY, USA) [57]. Provided that ∆E MM is the minimized energy of the ligand-protein complex, ∆G solv is the solvation energy, and ∆G SA is the binding energy of the surface area of compounds, with respect to MAO-A and MAO-B, ∆G bind values were computed, as follows: Obtained docking poses were minimized using Prime [57][58][59].

Conclusions
We