Design, synthesis, and evaluation of guanylhydrazones as potential inhibitors or reactivators of acetylcholinesterase.

Abstract Analogs of pralidoxime, which is a commercial antidote for intoxication from neurotoxic organophosphorus compounds, were designed, synthesized, characterized, and tested as potential inhibitors or reactivators of acetylcholinesterase (AChE) using the Ellman’s test, nuclear magnetic resonance, and molecular modeling. These analogs include 1-methylpyridine-2-carboxaldehyde hydrazone, 1-methylpyridine-2-carboxaldehyde guanylhydrazone, and six other guanylhydrazones obtained from different benzaldehydes. The results indicate that all compounds are weak AChE reactivators but relatively good AChE inhibitors. The most effective AChE inhibitor discovered was the guanylhydrazone derived from 2,4-dinitrobenzaldehyde and was compared with tacrine, displaying similar activity to this reference material. These results indicate that guanylhydrazones as well as future similar derivatives may function as drugs for the treatment of Alzheimer's disease.


Introduction
Acetylcholinesterase (AChE) is a very important enzyme used to control transmission between neurons 1,2 , when the process is either mediated or modulated by the neurotransmitter acetylcholine (ACh). ACh is released by the axon terminal or varicosities of the transmitter neuron into the extracellular spa ce to interact with the receptors of the other neuron. To maintain control of neurotransmission, it is necessary for AChE, after ACh executes its function, to catalyze ACh hydrolysis, converting ACh to choline (Ch) and acetate (Ac). After ACh hydrolysis, Ch is reabsorbed by the axon terminal to produce more ACh. If AChE is inhibited in the central nervous system, the concentration of ACh increases in the synaptic cleft, leading to cholinergic crisis, which affords several dangerous effects, such as convulsion and respiratory problems, which could lead to death.
AChE can be inhibited with reversible and irreversible inhibitors. The irreversible inhibitors are basically neurotoxic organophosphorus compounds, normally used as pesticides, insecticides, and chemical warfare agents 3,4 . These agents lead to the phosphorylation of Ser203, which is the most important active amino acid involved in ACh hydrolysis, at the active site of AChE, leading to complete inhibition of this enzyme 5 . Despite the use of chemical warfare agents being prohibited by the United Nations, these agents constitute one of the greatest threats in the modern world 6 . In parallel, many organophosphorus pesticides are used by all countries for agricultural production, a process that sometimes causes several toxic problems. To reactivate inhibited and phosphorylated AChE, it is necessary to use agents with the appropriate capacity to execute a nucleophilic attack on the phosphorus atom bound to Ser203, leading to the reactivation of the function of this serine. The AChE reactivator agents are mainly cationic oximes 7 , which are effective but, unfortunately, not appropriate against all neurotoxic organophosphorus agents 8 , thereby necessitating the development of new and more effective agents for the protection of all humanity.
Some reversible inhibitors of AChE have medical applications and are particularly important for the treatment of Alzheimer's disease (AD). When people develop AD, their neurons degenerate, leading to the low production of neurotransmitters, a process that induces serious memory problems. In this case, the inhibition of AChE increases the concentration of ACh in synaptic clefts, improving the neurotransmission process and brain function. For this reason, AChE inhibitors are very important agents for the treatment of AD, but some of these inhibitors are toxic, such as tacrine, requiring the development of new agents. Interestingly, some AChE reactivators also display competitive inhibition of the enzyme 8 , and the reversible inhibitor and AD drug galantamine protects animals from soman, sarin, and paraoxon intoxication 9 , suggesting that novel compounds may have dual application, for AD and organophosphorus intoxication.
Therefore, the development of new agents that interact with AChE as potential drugs for the treatment of AD or as reactivators of this enzyme for the detoxification of neurotoxic organophosphorus compounds is very important. This work describes the preparation of compounds with a certain similarity to pralidoxime and their evaluation as AChE inhibitors or reactivators using the Ellman's test, nuclear magnetic resonance (NMR) and molecular docking.

