Dual monoamine oxidase B and acetylcholine esterase inhibitors for treating movement and cognition deficits in a C. elegans model of Parkinson’s disease

Abstract

Parkinson’s disease (PD) is an age-associated neurodegenerative movement disorder that leads to loss of dopaminergic neurons and motor deficits. Approaches to neuroprotection and symptom management in PD include use of monoamine oxidase B (MAO-B) inhibitors. Many patients with PD also exhibit memory loss in the later stages of disease progression, which is treated with acetylcholine esterase (AChE) inhibitors. We sought to identify a dual-mechanism compound that would inhibit both MAO-B and AChE enzymes. Our screen identified a promising compound (7) with balanced MAO-B (IC₅₀ of 16.83 μM) and AChE inhibition activity (AChE IC₅₀ of 22.04 μM). Application of this compound 7 increased short-term associative memory and significantly prevented 6-hydroxy-dopamine toxicity in dopaminergic neurons in the Caenorhabditis elegans nematode. These findings present a platform for future development of dual-mechanism drugs to treat neurodegenerative diseases such as PD.

Introduction

Parkinson’s disease (PD) is a progressive neurodegenerative disease characterized by the loss of the dopaminergic (DAergic) neurons in the substantia nigra [1–6]. Loss of DAergic neurons leads to characteristic PD symptoms of bradykinesia, rigidity, and tremors [7]. As the disease progresses, patients also experience cognitive deficits and memory loss, which are pharmacotherapeutically treateded with acetylcholine esterase (AChE) inhibitors such as donepezil. These AChE inhibitors inhibit the catabolism of AChE in the synapse therey increasing the concentration of AChE. Idiopathic PD is commonly associated with aging (>55 years old), but a genetic origin contributes to incidence of PD in younger patients [3]. Several studies implicate environmental (e.g., insecticides) and genetic factors (e.g., LRRK2, PARKIN, PINK1, and DJ1) in PD development, while reactive oxygen species, which are in part produced by mitochondria, are implicated in PD progression [4, 5].

Since dopamine (DA) is the major neurotransmitter lost in PD, the movement symptoms of PD are treated with the dopamine precursor L-Dopa, alone or in combination with monoamine oxidase B inhibitors. Monoamine oxidase (MAO) is a flavin-containing enzyme that catalyzes DA metabolism. MAO is located on the outer mitochondrial membrane and can be found in two isoforms, A and B [8]. MAO-B is an important drug target for PD treatment; inhibition of this enzyme promotes elevated DA levels in the striatum and improves motor symptoms [9, 10]. Although initially used only to treat PD symptoms, it was later found that MAO-B inhibition plays a role in neuroprotection, especially against the effects of environmental toxins, with classic examples such as 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) [11, 12]. Several studies have suggested that MAO-B levels increase with age indicating that MAO-B enzymes may play a role in neurodegeneration [13, 14].

The concept of designing or developing a single compound which interacts with two drug targets is referred to as dual or multi-mechanism drugs [15, 16]. An example of this approach was the development of ladostigil (TV-3,326), which is a combination of rivastigmine (AChE inhibitor) and rasagiline (MAO-B inhibitor) [17]. Main advantages of this approach is enhanced therapeutic outome of the disease state, use of a single agent reduces drug interactions and decreased toxicity profiles. In the case of Parkinson’ disease, the onset of dementia leads to the use of AChE inhbitors, which increases the levels of acetylcholine associated with memory function [18]. Therefore, a dual mechanism MAO-B/AChE inhibitor would be of therapeutic advantage. In this study, we screened a series of compounds which have not been evaluated before to identify novel dual-mechanism drugs inhibiting both MAO-B and AChE, which could serve as hit compounds for new drug discovery programs to treat both movement and memory symptoms in Parkinson’s patients. These compounds were chosen based on a central scaffold six-membered ring fused with a five membered ring, that is found in both the MAO-B inhibitor zonisamide, as well as in AChE inhibitor donepezil (Figure 1).

Fig. 1.

Selected compounds were based on premise of common core (shown in blue) found between the MAO-B inhibitor zonisamide, and the AChE inhibitor donepezil.

Results and Discussion

Our goal was to identify novel dual-mechanism compounds with balanced targeting of both MAO-B and AChE to simultaneously treat the movement and memory symptoms in Parkinson’s disease patients. Previous studies have indicated that it would is feasible to design dual-mechanism drugs, as reviewed elsewhere [19]. The design of dual or multi-targeted ligands have several advantages, one of which is a single agent pharmacokinetic profile with potentially fewer off-target effects in contrast to polypharmacy [20, 21, 15].

