Synthesis and Biological Evaluation of Dantrolene‐Like Hydrazide and Hydrazone Analogues as Multitarget Agents for Neurodegenerative Diseases

Abstract Dantrolene, a drug used for the management of malignant hyperthermia, had been recently evaluated for prospective repurposing as multitarget agent for neurodegenerative syndromes, including Alzheimer's disease (AD). Herein, twenty‐one dantrolene‐like hydrazide and hydrazone analogues were synthesized with the aim of exploring structure‐activity relationships (SARs) for the inhibition of human monoamine oxidases (MAOs) and acetylcholinesterase (AChE), two well‐established target enzymes for anti‐AD drugs. With few exceptions, the newly synthesized compounds exhibited selectivity toward MAO B over either MAO A or AChE, with the secondary aldimine 9 and phenylhydrazone 20 attaining IC50 values of 0.68 and 0.81 μM, respectively. While no general SAR trend was observed with lipophilicity descriptors, a molecular simplification strategy allowed the main pharmacophore features to be identified, which are responsible for the inhibitory activity toward MAO B. Finally, further in vitro investigations revealed cell protection from oxidative insult and activation of carnitine/acylcarnitine carrier as concomitant biological activities responsible for neuroprotection by hits 9 and 20 and other promising compounds in the examined series.


Introduction
Dantrolene (DAN; Figure 1) is a drug specifically used in the management of malignant hyperthermia, a life-threatening pathology with fatal course. In a recent work, we disclosed new biological activities exerted by DAN, namely inhibition of monoamine oxidase (MAO) B human enzyme with K i value in the low micromolar range, acetylcholinesterase (AChE), and aggregation of beta amyloid-40 and hexapeptide tau protein sequence PHF6, i. e. two probes of amyloid aggregation in Alzheimer's disease (AD) brain. [1] It is well known the crucial role of MAO isoforms A and B as metabolizing enzymes in modulating the concentration of neurotransmitters, mostly in some severe and chronic neurodegenerative pathologies. This established reputation is strictly related to the substrate and tissue specificity of both isoforms: MAO A selective inhibitors are clinically administered as antidepressants, [2] while MAO B selective inhibition is commonly used for the treatment of the early symptoms of Parkinson's disease. [3] A new outcome of that study was the discovery of the activation by DAN of the carnitine/acylcarnitine carrier (CAC), with EC 50 of 9.3 μM for the purified recombinant wild type (WT) protein. This transporter acts through reductive activation and is involved in trafficking of acyl groups into the mitochondria, carried by l-carnitine. Treated with DAN, this transport system facilitates, under oxidative stress (OS) conditions, the restoring of ATP production and thus cell vitality, but also the export of endogenous acetyll-carnitine from mitochondria, with consequent neuroprotective effects.
The above promising results prompted us to synthesize a number of novel DAN analogues, with the aim of optimizing the inhibitory activity against MAO B and AChE, ultimately improving their pleiotropic pharmacological potential in the treatment of AD and related neurodegenerative syndromes. In this work we investigated in particular: i) the bioisosteric replacement of the NO 2 group, which may be toxicophore and precursor for the production of reactive oxygen species (ROS), with CN group; ii) the improvement of the aqueous solubility of DAN by replacing the hydantoin moiety with other (hetero) cyclic moieties bearing protonatable nitrogen(s); iii) the effect of molecular simplification of the three-ring scaffold in DAN, for detecting the minimal pharmacophoric features responsible for the dual activity on MAO A/B and CAC ( Figure 1). SARs were investigated as thoroughly as possible, even considering the effect of lipophilicity on the enzymes' inhibition potency.
Some of the prepared compounds retained the molecular motif of hydantoin, present in DAN and, as thiohydantoin, in the known hypoglycemic drugs rosiglitazone and pioglitazone ( Figure 2), which also behave as moderate MAO inhibitors. In turn, the molecular pruning gave simple hydrazone and hydrazide derivatives (2-rings series, compounds 14-21) whose structural pattern could be related to that of isocarboxazide ( Figure 2), an early irreversible MAO inhibitor.

