Dual functional cholinesterase and MAO inhibitors for the treatment of Alzheimer's disease: synthesis, pharmacological analysis and molecular modeling of homoisoflavonoid derivatives.

Abstract Because of the complexity of Alzheimer's disease (AD), the multi-target-directed ligand (MTDL) strategy is expected to provide superior effects for the treatment of AD, instead of the classic one-drug-one-target strategy. In this context, we focused on the design, synthesis and evaluation of homoisoflavonoid derivatives as dual acetyl cholinesterase (AChE) and monoamine oxidase (MAO-B) inhibitors. Among all the synthesized compounds, compound 10 provided a desired balance of AChE and hMAO-B inhibition activities, with IC50 value of 3.94 and 3.44 μM, respectively. Further studies revealed that compound 10 was a mixed-type inhibitor of AChE and an irreversible inhibitor of hMAO-B, which was also confirmed by molecular modeling studies. Taken together, the data indicated that 10 was a promising dual functional agent for the treatment of AD.


Introduction
Alzheimer's disease (AD), characterized by memory loss and other cognitive impairments, is currently one of the most difficult and baffling diseases to treat [1][2][3] . In the past decades, various pathogenesis hypothesis of AD have been proposed, such as cholinergic hypothesis 4,5 , amyloid cascade hypothesis [6][7][8] , oxidative stress hypothesis [9][10][11][12] , tau protein hypothesis 13,14 and so on. Among them, cholinergic hypothesis was a widely accepted theory, which suggests that the low level of ACh in specific regions of the brain is the major cause results in learning and memory dysfunctions. Based on the cholinergic hypothesis, one possible approach to treat AD is to restore the level of acetylcholine, such as using reversible inhibitors to inhibit acetyl cholinesterase (AChE). Up to now, AChE inhibitors, including tacrine (Cognex Õ , Figure 1), donepezil (Aricept Õ , Figure 1), rivastigmine (Exelon Õ , Figure 1) and galanthamine (Reminyl Õ , Figure 1) are approved by Food and Drug Administration to use in clinic, which could effectively improve memory in some AD patients 15,16 . The crystal structures of AChE with inhibitors revealed that there are two binding sites, a peripheral anionic site (PAS) and a catalytic active site (CAS) [17][18][19] . The PAS region of AChE, located at the rim of the gorge, was involved in the interaction with Ab peptide, which may induce the deposition of neurotoxic Ab fibrils 20,21 . Recent studies showed that AChE inhibitors, blocking both the catalytic and peripheral sites simultaneously, might not only alleviate the cognitive defect of AD patients by elevating ACh levels, but also act as diseasemodifying agents delaying amyloid plaque formation 22,23 . For example, donepezil, a mixed-type inhibitor, has a small effect on decreasing Ab aggregation induced by AChE (about 22%), which can be attributed to its higher affinity for the active site than for the PAS 24 .
Monoamine oxidases A and B (MAO-A and MAO-B), which could catalyze the oxidative deamination of neuroactive and vasoactive amines, are thought to play an important role in the metabolism of biogenic amines in the central nervous system and in peripheral tissues 25,26 . Recent studies have revealed that MAOs are associated with psychiatric and neurological disorders, including depression, Parkinson's disease (PD) and AD 27,28 . MAO-A inhibitors are used as antidepressants and anti-anxiety agents in the clinic, while MAO-B inhibitors are used as therapeutics for PD 29 . As oxidative stress has been implicated in central nervous system degenerative disorders, MAO-B inhibitors are also potential candidates as anti-AD drugs due to their regulation of neurotransmitters and capacity to inhibit oxidative damage 30 .
Because AD is a complex neurodegenerative disorder resulting from multiple factors, molecules that modulate the activity of a single protein target might not be able to significantly alter the progression of the disease. In recent years, a multi-target-directed ligand (MTDL) strategy has been developed to approach this multifaceted disease. Based on the MTDL drug design strategy, Sterling et al. incorporated the pharmacophores of rivastigmine (an AChE inhibitor, Figure 1) and rasagiline (an MAO inhibitor, Figure 2) to a single molecular entity to give a novel dual AChE and MAO-B inhibitors, ladostigil (Figure 2), the Phase II clinical trials of which is in progress 31 .
Homoisoflavonoids, which have been found to possess various biological properties, such as antifungal 32 , antiviral 33,34 , anti-proliferative 32 , antioxidant 35 and anti-inflammatory 36 activities, were also proven to be potent and selective MAO-B inhibitors 37 . Inspired by the outcomes of ladostigil, we have developed a series of tacrine-contained hybrids targeted to both AChE and MAO-B in earlier studies 38,39 . Herein, we described the synthesis and evaluation of a series of homoisoflavonoid derivatives as dual AChE and MAO-B inhibitors (Figure 3), which were expected to bind to MAO-B, PAS and CAS of AChE.