We screened a small pilot set of ten compounds in our collection for MAO-B and AChE enzyme inhibition (Figure 2, Table 1 and Table 2), based on common ring cores (Figure 1) seen between the MAO-B inhibitor, zonisamide and the AChE inhibitor, donepezil [19]. Serendipidously, we found eight compounds inhibited MAO-B enzyme activity, with only one compound also inhibiting MAO-A. For the tested comounds, the triazolepyridazine compounds displayed a wide range of activity in inhibiting MAO-B. For instance, the activity of compound 8 (MAO-B IC₅₀ = 5 μM) against MAO-B was slightly more potent than that of compound 7 (MAO-B IC₅₀ = 16.38 μM), which has an additional oxygen between the chlorinated benzene ring and the piperazine, and that of compound 9 (MAO-B IC₅₀ = 83 μM), which has an increased chain length between the halogenated benzene and the piperazine. Furhter analysis of compound 7 inhibition of recombinant MAO-B, showed a dose-dependendent decrease of the Michaeles-Menten substrate curves, and Lineweaver-Burke transormantion indicated a competitive mechanism of inhibtion (Figure 3). Lastly, diaylis of the enzyme-inhibitor complex resulted in recovery of the enzyme activity, corroborateing a compeative-reversible inhibition (Figure 4). Laslty, the positive control compounds zonisamide (MAO-B) and tranylcypromine (MAO-A) were used [22].

Fig. 2.

Structures of compounds evaluated in this study for dual MAO-B/AChE inhibition.

Table 1.

In silico physico-chemical properties of the compounds tested in this study.

	Compound ID	clogP	TPSA (Å2)	Drug likeness (Lipinski Ro5)
1	F0012–0121	4.40	98.47	Yes
2	F0536-0298	3.22	64.68	Yes
3	F0647-0340	3.26	66.07	Yes
4	F0647-0350	3.28	66.07	Yes
5	F2609-0119	3.55	64.68	Yes
6	F2616-0708	1.81	76.66	Yes
7	F5123-0117	1.58	75.86	Yes
8	F5123-0122	1.78	66.63	Yes
9	F5313-0560	1.80	66.63	Yes
10	F5313-0599	1.73	66.63	Yes

Abbreviation: total polar surface area (TPSA)

Table 2.

Activity profiles of the screened compounds. Data are shown as enzyme inhibition (IC₅₀) values in μM.

	Compound ID	MAO-A	MAO-B	AChE	BSA
1	F0012-0121	1.74	40.12	ia	12.66
2	F0536-0298	ia	19.71	ia	>100
3	F0647-0340	ia	ia	ia	95.12
4	F0647-0350	ia	0.55	ia	38.54
5	F2609-0119	ia	ia	ia	40.1
6	F2616-0708	ia	0.243	ia	>100
7	F5123-0117	ia	16.38	22.04	>100
8	F5123-0122	ia	5.55	35.28	>100
9	F5313-0560	ia	83.09	ia	>100
10	F5313-0599	ia	96.97	ia	>100
Zonisamide			22.6
Tranylcypromine		3.7
Indomethacin					19.72
Diclofenac					16.05

i.a.: inactive at dose tested.

Fig. 3.

Michaeles-Menten kinetic analysis of the inhibition of recombinant MAO-B by compound 7. Substrate concentration of kynuramine was microM. A) Michaeles-Menton substrate curve; B) Lineweaver-Burk plot of inhibition. Compound 7 was incubated at microM.

Fig. 4.

Effect of equilibrium dialysis on recombinant MAO-B activity activity in the presence of compound 7 tested at the IC₅₀ concentration. Catalytic active was recovered in part after the dialysis period. Each bar represents average ± SD where N = 4.

Docking studies were done to gain insight into the binding interaction with MAO-B, and the results indicated that compound 7 spans the entrance cavity and the substrate cavity of MAO-B (Figure 5A and 5B). The entrance cavity gate-keeper ILE199 was used to delineate the two cavities. A water-bridge interaction between compound 7 and LYS296 of MAO-B is predicted from the docking studies to contribute to the overall interaction between compound 7 and MAO-B enzymes. Finally, we validated our docking strategy by redocking the original co-crystal ligand of MAO-B, and found the root square mean deviation (RMSD) to be less than 2 Å, which provided support of using this method to study compound 7’s binding intereaction (Figure 6).

Fig. 5.

