Synthesis and Monoamine Oxidase Inhibitory Activity of Halogenated Flavones

Abstract

Small molecule neurotransmitters containing amines are metabolized by monoamine oxidase (MAO) in the nervous system. Monoamine oxidase inhibitors are a valuable class of drugs prescribed for the management of neurological disorders, including depression. A series of halogenated flavonoids similar to the dietary flavonoid acacetin were designed as selective MAO-B inhibitors. MAO-A and -B inhibition of 36 halogenated flavones were tested. The halogens (fluorine and chlorine) were placed at positions 5 and 7 on ring A of the flavone scaffold. All compounds were selective MAO-B inhibitors with micro- and nanomolar IC₅₀ values. Compounds 9f, 10a–c, 11a–c, 11g,h, and 11l displayed inhibitory activity toward MAO-B with IC₅₀ values between 16 to 74 nM. We conclude that halogenated flavonoids are promising molecules in pursuit of developing new agents for neurological disorders.

Parkinson’s disease (PD) patients are predominantly affected by clinical symptoms, such as rigidity, resting tremor, and bradykinesia, related to the gradual increase in misfolded α-synuclein protein leading to the formation of Lewy bodies. The subsequent progressive degeneration of dopaminergic neuronal synapses of the substantia nigra brain region leads to depletion of striatal dopamine and an increase in inflammation, followed by a recruitment of activated immune cells and a loss of selective permeability of the blood–brain barrier (BBB).^1,2 Mutations of the LRRK2, PARK7, PINK1, PRKN, or SNCA genes are associated with the disease and failure to degrade misfolded alpha-synuclein via the proteasome.³ While it is generally accepted that oxidative stress, neuroinflammation, and mitochondrial dysfunction (as well as deficiency of neurotrophic factors) are related to PD pathophysiology, the way in which these factors contribute to the onset of the disease state remains poorly defined.⁴ Oxidative stress in certain brain regions is one major contributing factor in PD pathophysiology, which is the outcome of an imbalance between reactive oxygen species (ROS) production and cellular antioxidant action. In the substantia nigra pars compacta (SNc) region of the brain, ROS can be generated in excess through dopaminergic processes, including dopamine metabolism and dopamine conversion to neuromelanin, and this ROS generation can be exacerbated through glutathione deficiencies and elevated levels of iron and calcium ions. Additionally, heightened levels of polyunsaturated fatty acids contained in brain membrane components are susceptible to lipid peroxidation under oxidative stress. Dopaminergic neurons are known to express a relatively high content of ROS-provoking enzymes, such as tyrosine hydroxylase (TH) and monoamine oxidase (MAO), and these enzymes are especially prone to oxidative stress.⁴ The use of natural antioxidants, including flavonoids or polyphenols, can be useful to combat this oxidative stress.⁵

Monoamine oxidases (MAO) are riboflavin proteins distributed on the cytoplasmic surface of mitochondria.^6,7 MAOs catalyze the oxidative deamination of monoamine neurotransmitters, including dopamine, 1-phenylethylamine, 5-hydroxytryptamine, and norepinephrine.⁸ In this process, molecular oxygen is the electron acceptor, which is reduced to hydrogen peroxide by these enzymes.⁹ In vertebrates, there are two MAO isoenzymes: MAO-A and MAO-B. These two enzymes display a high level of amino acid sequence identity (∼70%) yet still exhibit varying substrate specificities and expression patterns.^7,9 MAO-A is predominately found in the platelets, heart, and gastrointestinal (GI) tract. Thus, MAO-A promotes the breakdown of food products, which contain tyramine, thereby avoiding hypertensive issues due to potential excessive generation of tyramine (i.e., cheese reaction).⁸ Complementarily MAO-B is most active toward arylalkyl amines, for example, benzylamine, thus contributing in the metabolism of exogenous aromatic amines, which prevents their potential function as false neurotransmitters. The key amino acids for the recognition of substrates and enzymatic activity by the FAD cofactor in the aromatic cage of the active site of human MAO-B are Cys172, Tyr435, Tyr326, Tyr60, Tyr398, Ile198, Ile199, Tyr188, and Leu171.^6,7,10 Serotonin, epinephrine, and norepinephrine are deaminated preferentially by MAO-A, whereas MAO-B preferentially metabolizes β-phenylethylamine and benzylamine. The aminergic compounds dopamine, tyramine, and tryptamine are metabolically deaminated by both enzymes.