Materials and methods
Materials 1 H NMR and 13 C NMR spectra were recorded on a Bruker BioSpin GmbH spectrometer (Ö stliche Rheinbrückenstr, Germany) at 400.132 and 100.614 MHz, respectively, using tetramethylsilane as the internal standard. Coupling constants are given in Hz. MS spectra were recorded on a Agilent LC-MS 6120 instrument (Santa Clara, CA) with an ESI mass selective detector. Flash column chromatography was performed with silica gel (200-300 mesh) purchased from Qingdao Haiyang Chemical Co. Ltd. (Qingdao, China). The chemicals used in the chemistry section were procured from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China) and all of the chemicals are analytical reagents. The purities of the synthesized compounds were confirmed to be higher than 95% by analytical HPLC (Agilent technologies 1200 series system) with a TC-C8 column (4.6 Â 150 mm, 5 mm) and the compounds were eluted with CH 3 OH/water (0.1% TFA, w/v) in ratios of 70:30-40:60 at a flow rate of 1.0 mL/min.
General procedures for the preparation of 5-8 To a stirred suspension of compounds 4 (10 mmol) and K 2 CO 3 (15 mmol) in butanone (50 mL), KI (1 mmol) and a,!-dibromo alkane (12 mmol) were added. The mixture was stirred at room temperature for 1-4 h, filtered, and then evaporated under vacuum. The crude product was chromatographed on a silica gel column, eluted with EtOAc/petroleum ether as eluent to afford the proposed compound.

Kinetic characterisation of AChE inhibition
Kinetic characterisation of AChE was performed using a reported method. Test compound was added into the assay solution and pre-incubated with the enzyme at 37 C for 15 min, followed by addition of substrate (the final concentrations of substrate were 0.05, 0.0625, 0.10, 0.125, 0.25, 0.50 mM, respectively). Kinetic characterisation of the hydrolysis of ATC catalyzed by AChE was performed spectrometrically at 412 nm. A parallel control with no inhibitor in the mixture allowed the change in activities to be measured at various times. The plots were assessed by a weighted least square analysis that assumed the variance of V to be a constant percentage of V for the entire data set. The slopes of these reciprocal plots were subsequently plotted against the concentration of the inhibitors in a weighted analysis, and K i was determined as the ratio of the replot intercept to the replot slope.

Inhibition of MAO activity
The potential effects of the test compounds on hMAO activity were investigated by measuring their effects on the production of H 2 O 2 from p-tyramine, using the Amplex Red MAO assay kit (Molecular Probes, Inc., Eugene, OR) and recombinant human MAO-A or MAO-B (Sigma-Aldrich) according to published procedures 40 . Compounds were dissolved in DMSO (10 mM) and diluted in 0.1 M KH 2 PO 4 /K 2 HPO 4 buffer (pH 8.0) to the desired final concentration. All the compounds are soluble at the tested concentration. Adequate amounts of recombinant hMAO-A or hMAO-B (Sigma-Aldrich) were acquired and adjusted to 12.5 mg/mL for hMAO-A and 75 mg/mL for hMAO-B. Test drugs (20 mL) and MAO (80 mL) were incubated at 37 C for 15 min in a flat-black-bottom 96-well microtest plate in dark. The reaction was started by adding 200 mM Amplex Red reagent, 2 U/mL horseradish peroxidase and 2 mM p-tyramine for hMAO-A or 2 mM benzylamine for hMAO-B, at 37 C for 20 min. The reaction was quantified in a multi-detection microplate fluorescence reader based on the fluorescence generated (excitation, 545 nm; emission, 590 nm).
The specific fluorescence emission was calculated after subtraction of the background activity, which was determined from vials containing all components except the hMAO isoforms, which were replaced by a sodium phosphate buffer solution.