Docking studies of compound 7 with MAO-B and AChE enzymes. A) Compound 7 occupies the entrance (indicated by ILE199) and substrate cavities, B) forms a water-bridge to LYS296, C) and can have a docked pose in AChE. D) Compound 7 also interacts with TRP86, TRP286, or TYR341 in AChE via primary hydrophobic interactions.

Fig. 6.

Root-square mean derivative (RMSD) results of the control dcoking experiments. The co-crystalized ligand of each enzyme was docked (shown in green) into the orginal structure (original ligand shown in atom colors). An RMSD<1.5Å was found for each enzyme. A) MAO-B (2v5z.pdb); B) 4EY7.pdb.

Fewer compounds were found to effectively inhibit AChE (Table 2). We screened for inhibition of AChE enzyme activity using the well-established Ellman’s reaction as the primary readout in the assay. Only two of the compounds tested showed AChE inhibition activity, with inhibition of the AChE enzyme in the micromolar range, Compound 7 was the most potent inhibitor of AChE enzyme in this set with an AChE inhibition IC₅₀ of 22.04 μM. Further, the AChE inhibition activity of compound 7 was the most balanced and close to its MAO-B inhibition activity (MAO-B IC₅₀ = 16.38 μM), which would enable simpler formulation for in vivo pharmacology since both drug targets fall within the same range for target engagement. Docking studies showed that compound 7 is able to undergo aromatic-type interactions with Tyr341 and Trp286 in the AChE enzyme (Figure 5C and 5D). Additionally, the chlorine in compound 7 was able to undergo a salt-bridge-type interaction with the water in the catalytic cavity of AChE. Based on our goal of identifying a dual MAO-B/AChE inhibitor, we identified compound 7 as our lead for the downstream biological phenotypical characterization. Figure 6 shows a RMSD of less than 2Å for reproducing the co-crystal structure binding orientation, as control experiment of the docking procedure.

To gain perspective on the compounds activity profiles, we additionally tested for antioxidant properties of the compounds, which can augment the protection of neuronal cells during neurodegeneration. Since reactive oxygen species (ROS) play an import role in neurodegeneration, we screened the compounds using the oxygen radical absorbance capacity antioxidant assay (ORAC) with fluorescein as the fluorescent marker. In this assay, AAPH is used as free radical generator which reduce the fluorescein fluorescence intensity. Four of the tested compounds displayed moderate antioxidant activity (Figure 7). Taken together with their moderate antioxidant activity, the tested compounds could serve as lead compounds for the development of multifunctional drugs [15, 23].

Fig. 7.

Anti-oxidant activity of the compounds tested using the ORAC assay format. In this assay, the chemical AAPH generates radicals that decrease the fluorescence of fluorescein. In the presence of an antioxidant, the fluorescence persists during the experiment. Vitamin C is used as positive control, with a EC50 of 25.95 μM. Screen of compounds were done at 12 μM. Each data point is average ± SD.

Considering the role of pharmacokinetic (PK) properties in the drug discovery process [24], we evaluated several of these parameters using the online server SwissADME. Additionally, we evaluated the interaction to serum albumin (Table 1 and Table 2). These data suggest that the free fraction of the compounds will be available for target engagement if administered in vivo, i.e. if target engament required at <100 μM, then adequate compound will likely be unbound. As control compounds, we used indomethacin and ibuprofen, as these non-steroidal anti-inflammatroy drugs (NSAIDS) are know to be bound to serum albumin [25, 26]. The compounds screened were also predicted to be drug-like by adhering to Lipinski’s rule of 5. Aearly adoption of ADME evaluation in the drug discovery process can significantly decrease percentage of compounds failing in the early clincial trials [27].

Dual-mechanism evaluation of the lead compound 7 was done utilizing the soil nematode C. elegans, which is widely used in high content screening of disease models [28]. Compound 7 had the desired properties of target engagement (Table 1), with similar potencies for both MAO-B and AChE (Table 2), which will improve and accelerate down-stream future formulation and toxicology studies. We therefore evaluated the pharmacological profile on memory and neurotoxicity using the dopaminergic neurotoxiciy 6-OH-DA assay, as well as the butanone short-term memory association in vivo models.

Compound 7 prevented the neurotoxic effects of 6-OH-DA in C. elegans, demonstrated by the attenuation of loss of fluorescence as indication of loss of the GFP-labeled DAT neurons in worms treated with 6-OH-DA plus compound 7 (30 μM ) compared to worms treated with 6-OH-DA alone (Figure 8). Incubation of C. elegans worms with 30 μM, chosen based on the rounded IC₅₀ value of the two enzymes of compound 7 increased their butanone short-term associative learning, suggesting that this compound may be useful in further studies regarding PD-associated dementia (Figure 9).