Several nonselective MAO inhibitors have been FDA-approved to treat depression, including isocarboxazid, phenelzine, and tranylcypromine, among others. Meanwhile, selective MAO-B inhibitors, including rasagiline, selegiline, and safinamide, are drugs developed to treat PD.⁷ Although MAO-B inhibitors are used to treat PD, they act irreversibly, which has drawbacks and is spurring new investigations toward the design of reversible MAO-B inhibitors.¹¹ Chalcones and flavones were reported as potent and selective inhibitors of MAO-B.¹¹⁻¹⁶ Recently, Iacovino et al. synthesized a series of noncytotoxic chalcones with nanomolar K_i values and high selectivity toward MAO-B isoenzymes.¹⁶ Dietary flavones, including acacetin, baicalein, luteolin, apigenin, and nobiletin, showed beneficial effects in protecting dopamine neurons in PD models in vitro and in vivo.^4,17 Low bioavailability of flavonoids results from poor absorption, rapid excretion, and the extensive conjugation and biotransformation that occur during their absorption.^18,19 Novel strategies to enhance the bioavailability of dietary flavonoids have been developed, which comprises unusual delivery systems, including solid dispersion, emulsions, crystal engineering preparation, phytosome and cyclodextrin complexes, nanocrystal, nanoparticles, polymeric micelles, and liposomes, among others.¹⁹ Additionally, derivatization of the flavonoid or related compounds core to enhancing bioavailability has been explored, including the use of aliphatic rings, heterocyclic rings, alkyl, and halogen atoms.²⁰⁻²²

Previously, we designed a series of acacetin 7-O-methyl ether derivatives with high selectivity toward MAO-B on the basis of fragment-based drug design.²³ In the current study, we explore the synthesis of acacetin 7-O-methyl ether derivatives with halogen atoms (fluorine and chlorine) on ring A at position 5 or/and 7 of the flavone nucleus and evaluate their MAO-A and -B inhibitory assays. Furthermore, computational simulations were used to study their binding pocket interactions.

On the basis of our previous predicted binding model and synthetic design, we found that propyloxy, isopropyloxy, isobutyloxy, and propargyloxy substituents at the 4′-position in the acacetin 7-O-methyl ether showed high inhibition toward MAO-B accompanied by approximately 1000-fold higher selectivity in comparison with MAO-A.²³ Thus, 36 novel flavone derivatives with fluorine or chlorine at positions 5 or 7 in ring A were synthesized, 9–11 (a–l). The synthetic route to access the novel halogenated flavones was based on Claisen–Schmidt condensation of the aromatic aldehydes with the appropriate 2′-hydroxyacetophenone (Scheme 1) to form the chalcone precursor (Scheme 2). Those chalcones were subjected to cyclization to form the targeted flavones (Figure 3).^23,24

Scheme 1. Synthetic Route for the Synthesis of the Acetophenone Precursors 3–6.

Scheme 2. Synthetic Route for the Synthesis of the 2′-Hydroxychalcones.

Figure 3.

Binding pose of compounds 10a (orange) vs 10g (green) in the orthosteric pocket of monoamine oxidase B (gray) with FAD coenzyme (magenta) [PDB ID: 2V60]. Steric clash between phenol of TYR435 and compound 10g is emphasized by dotted red line and mesh ligand surfaces.

The 3-halogenated-methoxyphenols were precursors for the 2′-hydroxyacetophenone units. Thus, 3-fluoro-methoxyphenol was acetylated with acetic anhydride and pyridine to form 3-fluoro-5-methoxyphenyl acetate (1). Compound 1 was then subjected to a Fries rearrangement using boron trifluoride diethyl etherate as a Lewis acid.^25,26 Interestingly, both isomeric acetophenone compounds (3) and (5) were formed in an approximately 1:1 ratio. The same was found for 3-chloro-5-methoxyphenyl acetate (2) in forming the acetophenones (4) and (6) (see Scheme 1).