Chemistry
The synthetic pathways chosen for the preparation of the designed compounds are those explored by Snyder and coworkers with slight modifications. [4] The chemical scaffolds were selected to investigate a range of molecular diversity around the structure of DAN, in terms of variation of stereoelectronic and hydrophobic properties. The general strategy of structural modifications is shown in Figure 1, whereas the syntheses of compounds are shown in Schemes 1-3. Condensations from suitable aldehydes and amino/hydrazide derivatives were performed in DMF/water or acetone/water mixtures, with acidic catalysis and agitation at room temperature. Final compounds were obtained in moderate (20-50 %) to good (70 %) yields, following a simple workup including filtration and purification through either crystallization or column chromatography.

Inhibition of MAOs and AChE
The results of in vitro inhibition tests on human MAOs and AChE of the newly synthesized DAN analogues are summarized in Tables 1 and 2, along with the inhibition data of pargyline and galantamine used as positive controls for MAO (B-selective) and AChE, respectively. Inhibition assays were also performed on human butyrylcholinesterase (BChE), where all compounds resulted inactive or poorly active (data not shown). IC 50 values were calculated for compounds displaying > 60 % inhibition in one-point (10 μM) concentration assay. Compared with previously obtained results of DAN, [1] its CN congener 7 showed a significant loss of inhibition potency toward all three enzymes (approximately a four-fold decrease for MAO B). On the contrary, the thiophene hydrazides 1 and 4 proved to be equipotent with DAN for MAO B inhibition, with a noteworthy increase of MAO A activity and loss of MAO B/A selectivity for the CN congener 4. The same modification in compounds 2 and 5 resulted in almost overlapping activities, even with a decrease of MAO B potency.
Compound 3, which along with 1 and 2 maintains the 4nitrophenyl substituent, was prepared with the aim of introducing a strong basic moiety likely able to improve the affinity to cholinesterases. Unfortunately, it resulted a poor AChE inhibitor, while retaining fair MAO inhibition, although unselective, in the low micromolar range. The same activity profile was featured by the hydrazide 6 and 8, where the structural simplification of the hydantoin ring of the close analogue 7 to acethydrazide (6) and phenylacethydrazide (8) allowed the restoration of MAO inhibitory activity.
The introduction of terminal moieties with larger structural diversity, as in case of compounds 9-13, produced contrasting results. Concerning MAO B inhibition, we obtained an interesting submicromolar value of IC 50 for phenol derivative 9, which turned out as the most potent MAO B inhibitor of this series, also displaying 5-fold selectivity over MAO A. Inhibition kinetics assessed for 9 a competitive mechanism, with K i equal to 0.50 � 0.06 μM ( Figure 3). Taking into account the possible hydrolytic degradation of imine (see stability studies below), inhibition kinetics were determined without preincubation with substrate, as usually done for MAO inhibition experiments.
The derivatization of furaldehyde bridge as 4-substituted arylhydrazones (compounds 10-12) resulted in a decrease of activity towards MAO isoforms. Nevertheless, sulfonamide derivative 12 scored an unexpectedly remarkable inhibitory activity towards AChE with an IC 50 value of 1.67 μM.
While compound 4 was the most active inhibitor of MAO A (IC 50 1.46 μM), the replacement of the central furan ring with a larger and sterically hindered ring such as methoxybenzene in 13, resulted in a strong decrease of this inhibitory activity, while retaining a similar low value of IC 50 for MAO B. This potency reversal represents an interesting example of isoform selectivity, but seems also contrasting with previous results, [5] assessing a preferential MAO A affinity for molecules bearing bulky bridging substituents.
A number of structurally simpler derivatives, based on tworing scaffolds, were also synthesized with the aim of gaining information on the minimal pharmacophore features [6] of the compounds under examination. The effects of such a molecular pruning on target interactions are shown in Table 2. Removal of nitrophenyl group (compare 14 vs. DAN and 7, 15 vs. 1 and 4, 16 vs. 3, 17 vs. 8, 18 vs. 12) resulted in a marked loss of activity towards MAOs, while only in few cases AChE inhibition was sparingly retained, particularly in the case of isonicotinic acid derivative 19, analogue of 2 and 5. It is worthy of note that sulfonamide 18 lost the appreciable AChE inhibitory activity of its analogue 12, likely because of the lack of favorable hydrophobic/aromatic interactions by the nitrophenyl moiety in the enzyme binding site. This hypothesis is strengthened by the result of compound 20, which can be considered as a superior homologue of 18. The increase of lipophilicity, along with the introduction of an electron-donating substituent, through the replacement of furan with the methoxybenzene ring determined not only the restoring of the activity towards hAChE but also, in line with expectations, the restoring of the affinity towards MAO targets. Particularly in MAO B inhibition (IC 50 = 0.81 μM), sulfonamide 20 resulted as a potent and B/A selective candidate within the entire molecular set herein presented.
To investigate a possible correlation between MAO inhibitory potency and lipophilicity, [7] we calculated log P with three different programs (ChemDraw 15.0; ALOG PS 2.1; ChemSketch 2017, see Table S1 in Supporting Information) and compared the calculated values with experimental lipophilicity indexes as assessed by reversed-phase (RP) HPLC in isocratic conditions for the majority of compounds. The log of capacity factors (log k') measured by RP-HPLC using a mixture of methanol/PBS (60 : 40, v/v) as the mobile phase are reported in Tables 1 and 2, along with Clog P values calculated with ChemDraw. These calculated lipophilicity descriptors correlated with the experimental log k' values (n = 13, r 2 = 0.901) better than the other log Ps calculated by ALOG PS and ChemSketch computational tools (r 2 equal 0.705 and 0.568, respectively). As shown by the scatter plots in Figure 4 and Figure S1 in Supporting Information, there is no evident correlation trend between MAO A/B inhibition data and lipophilicity calculated by the expert system implemented in ChemDraw for compounds achieving finite IC 50 values.
While the graphical analysis proved the absence of a general correlation trend, the experimental log k' values (Tables 1 and 2), albeit limited to just over half of the compounds studied, suggest a certain effect of lipophilicity on the inhibition potency. Indeed, within the physicochemical property space explored, the strongest MAO A/B inhibitors possess higher lipophilicity (i. e., with values comprised in the 0.0-1.0 range), whereas the most active AChE inhibitors scored log k' values close to or lower than 0.