Chemistry
Solvents (ethyl alcohol 95%, diethyl ether, methanol, chloroform) were purchased from VETEC (Duque de Caxias, Brazil) (used with further purification), and reagents were purchased from Merck and Sigma-Aldrich (São Paulo, Brazil) (used without further purification). Reactions were monitored by TLC using DC-Alufolien Kieselgel 60 F254 (Merck, Darmstadt, Germany). The NMR spectra were determined on a Varian 600 MHz spectrometer, using dimethyl sulfoxide-d6 (DMSO-d 6 ) as solvent and tetramethylsilane as internal standard. The infrared (IR) spectra were measured using a Spectrum 100 spectrometer. The absorption values are expressed in wave number, using inverse centimeters (cm À1 ) as units. The mass spectra were obtained on a Waters (Milford, MA) spectrometer, model Q-Tof micro, using positive mode detection (MS ES + ) and acetonitrile as solvent. The IR, HRMS and NMR spectra are available in the Supplemental file.

Synthesis of the compounds
N-methylation method for pyridine-2-carboxaldehyde 2-Formyl-1-methylpyridinium iodide was prepared using 5.0 g pyridine-2-carboxaldehyde (0.05 mol) and CH 3 I (14 g) with stirring and heating under reflux for 2 h. The orange crystals formed were filtered under nitrogen and washed with dry diethyl ether.

Method of synthesis for 2-PAM (1)
The method for the preparation of 1-methylpyridine-2-oxime iodide (1) was developed in two stages. First prepared was the iodide of 2-formyl-1-methylpyridine. Then, the iodide of 2formyl-1-methylpyridine (1.4 mol) diluted in ethanol was reacted with hydroxylamine (0.0014 mol) in the presence of pyridine (1 mL) as catalyst. This reaction was stirred and heated under reflux for 40 min. Afterwards, distilled water (1 mL) was added, and the product was filtered.

Pralidoxime (1)
Yield 26%; Yellow solid. mp: 212-214 C (Lit. 224-225 C) 10  General procedure for hydrazone (2) The method for the preparation of 1-methylpyridine-2-hydrazone iodide (2) was developed in two stages. First prepared was the iodide of 2-formyl-1-methylpyridine. Then, the iodide of 2formyl-1-methylpyridine (0.004 mol) diluted in methanol was reacted with hydrazine hydrate (0.004 mol) in the presence of HCl (0.6 M) as catalyst. This reaction was stirred and heated under reflux for 30 min and then was cooled to 5 C and filtered. The red liquid filtrate was evaporated, and the product was recrystallized from diethyl ether.

AChE inhibition assay by Ellman's method
Inhibition of AChE by guanylhydrazones and hydrazone was performed by Ellman's colorimetric method modified for reading in a microplate spectrophotometer (SpectraMax 250, Molecular Devices, Sunnyvale, CA, USA) and at pH 7.4 15,16 . In this study, instead of using the homogenate of rat brain, enzyme purified from Electrophorus electricus was used. Equine serum butyrylcholinesterase was used to evaluate selectivity among the enzymes. To evaluate the inhibitory capacity of the compounds, a final volume of 200 mL was used and to a microplate were added