Chalcones 7 and 8 (a–l) were synthesized through Claisen–Schmidt condensation by reacting acetophenones 3–6 and the corresponding 4-benzaldehydes (Scheme 2) using sodium hydroxide in ethanol.^23,27 The commercially available benzaldehydes used were 4-propoxybenzaldehyde, 4-isopropoxybenzaldehyde, 4-isobutyloxybenzaldehyde, 4-propargyloxybenzaldehyde, 4-allyloxybenzadehyde, and 4-isobutylbenzaldehyde.

The flavone derivatives 9 and 10 (a–l) were achieved by cyclization of chalcones 7 and 8 (a–l) using iodine crystals in the minimal amount of dimethyl sulfoxide (DMSO) (Scheme 3). The methoxylated flavones 10a–l were subjected to a demethylation reaction with boron tribromide to obtain the flavones 11a–l (Scheme 3).²⁸

Scheme 3. Synthetic Route for the Synthesis of Targeted Flavones Derivatives.

The 36 flavone derivatives 9–11 (a–l) were evaluated for their potential to inhibit MAO-A and -B and for their selectivity. The MAO-A inhibitors harmine and clorgyline and the MAO-B inhibitors deprenyl and safinamide were used as positive controls (Table 1). All compounds evaluated exhibited selective inhibition toward MAO-B. The compounds most active for MAO-A were 10c and 11f with IC₅₀ values of 1.09 and 1.03 μM, respectively. In contrast, methoxylated flavones with fluorine at position 5 (9a–e) showed no activity for MAO-A at the highest concentration tested (IC₅₀ values > 100 μM). As a group, the chlorinated flavones showed lower activity for MAO-A, including eight analogues with IC₅₀ values greater than 100 μM and three more above 50 μM. Compounds 10i and 11k were the exceptions with MAO-A IC₅₀ values of 2.33 and 0.40 μM, respectively (Table 1).

Table 1. Comparisons between MAO-A and MAO-B Inhibition by Halogenated Flavones.

fluorine flavone	MAO-A (IC50 μM)	standard error	MAO-B (IC50 μM)	standard error	selectivity index MAO-A/MAO-B	chlorine flavone	MAO-A (IC50 μM)	standard error	MAO-B (IC50 μM)	standard error	selectivity index MAO-A/MAO-B
9a	>100	0.188	0.130	0.035	>769.23	9g	>100	0.142	0.124	0.069	>806.45
9b	>100	NDa	0.104	0.052	>961.53	9h	65.98	0.085	0.129	0.095	512.67
9c	>100	0.061	0.135	0.042	>740.74	9i	9.891	0.121	0.332	0.039	29.77
9d	>100	0.155	27.620	0.091	>3.62	9j	38.13	0.166	17.740	0.100	2.15
9e	>100	0.257	0.198	0.058	>505.05	9k	>100	0.231	0.809	0.059	>123.6
9f	36.75	0.126	0.016	0.042	2274.13	9l	34.63	0.114	0.243	0.040	142.28
10a	23.85	0.044	0.020	0.090	1170.26	10g	14.26	0.308	1.100	0.069	12.96
10b	9.144	0.094	0.026	0.032	352.64	10h	80.92	0.025	0.727	0.052	103.24
10c	1.096	0.031	0.056	0.035	19.56	10i	2.339	0.110	0.320	0.041	7.31
10d	>100	0.097	10.170	0.044	>9.83	10j	>100	0.039	7.392	0.063	>13.52
10e	8.231	0.059	0.367	0.065	22.43	10k	>100	0.074	3.156	0.058	>31.68
10f	36.86	0.054	0.308	0.064	119.68	10l	>100	0.207	4.803	0.061	>20.82
11a	57.22	0.088	0.074	0.025	771.16	11g	>100	0.058	0.019	0.029	>5263.15
11b	12.04	0.073	0.017	0.023	711.58	11h	>100	0.121	0.067	0.017	>1492.53
11c	67.27	0.106	0.054	0.021	1237.72	11i	>100	0.041	0.159	0.072	>628.93
11d	96.75	0.044	0.187	0.034	518.21	11j	79.89	0.057	0.336	0.095	238.12
11e	5.233	0.175	0.105	0.034	49.93	11k	0.402	0.188	0.173	0.021	2.33
11f	1.035	0.115	0.149	0.036	6.93	11l	>100	0.051	0.027	0.025	>3703.7
harmine (reversible inhibitor MAO-A)	0.042	0.062	ND			deprenyl (irreversible MAO-B inhibitor)	ND		0.027	0.016
clorgyline (irreversible inhibitor MAO-A)	0.018	0.026	ND			safinamide (reversible MAO-B inhibitor)	ND		0.014	0.009

^aND = Not determined.