Reversibility and irreversibility experiments
To evaluate whether compound 10 is a reversible or irreversible hMAO-B inhibitor, an effective centrifugation ultrafiltration method (so-called ''repeated washing'') was used 40 . The experiment was performed according to a reported method. Adequate amounts of recombinant monoamine oxidase with or without test drugs were incubated at 37 C for 15 min. An aliquot from this incubation was stored at 4 C for later use in the measurement of MAO activity. Another aliquot was transferred to an Amicon Ultra-0.5 Centrifugal Filter Unit with Ultracel-30 membrane (Millipore) and centrifuged (4 C, 9000g, 20 min) three times. The enzyme was obtained and used for the subsequent measurement of activity using a method similar to that described above (the section ''Inhibition of MAO activity''). The corresponding values of percent hMAO-B inhibition were separately calculated for samples with and without repeated washing.

Docking study
The simulation system was built based on the structure obtained from the Protein Data Bank (PDB: 2V60 for MAO-B, 2CMF for AChE). The heteroatoms and water molecules were removed, and all hydrogen atoms were subsequently added to the protein.
Then forcefield was assigned to the enzyme. The ligand binding site was defined as 13 Å from the original ligand. Prior to the docking calculations, the original ligand was removed. The 3D structures of compound 10 were generated and optimized with the Discovery Studio 2.1 package (Accelrys Inc., San Diego, CA).
The CDOCKER program of the Discovery Studio 2.1 software, which allows full flexibility of ligands, was used to perform docking simulations. The docking and subsequent scoring were performed using the default parameters of the CDOCKER program. CDOCKER_INTERACTION_ENERGY is used like a score where a lower value indicates a more favorable binding.
Compounds 5-18 were synthesized by the following procedures. Intermediates 2, 3 and 4 were synthesized according to reference 39 .

In vitro inhibition of AChE and BuChE
The AChE (E.C.3.1.1.7) inhibitory effects of the homoisoflavonoid derivatives were determined by the spectroscopic method described by Ellman et al. using rivastigmine as the standard 41 . As shown in Table 1, most of the hybrids demonstrated potent inhibitory activity against AChE and BuChE with IC 50 values in the sub-micromolar to micromolar range. Compound 9, a piperidine moiety linked with homoisoflavonoid by two carbon spacers, gave the most potent inhibitory activity (IC 50 ¼ 0.74 mM), which was about 135-fold more potent than lead compound 4 (IC 50 4100 mM). A simple structure-activity relationship analysis showed that the AChE inhibitory potency closely related to the length of the alkylene chain. When the length of the alkylene chain increased to 5, the AChE inhibitory activity decreased dramatically (12, IC 50 ¼ 14.1 mM). From the IC 50 values of the tested compounds, it appears that a 3-carbon spacer between homoisoflavonoid moiety and amino group seems to be the proper length for balancing the inhibition activity against AChE and MAO-B. Moreover, amino group is also important for the inhibitory activities. Among the three aliphatic substituted amino derivatives, the N,N-dimethyl amino group (13, Scheme 1. Reagents and conditions: (a) 3-chloropropionic acid trifluoromethanesulfonic acid; (b) NaOH, 0 C; (c) p-methoxybenzaldehyde, piperidine, 80 C; (d) a,!-dibromo alkane, K 2 CO 3 , KI, butanone, rt; (e) amines, acetonitrile, reflux. The in vitro BuChE (E.C. 3.1.1.8) inhibition activities were also determined using the same method. Most of the tested compounds demonstrated much higher inhibitory potency against BuChE than homoisoflavonoid 4. Compound 10, which exhibited a good balance of AChE/MAO-B inhibitory activity, was chosen for further studies.