Cell-based assay of neuroprotection
In order to get further information about the potential neuroprotective effects exerted by the most active compound of the molecular series (phenol derivative 9) toward hMAO B, we tested it in a 2',7'-dichlorofluorescein diacetate (DCF-DA)  fluorescence-based assay measuring the production of ROS induced by dopamine in cultured SH-SY5Y cells.
The immortalized neuroblastoma cells are a widely used model for neuroprotection assays, while dopamine at a concentration of 10 nM acts as an inducer of oxidative stress, [8,9] being metabolized by MAOs to form hydrogen peroxide. Once hydrolyzed and oxidized, DCF acts as the fluorescing probe of the oxidative stress (OS) state of cells. In the presence of an antioxidant, or even of a MAO inhibitor, ROS burden is lowered and DCF fluorescence decreased. Figure 5 shows that 9 was effective in reducing ROS oxidation of DCF. Its activity is superimposable to that of phenelzine, a nonselective MAO inhibitor used as a positive control in this test. [10] The same experiment performed on dantrolene in our previous work gave similar results. [1] Hydrolytic stability in buffered solution The hydrolytic stability of some representative imino/hydrazone/hydrazide derivatives was determined in buffered aqueous media. The stability studies were carried out on compounds 3, 4 and 9, the most active MAO B inhibitors of the 3-rings scaffold series, at a single concentration of 20 μM in 10 mM phosphate buffered saline (PBS) at physiological pH of 7.4, at 37°C and for 6 h incubation time. The degradation profiles in Figure 6A confirmed a substantial stability for 3 and 4, while the imine 9 resulted, at the end of the 6 h of incubation, in a degradation percentage of 76 %. This result is in line with the nature of the functional group of this molecule: in fact, 9 is the only Schiff base within the synthesized series and therefore by itself less stable than the other hydrazone analogues. The hydrolytic degradation of 9 in its starting reagents was confirmed by the comparison of the chromatograms of 9 and 2chloro-4-hydroxyaniline, obtained in the same chromatographic conditions ( Figure 6B).