Examination of MAO-B potency revealed that 29 flavone derivatives showed an IC₅₀ < 1 μM, and 10 of these compounds had an IC₅₀ < 100 nM. More of the fluorinated flavones showed MAO-B inhibition < 100 nM than the chlorinated analogues (7 vs 3) with similar or lower IC₅₀ values than the positive controls. However, the chlorinated compounds showed better selectivity indices than the fluorinated compounds (>5263–1492 vs 2274–1170), respectively (Table 1).

The presence of halogens at position 5 in series 9 showed IC₅₀ values for MAO-B < 100 nM, except for compound 9f in which fluorine at C-5 showed a potent IC₅₀ value of 16 nM, and replacing fluorine with chlorine in 9l caused a 15-fold loss of potency for MAO-B (Table 1). The fluorinated flavones in the methoxylated series 10 in which a methoxy group occupies position 5 of the flavone core showed higher affinities than the equivalent chlorinated flavones; thus, analogues with propoxyl, isopropoxyl, and isobutyl substituents at the 4′-position (10a–c) showed the highest potencies with MAO-B IC₅₀ values of 20, 26, 56 nM, respectively (Table 1, Figure 1).

Figure 1.

Log curve of inhibition for top 11 flavones, 8 fluorinated flavones, and 3 chlorinated flavones compared with the inhibitors deprenyl and safinamide (n = 3–6). IC₅₀ was calculated by nonlinear regression log(inhibitor) vs normalized response using software GraphPad Prism v.5.03.

All series 11 fluorinated or chlorinated analogues displayed IC₅₀ values lower than 187 nM for MAO-B. These results gave us extra validity to our previous design in which it is important to have a free hydroxyl group at position 5 in the acacetin 7-O-methyl ether core.²³ The propyloxy derivative at position 4′ (11g) showed the lowest value for the chlorinated series with an IC₅₀ of 19 nM and the highest selectivity for MAO-B (>5000-fold) for all tested compounds (Table 1). Interestingly, changing the chlorine with fluorine (11a) decreased the affinity for MAO-B by 4-fold and increased the inhibition for MAO-A (Table 1). The fluorinated isopropyloxy derivative 11b showed a potent inhibition of MAO-B (17 nM) and a selectivity index of 711. It is also noteworthy that the fluorinated flavones series 10 and 11 (a–f) showed higher IC₅₀ values for MAO-A and yet more modest SI values, which may be due to the fact that fluorine atoms are prone to have weak intermolecular hydrogens bonds with different residues.²⁹ The two propargyloxy derivatives (d) were found to have a low solubility, which likely caused their highest values in the series (Table 1). In general, there was not a marked effect (Table 1) in the presence or absence of an oxygen substituent as in the isobutyl derivatives 9–11 (c or i) in comparison with the isobutyloxy derivatives 9–11 (f or l), which suggests that the presence of oxygen at position 4′ does not markedly affect binding.

We next evaluated potential cytotoxic effects of the synthesized flavones because any kind of cytotoxicity is undesirable in molecules for neurological disorders such as Parkinson’s. Thus, 11 flavones with significant inhibitory activity for MAO-B (with IC₅₀ values < 105 nM, Table 1) were also evaluated for their antiproliferative activity in three cell lines, MDA-MB-231 cells (breast cancer), HEK293 (human embryonic kidney cells), and SH-SY5Y (neuroblastoma). Table 2 shows the effect on viability represented in GI₅₀ (growth inhibitory dose 50%) values, which were determined using the MTT assay after 3 days of treatment.

Table 2. Cell Growth Inhibition (72 h) in the Presence of Selected Flavones.