AChE inhibition assay by the NMR method
The EeAChE activity and inhibition determination by NMR was performed on a 600 MHz spectrometer using 5-mm NMR tubes, as described in our previous article on this topic 17 . To determine EeAChE activity, 2 mL of its solution of 0.2 mM concentration in pH 7.4 phosphate buffer with D 2 O and in the presence of 1% bovine serum albumin was used. At the starting time of the reaction, this solution was mixed with 30 mL of acetylcholine (0.1 M) in the same solvent, and the complete mixture was diluted to a volume of 600 mL with the D 2 O pH 7.4 phosphate buffer in the 5-mm NMR tube. This sample was immediately introduced to the magnet for locking and shimming, allowing for the observation of the first 1 H spectra exactly 5 min after the introduction of acetylcholine. All of the next 1 H spectra were obtained every 5 min with a single scan over 80 min. The intensity of the methyl signals of acetylcholine (2.24 ppm) and acetic acid (2.16 ppm) was determined by scanning, affording information about EeAChE kinetics. The AChE inhibition test was executed with the same procedure, including the addition of 5 mL of each potential EeAChE inhibitor (12 mmol), always before the addition of acetylcholine, using exactly the same time periods. The concentration of acetylcholine and Ac from each 1 H spectrum was determined by the integration of their methyl signals, using exactly the same integration length in all cases. The time used for the pure enzyme to afford 50% of Ac was also used to determine the respective Ac production (%) in the presence of the tested inhibitors. The comparison of the Ac concentration at this time with the pure AChE allows the determination of the percent inhibition of each tested compound. These results were obtained in triplicate.

Molecular modeling
The structure conformation, dipole moment, and determination of atomic electronic charges were obtained with Spartan'06 using the B3LYP system and the 6-311++G** basis set 18 .

Docking
The ligands' three-dimensional structures were built using the Spartan'10 program 19 , and the conformer distribution with molecular mechanics using 100 conformers were examined. Their partial atomic charges were calculated using the RM1 semiempirical method, and the molecular energies were calculated using Hartree-Fock 6-31 G*.  23 .

Synthetic aspects
One previous study using molecular modeling with ab initio methods suggested the evaluation of anionic nucleophiles different from oximates, for example hydrazonates, as AChE reactivators 24 . According to this molecular modeling study, we decided to prepare a hydrazone with a structure similar to pralidoxime to test its capacity as an AChE reactivator or inhibitor. Because hydrazones normally display low acidity, the binding of hydrazonates in the active site of AChE may be relatively difficult, indicating that these compounds would not be very effective reactivators of this enzyme when inhibited by organophosphorus agents. However, other hydrazones could display more acidity than the NH 2 group but certainly would be less acidic than the respective oximes. To test this possibility, the only hydrazone selected for preparation and evaluation was 1-methylpyridine-2-carboxaldehyde hydrazone (2) (Figure 1), with a structure very similar to pralidoxime (2-PAM, 1), which is one of the simplest cationic oximes used commercially for the reactivation of phosphorylated AChE and which certainly interacts with the active site of this enzyme. Considering the possibility that other cationic or neutral compounds similar to pralidoxime could interact well with the AChE active site and with inhibition capacity, we decided to prepare and test the compounds 1-methylpyridine-2-carboxaldehyde guanylhydrazone (3), pyridine-2-carboxaldehyde guanylhydrazone (4), benzaldehyde guanylhydrazone (5), 4nitrobenzaldehyde guanylhydrazone (6), 4-chlorobenzaldehyde guanylhydrazone (7), 4-methylbenzaldehyde guanylhydrazone (8), 4-dimethylaminobenzaldehyde guanylhydrazone (9), and 4-hydroxybenzaldehyde guanylhydrazone (10), as shown in Figure 1. Guanylhydrazones were selected because they contain one cationic site and therefore are appropriate to interact with the anionic region of the AChE active site and compete with acetylcholine. Several guanylhydrazones have been used as pharmacological agents for diverse diseases, such as cancer 25,26 , Chagas disease 27,28 , malaria and others 29 . These compounds were also studied by NMR and theoretical chemistry 30 .
To determine if hydrazones could act as inhibitors or reactivators of AChE, the first step was the N-methylation of pyridine-2-carboxaldehyde using methyl iodide as solvent with stirring and under reflux for 2 h. This compound was used to prepare 2-PAM (1) in 26% yield by reaction with hydroxylamine hydrochloride in ethanol with pyridine under reflux for 40 min. The cationic hydrazone 2 was prepared by the reaction of N-methylpyridine-2-carboxaldehyde with hydrazine hydrochloride under reflux in 95% ethanol for 2 h, affording the compound in 36% yield. The N-methylpyridine-2-carboxaldehyde was also used to prepare the bis-cationic guanylhydrazone (3) by reaction with aminoguanidine hydrochloride in ethanol and HCl under reflux for 2 h, affording the compound in 44% yield. All other guanylhydrazones (4-10) were prepared by reaction of the respective aldehydes with aminoguanidine hydrochloride using 95% ethanol as solvent and heating under reflux for 2-7 h, leading to mono-cationic compounds obtained in 40-94% yields. All compounds were characterized by IR, NMR and mass spectrometry.
The only previous attempt to synthesize compound 2 was made in 1961 by Poziomek, who reported that at the end of the synthesis procedure, they obtained incomplete dehydration of the molecule and not the desired product 31 . Our procedure led to the desired hydrazone after recrystallization from diethyl ether, and this compound has not been described in the literature. Compound 3 is a new unpublished agent. Compounds 4-10 have been described in the literature 11,12,32,33 , but never for their application in AChE inhibition or reactivation, and thus, this is a new type of application for guanylhydrazones.