Activation of carnitine/acylcarnitine carrier (CAC) transport
CAC (SLC25 A20) is essential for the transport into mitochondria of acyl moieties as acylcarnitines, where they are processed by β-oxidation pathway. The protein contains six cysteine residues, but two of them, namely C136 and C155, based on the redox state of the protein, are crucial for the regular function of the carrier. [11,12] In fact, the transporter is active when the two cysteines are in reduced form, while it is inhibited when a C136-C155 disulfide bridge is formed in conditions of OS. One or both cysteines represent also specific targets for various chemical and physiological thiol reducing agents, [13][14][15][16][17] allowing to modulate the transport activity of the carrier. We demonstrated that DAN led to a significant recovery of the CAC transport activity of the oxidized protein. [1] Herein, we also investigated whether some newly synthesized derivatives,  The tested molecules were able to improve the transport activity of CAC compared to the control and at least three of them (9, 14 and 19) showed EC 50 lower than that previously calculated for DAN (9.3 μM) after the same exposure duration (30 min), highlighting similar pharmacological effect. On the contrary, the efficacy of these compounds, i. e., the power of the molecules to achieve maximum effect, is about 5 times lower than DAN. [1] To demonstrate that the action of DAN analogues was exerted on the cysteine residues of the CAC, 1 or 50 mM dithioerythritol (DTE), a strong reducing agent, was added to the reconstitution mixture (see Experimental section) in order to mimic the protein at different states of oxidation. The bar plot in Figure 8 shows that the tested molecules enabled the protein to recover a significant transport activity, compared to the control, when the protein is more oxidized, i. e., in the presence of 1 mM DTE.

Conclusion
From our previous investigation of the multitarget activity exerted by DAN, [1] we evidenced the potential of repurposing of this orphan drug toward AD-related targets, but also its intrinsic limitations, particularly its low aqueous solubility, that may discourage further pharmacological evaluation. The DAN-like molecules herein reported were designed and synthesized with the aim of extending the knowledge of the SARs and improving the pharmaceutical potential of this class of compounds. In particular, our efforts were focused on modifying the structure of DAN for increasing the aqueous solubility and replacing the potentially toxicophore nitro group, as well as following an approach of molecular simplification. The in vitro screening proved that some of the newly investigated analogues improved the MAO inhibitory potency, mostly for the 3-ring series (compounds 1-13), although with low isoenzyme selectivity. The phenol derivative 9 emerged as an outstanding, reversible MAO B inhibitor (K i 0.50 μM) with fair B/A selectivity. The drop of MAO A activity of compound 13, compared to that of its close congener 4, is a matter of evidence that would deserve further investigations, considering the high B/A selectivity obtained with this homologation. Among the 2-ring series, only the sulfonamide hydrazine 20 resulted in a strong and selective MAO B inhibitor (IC 50 0.81 μM). As far as the AChE inhibition is concerned, appreciable inhibition values were obtained only sparsely, in line with results recently published for a related class of DAN analogues. [18] While only sulfonamide 12 resulted in good AChE inhibition, the best selectivity was achieved with compounds 14 and 15 resulting from the molecular simplification study. The limited molecular size of both 2-ring and 3-ring scaffold hampered a significant inhibition of BChE.  The new compounds confirmed a good stability in buffered conditions, with the obvious exception of imine 9, and a safe cellular activity in contrasting ROS cytotoxicity from oxidative degradation of dopamine. Finally, some tested compounds confirmed the activation of the mitochondrial carnitine trafficking mediated by CAC, thus giving further evidence to the multitarget profile of this class of molecules. The evaluation of possible activity on ryanodine receptors will be a major concern to be faced in the near future, in order to achieve a more complete activity profile of these DAN analogues.