Kinetic study of AChE
In order to investigate the mechanism of homoisoflavonoid derivatives against AChE, kinetic study was performed using compound 10 and the steady-state inhibition data for AChE are shown in Figure 4. Lineweaver-Burk reciprocal plots revealed that there was an increasing slope and an increasing intercept at higher inhibitor concentrations, indicating a mixed inhibition. Based on the kinetic analysis, we concluded that homoisoflavonoid derivatives exhibited as dual binding sites AChE inhibitors, which target catalytic and peripheral sites of AChE simultaneously.

Molecular modeling studies of AChE
To explore the interaction mode of homoisoflavonoid derivatives with AChE and the structure-activity relationships, molecular docking simulations were performed with the Discovery studio 2.1 software based on the structure of the complex of Torpedo

In vitro inhibition studies of hMAO-A and hMAO-B
To confirm the multipotent biological profile of homoisoflavonoid derivatives, inhibitory activity against MAOs was determined using ladostigil as reference compound and the results are summarized in Table 1. Most of these target compounds exhibited high selective inhibition to hMAO-B in the micromolar range. Among them, compound 12 presented the most potent inhibition for hMAO-B with IC 50 value of 0.65 mM, which is 2-fold and 57fold more potent than lead compound 4 and ladostigil, respectively. A structure-activity relationship analysis showed that the length of carbon spacer also played an important role in inhibition activity. However, unlike that of AChE, a longer linker resulted in better inhibitory activity, the inhibition activity against MAO-B increased as the increasing of the carbon spacer length. For example, compound 9, in which piperidine group linked with homoisoflavonoid via a 2-carbon spacer, provides IC 50 value of more than 25 mM. Its homologous compound 12, with 5-carbon spacer, gave 0.65 mM of the IC 50 value, which was 40 times as that of compound 9. On the other hand, all of the target compounds showed very weak inhibition activity against MAO-A. Overall, among the synthesized compounds, compound 10 exhibited best balance of inhibition for both AChE and MAO-B.

Reversibility and irreversibility study of hMAO-B
The reversibility and irreversibility of inhibitory activities of 10 was assessed according to a method of reference 40 . Pargyline, known as irreversible MAO-B inhibitor, was used as a reference compound. Data are listed in Table 2. Similarly with pargyline, MAO-B inhibition of 10 was irreversible for the enzyme activity, which did not restore after repeated washing.

Molecular modeling studies of MAO-B
To evaluate the binding modes of the homoisoflavonoid derivatives with MAO-B, docking simulation of compound 10 was carried out using the CDOCKER program in the Discovery studio 2.1 software based on the X-ray crystal structure of human   MAO-B (PDB entry 2V60). As shown in Figure 6, the benzyl group of homoisoflavonoid moiety is properly oriented to the flavin adenine dinucleotide (FAD) cofactor adopting between Tyr398 (4.25 Å ) and Tyr435 (4.33 Å ) form face-to-face p-p stacking interactions in a ''sandwich'' form. Besides, the pyridine moiety and carbon linker are embedded into a hydrophobic pocket delimited by Pro102, Pro104, Trp119, Leu164, Leu167, Phe186, Ile316.

Conclusion
In summary, we have designed, synthesized and evaluated a series of homoisoflavonoid derivatives as dual functional cholinesterase and MAO inhibitors for the treatment of AD. Most of these compounds were potent inhibitors of ChE, which are more active than lead compound 4. Meanwhile, compound 12 presented the most potent inhibition for hMAO-B with IC 50 value of 0.65 mM. Compound 10, with a pyridine group linked with homoisoflavonoid moiety by a three carbon linker, provided a good balance of AChE and hMAO-B inhibition activities, with IC 50 value of 3.94 and 3.44 mM, respectively. Kinetic studies revealed that compound 10 was a mixed-type inhibitor of AChE and an irreversible inhibitor of hMAO-B. Molecular modeling studies indicated that 10 could interact with both CAS and PAS of AChE, which is consistent with kinetic studies. Based on these results, compound 10 is a promising ''one-compound-multi-targets'' candidate for AD treatment and needs further study.