Assessment of inhibition of EeAChE
The tests of AChE inhibition with these compounds were carried out with two methods -the Ellman's test 15,16 and the recently published NMR method 16 using tacrine as a reference standard. The selected AChE was from E. electricus (EeAChE), which displays an identical active site and very similar activity to that of the human enzyme 34,35 . We have also used the Ellman's assay for EqBuChE inhibition, to assess the compounds' selectivity. All results obtained for cholinesterase inhibition using the Ellman's test and NMR are shown in Table 1.
Analysis of the Ellman's test indicates that all compounds display EeAChE inhibition activity, confirming that guanylhydrazones and hydrazones are potential inhibitors, especially because all contain cationic groups that are capable of interacting with the active site of this enzyme.
To confirm the results obtained by the Ellman's test, 1 H NMR experiments were also carried out 17 . The NMR method required a low concentration of enzyme (1.0 pM) in phosphate buffer at pH 7.4 and with the presence of 5% albumin to maintain the enzyme activity for a long time because EeAChE is a very fast enzyme in its hydrolysis of ACh. To perfectly control the time for each 1 H NMR spectrum, it was also necessary to introduce the ACh into the 5-mm NMR tube as the last ingredient and then take 5 min to lock, gradient auto-shim and obtain the first spectra with one single transient. Continuing in this manner with additional spectra taken every 5.0 min led to the acquisition of 20 spectra during 100 min. The substrate (ACh) and product (Ac) concentrations from each spectrum were obtained from the integration of the corresponding methyl signals at 2.15 and 1.91 ppm, respectively. These experiments, which were executed first with pure EeAChE and followed with compounds 2-10, are shown in Table 1. In general terms, despite the results of both methods being similar, there are also important differences with some compounds. For example, the results for compound 5 are very similar (14.06 by Ellman's and 13.9 by NMR), but for compound 9, the results are different (16.59 by Ellman and 10.4 by NMR). However, because the NMR test allows for perfect signal determination of the AChE substrate (ACh) and products (Ch and HOAc), as well as their correct concentration by integration, this method is as appropriate as the Ellman's test; because the NMR method includes the structural determination of all agents, substrates and products, it is also a very good method.
Comparing the effects on AChE for the guanylhydrazones prepared from benzyl aldehydes in Table 1, it is observed that the compound containing one para-NO 2 group (6) is the most effective, and the second most effective is compound 7, which is parachlorinated. The third most effective compound is the paramethylated compound (8), and the less effective compounds are 5, which does not contain an R group, and 9 and 10, whose para-R groups, Me 2 N and HO, respectively, are electropositive. This information indicates that the best benzylidene guanylhydrazones for the inhibition of AChE should contain strong electron withdrawing R groups. To confirm this effect, all guanylhydrazones were studied by molecular modeling (B3LYP 6-311 + G**) to determine the correlation between some calculated properties of each compound and their percentage of AChE inhibition. In this case, the correlation of their dipole moment (Debye) with inhibition afforded an R 2 of 0.9361, as shown in Figure 2.   The correlation of the electronic charge of the nitrogen atom bound to the carbonyl aldehydes with AChE inhibition also afforded an R 2 ¼ 0.8310. These results confirm that the electron withdrawing characteristics of the R groups at the para-position on the benzene ring affect the AChE-inhibiting capacity of these guanylhydrazones.
To test this possibility after obtaining the previous results, the two compounds with more electronegative groups on the benzylidene ring, 2,4-dinitrobenzaldehyde guanylhydrazone (11) and 2,4-dichlorobenzaldehyde guanylhydrazone (12), were prepared and tested (Figure 3). The most electronegative compound, 11, displayed EeAChE inhibition of 63.79% by the Ellman's test, with an IC 50 of 11.25 mM, and 75.6% by NMR, thus being the most effective compound, being superior to compound 2. Compound 11 was also more selective for EeAChE over EqBuChE. However, the bis-chlorinated compound, (12), which is less electronegative, affords 45.49% inhibition by the Ellman's test and 41.3% inhibition by NMR, therefore being the fourth most effective inhibitor. In this case, tacrine (13) was used as the reference AChE inhibition agent.
The results are shown in Table 2, and Figure 4 is the graphic of their NMR tests. The inclusion of the dipole moments of two additional compounds 11 and 12 with their AChE activities in the correlation graph of compounds 5-10 afforded an R 2 of 0.9119. These results confirm that including more electronegative groups on the benzylidene ring increases the EeAChE inhibition capacity.
These results obtained by NMR are very similar to those obtained by the Ellman's test, confirming their capacity for Figure 5. Correlation between the AChE inhibition results using the Ellman or NMR tests.  (11) and tacrine (13) by NMR. EeAChE inhibition. It is observed that 2,4-dinitrobenzaldehyde guanylhydrazone (11) displays a very efficient capacity to inhibit EeAChE (75.6%), indicating that other similar derivatives could be more effective and that guanylhydrazones and their analogs can be used as prototypes in the preparation of new drugs. The correlation of AChE inhibition by the Ellman's test with that using NMR, as shown in Figure 5, indicate that these methods afford relatively similar results (R 2 ¼ 0.8914).