	GI50 (μM)
compound	MDA-MB-231	HEK293	SH-SY5Y
9f	>20	10.20 ± 0.05	>20
10a	>20	8.58 ± 0.06	15.53 ± 0.05
10b	16.07 ± 0.06	9.90 ± 0.04	14.38 ± 0.05
10c	10.04 ± 0.06	6.80 ± 0.04	17.47 ± 0.04
11a	9.424 ± 0.07	11.05 ± 0.07	>20
11b	>20	>20	17.01 ± 0.08
11c	>20	>20	>20
11e	14.75 ± 0.04	17.46 ± 0.07	>20
11g	8.684 ± 0.04	9.59 ± 0.06	>20
11h	>20	>20	>20
11l	18.23 ± 0.04	7.02 ± 0.04	>20

The selected flavones showed moderate effects on viability with a range of GI₅₀ values from 7.02 to >20 μM in all cell lines tested. The flavones showed, in general, more cytotoxicity on the MDA-MB-231 and HEK293 lines (Table 2). Interestingly the flavones showed minimal toxicity on the neuroblastoma cell line SH-SY5Y in comparison with the other two cell lines. Series 11 displayed no antiproliferative effect on the SH-SY5Y cell line at the doses tested, except 11b (isopropyloxy substituent at 4′) with a GI₅₀ value of 17 μM.

Molecular docking was carried out (Table 1S) to better understand the binding mode of the flavone derivatives. Assessment of the virtually docked binding poses for series 9 to 11 was estimated with Schrodinger Maestro 13.1 through an induced fit docking program (Figures 1S–4S). All the inhibitors were found to occupy the binding pocket of MAO-B. Compounds belonging to series 11 exhibited a differential binding pose in MAO-B compared with compounds from series 9 and 10 (see the Supporting Information). When the methoxy group on the ligand is deprotected to form a phenol, the result was a flip in orientation within the MAO-B binding pocket, which led primarily to altered hydrogen bonding networks and aromatic π–π stacking interactions. As seen in the ligand density image, compound 11g extends further toward the FAD coenzyme, while compound 9g and 10g are located more parallel to the coenzyme (Figure 2). The differences in binding pocket interactions are best visualized between compounds 10g and 11g, where the phenol on 11g forms a new hydrogen bond with Tyr435, along with a new hydrogen bond between the ligand carbonyl and the thiol group of Cys172. This pattern of forming new hydrogen bonds with Tyr435 and Cys172 was consistent among all compounds from 11 (Figure 1S). The water molecule network also seemed to be altered by this flip in orientation, as the chlorine atom seemed to facilitate a new halogen–hydrogen bond between the ligand and the FAD coenzyme. Aromatic hydrogen-bonding networks were also disrupted in series 11. Tyr326 formed a conserved network of π–π stacking and aromatic H-bonding with the carbonyl of ligands 9g and 10g. However, in the case of 11g, there is an optimal aromatic π–π stacking interaction between the lone aromatic ring of the ligand and Tyr326. Furthermore, this residue (Tyr326) seems to be an important determinant of MAO-B activity.^30,31

Figure 2.

Binding poses of compounds 9g (orange), 10g (green), and 11g (pink) in the orthosteric pocket of MAO-B (gray) with FAD coenzyme (magenta) and water molecules (aqua); GLN215 hidden for clarity [PDB ID: 2V60].

Therefore, on the basis of our computational models, ligands that form hydrogen-bonding networks with Tyr435 and Cys172 exhibit greater inhibition. Binda et al. reported that Tyr435 is an integral part of the “aromatic cage” that holds water molecules in proximity to the coenzyme, thereby forming the hydrophilic recognition pocket for amines.⁹ Mutagenesis studies in MAO-A and MAO-B binding pockets confirmed that Tyr435 and Tyr398 are both critical binding residues for substrate recognition.^32,33 This could explain why compounds from series 11 consistently led to the inhibition of the enzyme. It is also interesting to note that the docking scores are very close between compound series, which suggests that the affinity for these compounds toward MAO-B may be similar. In other words, while this conformational difference may not lead to enhanced affinity, it may be one explanation for the relative increase in activity observed for series 11.