Fig. 8.

Protective effect of compound 7 against 6-OH-DA toxicity in C. elegans. A) Control; B) 6-OH-DA (20 μM); C) compound 7 (30 μM) and 6-OH-DA (20 μM). All figures are 40X magnification. Top panels show DAT::GFP-labeled neurons indicated with white arrows, and bottom panels show bright-field images.

Fig. 9.

Compound 7 improves short-term associative memory in C. elegans. Number of animals attracted to butanone after the initial aromatic memory training period when treated with compound 7 (30 μM). Each bar represents avg. ± S.D. *p<0.05.

Conclusion

We identified a novel dual-mechanism inhibitor of both MAO-B and AChE enzymes. This compound displayed several key characteristics that are favorable for the development of neuroprotective drugs, such as mild antioxidant activity, in silico predicted blood-brain barrier permeability, and passing a drug-likeness filter. The lead compound from this study also has promising in vivo activity for improving cognition changes in late-stage Parkinson’s disease dementia, while attenuating neurogeneration in the C. elegans model. Future studies will focus on chemical diversification of the compound 7 scaffold to explore structure-activity trends, with the goal of exploring these novel compoudns as possible candidates for treatment in Parkinson’s disease.

Experimental

Chemicals

Compounds tested in this study were obtained from Life Chemicals. The screening libray compounds were obtained as 1 mg powder in vials, <95% pure, as confirmed by NMR or/and LCMS. See supplmental data for spectra. All other compounds were purchased form commercial sources.

MAO Enzyme Assay

MAO enzyme inhibition assays (MAO-A and MAO-B) were performed using recombinant human enzyme (BD Genetest) as described previously with minor modifications.⁵ To measure MAO enzyme activity, kynuramine was employed as a substrate.⁶ As kynuramine is metabolized by MAO, it forms a fluorescent metabolite 4-hydroxyquinoline. Assays were performed in a 96-well plate format using a BioTek Synergy 4 plate reader (λ_ex= 310 nm; λ_em= 380 nm) and a monochromator. For the enzyme assays, MAO-A and MAO-B were each dissolved sepearately to final concentrations of 6 μg/mL and 15 μg/mL, respectively, in 0.1 M phosphate buffer (pH 7.4). Kynuramine was dissolved in water to final concentrations of 40 μM and 20 μM for the MAO-A and MAO-B assays, respectively. Small molecules being tested in this study (obtained from Life Chemicals, Figure 1) were dissolved in DMSO to make stock solution of 10 mM at a final concentration of 2% DMSO. Compounds were incubated with the MAO enzyme and kynuramine for 20 min, after which the reaction was quenched with the addition of 2N NaOH. IC₅₀ values were calculated using Prism 5 statistical software (www.graphpad.com) from one-site binding using eight different concentrations spanning five log units, each performed in duplicate. Tranylcypromine and zonisamide were used as positive controls [29, 30].

Enzyme kinetics

To determine the effect of compounds on substrate utilization of recombinant MAO-B, the enzyme assay was incubated with different concentraions of kynuramine. The compound of intersest was used at different concentraions, and the resulted substate-velocity Michaeles-Menten curves were transforemed to the reciprocal Lineweaver Burke plots, and evaluated for competitive or non-competative interaction [31].

MAO-Reversibilty Assay

Recombinant MAO-B enzyme at the same concentration as for the inhibition assay, was incubated with compound 7 for 30 min, afterwhich the mixture was dialysed against phosphate buffer saline (pH 7.4). Control enzyme (DMSO vehicle control) was run simoultanously and the activity determined before and after dialysis [31].

Acetylcholine esterase assay

The inhibition of AChE was determined was previously described [32]. AChE (827 U/mL) was dissolved in 100 mM phosphate buffer pH 7.4. In a 96 well plate, each well contained 100 μL of the compound tested, 50 μL of a 0.65 mg/mL acetylthiocholine, 50 μL of AChE, and 50 μL of DTNB/Ellman’s reagent. After 10 min, the yellow color was measured at 412 nm in a Cytation 5 BioTek multimodal platereader.