Molecular docking studies
To confirm this information, the interaction of compounds 2 and 6 in comparison with tacrine (13) was studied using AutoDock 4.2 software 21 . Because the structure of E. electricus AChE (EeAChE) has not been reported as a complex with any ligand, we used the structure of Torpedo californica AChE (TcAChE), obtained from the Protein Data Bank under the code PDB ID:1ACJ 20 , for docking. To determine the effectiveness of AutoDock 4.2, the re-docking of tacrine in the TcAChE site was executed to compare with the reported X-ray crystallographic structure of this enzyme-tacrine complex 36 . The root mean square deviation obtained for the re-docking of tacrine inside of the 3D structure of TcAChE was 0.53, confirming the effectiveness of AutoDock 4.2. After this, the docking of compounds 2, 6, 11, and 12 was performed, and their intermolecular binding energies were determined. The data obtained, in comparison with tacrine, are shown in Table 3.
These docking results indicate that these compounds are relatively effective as AChE inhibitors and demonstrate that compounds 2 and 12 have similar action but are less effective than tacrine. Despite the docking results being relatively different from the experimental results obtained by the Ellman and NMR tests  (Tables 1 and 2), these results confirm that one hydrazone and several guanylhydrazones are potential AChE inhibitors. The superposition of the lowest energy poses of tacrine and compound 11 interacting in the active site of TcAChE is shown in Figure 6. These results show that tacrine and compound 11, which is the most effective experimental AChE inhibitor of the tested compounds, interact with the active site of this enzyme, indicating that both display a very similar inhibitory capacity.
Despite the importance of docking studies, these results indicate that theoretical calculations usually are not in parallel with experimental information. The interaction of compounds 2, 6, 11, and tacrine (13) are shown in Figure 7.
We have also performed a preliminary in silico screening of drug properties of the most active compounds, using the web tools Osiris Property Predictor (http://www.organic-chemistry.org/prog/ peo/) and lazar (http://lazar.in-silico.ch/predict). Compounds 2, 6, and 11 all obey the rule-of-five principle and showed low toxicity risk, with confidence values below 0.25 in all assessments of carcinogenicity and mutagenicity. The pyridinium compounds, including pralidoxime (1) and 2, are not expected to permeate the blood-brain-barrier efficiently. However, compounds 6 and 11 showed CLogP of 0.11 and À0.81, respectively, which is compatible with brain penetration and eventual application in AD.