We also sought to better understand how a single atomic change (i.e., exchanging fluorine for a chlorine atom in the ligand) might affect the activity of these ligands with MAO-B. We observed that for series 10, the fluorinated compounds (a–f) generally exhibited greater inhibition of MAO-B activity than their chlorinated analogues (g–l) with a dramatic difference in activity between compounds 10a and 10g (Figure 3). Our models showed that this result might be best explained by the size difference between the two halogens. As seen in Figure 3, the fluorine atom appears to be small enough to fit in the binding pocket without creating any major stereochemical clashes. However, when this fluorine atom is replaced with a chlorine, our models show this larger halogen sterically clashes with Tyr435 of MAO-B. Interestingly, this result was not observed in compounds from series 11, which preferred to form hydrogen bonds with Tyr435 because of the presence of the free phenolic hydroxy group.

Compound 11k displays a unique binding pose within MAO-A compared with that of all other compounds. It is interesting to note that this molecular orientation resembles the conformation of the 11 series within the binding pocket of MAO-B (see Figure 4). Therefore, it is logical to hypothesize that this orientation of the ligand within the MAO-A binding pocket might elicit a conformation of the enzyme that leads to greater inhibition, which might explain the lower selectivity index of this flavone (Table 1). This rotation within the binding pocket, apparently induced by changing the -propyloxy constituent to an -allyloxy group, might also allow the chlorine atom to fit better by eliminating potential stereochemical clashes with the nearby Tyr407 and Tyr444 residues and water molecules.

Figure 4.

Binding pose of compounds 11k (orange) vs 11g (green) vs 11l (pink) in the orthosteric pocket of monoamine oxidase A (gray) with FAD coenzyme (magenta) and water molecules (aqua) [PDB ID: 2Z5X].

The conformational differences elicited by 11k within the MAO-A binding pocket can best be visualized through a comparison of Tyr69, Leu337, Ile335, Ile325, and Ile180 residues. Tyr69 typically forms an aromatic hydrogen bonding network with FAD, which seems to be disrupted in this case by the additional torsion caused by the newly formed aromatic hydrogen bond with compound 11k. The remaining hydrophobic residues (Leu337, Ile335, Ile325, and Ile180) also seem to be important determinants of MAO-A activity since they are conserved within the binding sites of rat MAO-A and also human MAO-B.³¹ These differences observed in the docking of 11k within MAO-A could explain the difference in selectivity for this compound and provide additional hints related to the mechanism of monoamine oxidase specificity.

Our study indicated that the flavone series 11 in which a hydroxyl is located at position C-5; a halogen, either fluoride or chloride, is at position C-7 in the ring A of the flavone; and ramified alkyl substituents, either isopropyloxy, isobutyloxy, or isobutyl, at C-4′ in the ring B of the flavone must be considered to further develop novel MAO-B-selective inhibitors. Additional studies will further investigate these analogues.

Currently, there are at least three drugs in commercial use for the treatment of PD: selegiline, rasagiline, and safinamide. However, these drugs produce highly undesirable and, in some cases, fatal side effects. Also, neurological diseases such as PD are chronic degenerative diseases in which patients require long periods of treatment, which causes a development of tolerance to drugs. For this reason, in this research work, the development of halogenated flavonoids derived from natural products with a low toxicity profile and high MAO-B inhibitory capacity has been sought for exploration and further development as a potential long-term treatment in patients with PD.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.3c00573.

• Full details on materials and methods; biological, synthetic, and computational methods used in this work; copies of ¹H and ¹³C NMR for target compounds; and HPLC spectra for representative compounds (PDF)

Author Contributions

J.I.C.-A. synthesized the halogenated flavonoids. J.I.C.-A., together with M.D.W., developed the in vitro evaluations of MAO-A and MAO-B inhibition for flavonoids. Z.S. developed the molecular docking experiments for flavonoids and developed the HPLC method. F.L. developed the original idea and supervised the process of this research. All authors reviewed, corrected, and contributed to the discussion of the manuscript.

This research was funded in part by grant number 5P20GM109091 from the National Institute of General Medical Sciences (NIGMS), a component of the National Institutes of Health (NIH); its contents are solely the responsibility of the authors and do not necessarily represent the official view of the NIGMS or the NIH. It was also funded by ASPIRE: Advanced Support for Innovative Research Excellence Program of the University of South Carolina given to F.L.

The authors declare no competing financial interest.