Cheminformatics and Docking Studies

In silico prediction of logP, total polar surface area (TPSA), and drug likeness (Lipinski Rule of 5) were determined using the SwissADME server [33]. Docking studies were performed using MOE 2019 (Chemical Computing Group). The protein structure referenced for MAO-B was 2v5z.pdb and for AChE 4ey7.pdb. In the MAO-B enzyme file two chains are crystalized and chain B was deleted before docking was carried out. Protein was prepared before docking by protonation of amino acids at pH 7.4. The binding site was identified as the area where co-crystallized ligand was located. Since MOE recognizes FAD as a part of the ligand set, we first designated the true ligand as so that MOE could use it in the docking run. Only the top-returned binding pose (most negative binding energy) of each ligand was further evaluated. The induced-fit binding mode was used to study the ligand-protein interaction, and solvent was allowed part of the docking calculation. As control docking experiment, the root-square mean deviation (RMSD) was calculated for each of the two enzymes. A RSMD < 2 Å was considered adeqote for our studies [34].

Antioxidant Assay

The oxygen radical absorbance capacity (ORAC) assay was used to assess the potential of compounds to sacavage peroxyl radicals based on previous methods [35]. AAPH (2,20-azobi(2-amidinopropane)-dihydrochloride) radical generation was initiated by the addition of a 72 mM stock to the drug solution to give a final concentration of 36 mM AAPH in each well which contained a 80 uL of a 112 nM fluorescein solution. The resultant fluorescence was measured 1 hour later using the 485/525 ex/em filters on a Cytation 5 BioTek multimodal platereader. Vitamin C was used as positive antioxidant control.

Bovine Serum Albumin Binding

Binding of the test compounds to bovine serum albumin (BSA) was determined as described previously [36]. In brief, the drugs were dissolved in 100 μL of 100 mM phosphate buffer pH 7.4 and serially diluted in a 96-well black plate. We added 100 μL of BSA to each well to reach a final concentration of 7.5 μM BSA. Fluorescence quenching was measured at excitation 280 nm and emission 340 nm using a Synergy 4 multimodal plate reader (www.biotek.com). Sodium diclofenac and indomethacin were used as control compounds.

Behavior Assay to Assess Memory in Caenorhabditis elegans

We used the Bristol N2 (wild type) strain of C. elegans obtained from the Caenorhabditis Genetics Center at the University of Minnesota, which is funded by the NIH Office of Research Infrastructure Programs (P40 OD010440). Worms were maintained under standard conditions (20 °C). Effects of compound 7 on memory in C. elegans were determined as described previously [37]. Briefly, worms were staged using bleach, and L1 stage worms were kept on nematode growth media (NGM) plates seeded with the bacterium E. coli OP50. When worms reached the L4 stage, they were washed from the plates using 3–4 mL standard M9 nematode buffer, collected in a conical tube, and allowed to settle under gravity. Worms were then washed twice with M9 buffer. After washing, worms were starved for one hour, then transferred to an NGM plate seeded with OP50 with 10% butanone in EtOH streaked (10 μL) on the plate lid. Worms were kept on plates for one hour, removed as before, washed twice with M9 buffer, and placed on NGM plates with OP50 for one hour. Two spots were marked on different sides of the NGM plate, and 2 μL of 10 mM sodium azide were placed on each spot to paralyze the worms. Ethanol and 10% butanone in ethanol were also spotted on the marks as chemo-attractants. A total of 100–200 worms were then placed at an origin marked spot and allowed to move around for one hour. Then, the number of worms at each spot of either the ethanol or butanone was counted. The assay was repeated twice, with two replicates in each trial.

Parkinson’s Disease 6-Hydroxydopamine (6-OH-DA) Toxicity Assay in C. elegans

EGL1 worms with a dopamine transporter and green fluorescent protein co-expresssed in the dopaminergic neurons with a co-expression of the dopamine transporter (DAT) and a green fluorescent protein (GFP) (DAT::GFP), were also obtained from the Caenorhabditis Genetics Center. Worms were kept and maintained under standard conditions (20 °C) as previously described [37]. To determine whether the lead compound from our study, compound 7, could protect against 6-OH-DA neurotoxicity, we plated worms at a density of ~50 worms per well on a 96-well plate using M9 buffer. OP50 was added to each well to prevent starvation in addition to compound 7 (30 μM) and 6-OH-DA (20 μM). Worms were incubated for 4 days and then assessed for neurodegeneration using fluorescence microscopy. The level of green GFP fluorescence is used as marker for dopaminergic neuronal cell less, where a loss of these neurons will lead to a decrease or even loss of the GFP fluorescence.