EeAChE reactivation
To perform the AChE reactivation tests, EeAChE was inhibited with paraoxon (100 nM) by the phosphorylation of Ser200. The complete inhibition of this enzyme was also confirmed by the Ellman's test. To test the activity of these compounds, pralidoxime (2-PAM, 1) was used as a reference material. The Ellman's tests for the reactivation of EeAChE inhibited with paraoxon confirmed that pralidoxime (1) is the only good reactivator, with the guanylhydrazones being ineffective, as shown in Table 4. As expected, 1 completely reactivated EeAChE, considering that it directly inhibited the enzyme by 24.3% at 100 mM (Table 1), so that the maximum expected activity in the reactivation assay would be 75.7%. Likewise, for compounds 2-12, only partial recovery of activity should be expected after reactivation, due to their own inhibitory properties.
The reactivation of EeAChE inhibited with paraoxon was also assayed with incubation times of 30 and 60 min but did not show significant alteration of the results.
The structural modeling ( Figure 6) indicated that the compounds can interact with the EeAChE active site. Thus, we expected that the nucleophilic groups NH 2 of 2 and the substituent OH of 10 could be deprotonated by active site residues to attack the phosphate group bound to Ser200, leading to enzyme reactivation. Hydrazones are usually not acidic compounds, but compound 2, which contains the N-methyl group on the pyridine ring, is more acidic than average. Compound 3 also failed to reactivate, in spite of the cationic pyridine ring, which would also increase the acidity and nucleophilicity of the NH group. The expected reactivation mechanism for the cationic compounds 4-12 relied on deprotonation of their guanidine groups by active site anionic amino acid residues. This would lead to neutral intermediates with some nucleophilicity at the nitrogen atoms; however, the mechanism was not effective. These negative results are relevant for the design of new agents with more activity for defense against neurotoxic organophosphorus compounds.

Conclusions
Our test for the reactivation of paraoxon-inhibited AChE with 1methylpyridine-2-carboxaldehyde hydrazone (2) indicates that in general hydrazones may be ineffective. This result suggests that for the reactivation of phosphorylated AChE, it is necessary to use better nucleophiles, especially compounds that could be easily deprotonated by the diverse basic groups of the amino acids in this enzyme.
The analysis of guanylhydrazones by the Ellman's test indicates that they are also very poor AChE reactivators. However, these agents display interesting AChE inhibition capacity, being more effective when they possess electronegative groups on their benzylidene ring. The polarization of guanylhydrazones increases in the presence of electronegative groups, with the most effective being nitro groups, an effect that was confirmed by the Ellman's test, NMR and molecular modeling. This condition indicates that guanylhydrazones with high polarization interact better with the active site of AChE. With these results, more effective AChE inhibitors are being designed as new potential agents for the treatment of AD, indicating that guanylhydrazones and analogs could be good effective agents for AD.