Experimental Section
Chemistry Chemicals, solvents and reagents used for the syntheses were purchased from Sigma-Aldrich (Milan, Italy) or Alfa Aesar (Haverhill, Massachusetts, USA) and used without any further purification. The purity of all intermediates, checked by 1 H NMR and HPLC, was always higher than > 95 %. Column chromatography was performed using Merck silica gel 60 (0.063-0.200 mm, 70-230 mesh). All reactions were routinely checked by TLC using Merck Kieselgel 60 F254 aluminum plates and visualized by UV light. Nuclear magnetic resonance spectra were recorded on a Varian Mercury 300 instrument (at 300 MHz) or on Agilent Technologies 500 apparatus (at 500 MHz) at ambient temperature in the specified deuterated solvent. Chemical shifts (δ) are quoted in parts per million (ppm) and are referenced to the residual solvent peak. The coupling constants J are given in Hertz (Hz). The following abbreviations were used: s (singlet), d (doublet), dd (doublet of doublet), t (triplet), q (quadruplet), qn (quintuplet), m (multiplet), br s (broad signal); signals due to OH and NH protons were located by All compounds were synthesized following Snyder's procedures [4] with slight modifications. Compounds 2, [19] 14, [20] 15 [21] and 19 [22] have already been described; their analytical data agreed with those reported in quoted references.

Synthesis of compounds 1-12
A suspension of 4-cyano or 4-nitroaniline (2 mmol) in 10 mL 6 N HCl was heated until the solid is solubilized, then the mixture cooled to 0°C. A solution of NaNO 2 (0.14 g, 2 mmol) in 1 mL of water was added and the mixture left under stirring for 30 min. In the order, solutions of 2-furaldehyde (0.19 g, 2 mmol) in 2 mL of acetone and CuCl 2 (0.04 g, 0.3 mmol) in 1 mL of water were added and the mixture was kept under stirring for 5 h. The precipitate formed was filtered and washed with distilled water. This intermediate compound was then solubilized in DMF (2 mL) and slowly added to an aqueous solution of the appropriate amine/ hydrazine/hydrazide derivative (1.8 mmol). A catalytic amount of 4 N HCl was added and the solution was left under stirring at room temperature for 24 h. The mixture was extracted with CHCl 3 (3 × 15 mL), the organic phase abundantly washed with H 2 O to remove the DMF and dried with Na 2 SO 4 . The solvent was evaporated under reduced pressure to give the crude product that was purified by column chromatography with hexane/ethyl acetate: 6/4 or 7/3 v/v as the mobile phase. N'-((5-(4-nitrophenyl)furan-2-yl) N-(4-methylpiperazin-1-yl)-1-(5-(4-nitrophenyl)

Synthesis of compound 13
The procedure was the same as for compounds 1-12, using 3methoxybenzaldehyde (0.27 g, 2.0 mmol) instead of 2-furaldehyde, and 2-thiophenecarboxylic acid hydrazide (0.43 g, 3.0 mmol). After evaporation of organic extract, the crude product obtained was purified by crystallization from absolute ethanol.

Synthesis of compounds 14-21
2.0 mmol of aldehyde (2-furaldehyde for 14-18; 3-methoxybenzaldehyde for 19 and 20; 3,4,5-trimethoxybenzaldehyde for 21) were solubilized in 5 mL of acetone and slowly added to the aqueous solution of the appropriate amine/hydrazine/hydrazide derivative, with a catalytic amount of HCl 4 N. The mixture was stirred at room temperature for 24 h, then extracted with ethyl acetate (3 × 15 mL) and the organic layer dried over anhydrous Na 2 SO 4 . Compound 21 was filtered after precipitation from the acetone/water mixture. The solvent was evaporated under reduced pressure to give the crude product that was purified by crystallization from absolute ethanol.  N'-((furan-2-yl)

Chromatographic measures
Stability studies in buffered solution were performed following an already described procedure, [23] on a Phenomenex C18 column (150 × 4.6 mm i.d., 3 μm particle size; Phenomenex, Castel Maggiore, Italy) using a mobile phase consisting of a mixture of methanolwater (75 : 25 v/v, with aqueous formic acid 0.1 %). The used flow rate was 0.500 mL/min while the injection volume was 20 μL. Wavelength of UV-Vis detector was adjusted at 254 nm. The chemical stability was evaluated in a phosphate buffer solution pH 7.4 (10 mM HPO 4 2À /H 2 PO 4 À ; 100 mM NaCl) at 37°C. Five different concentrations from 0.5 μM to 20 μM were studied in a 2 h time range (data not shown). Each concentration (0.5, 1, 5, 10 and 20 μM) was tested in triplicate starting from 3 different stocks solution (1 mM) prepared separately. The time range was extended to 6 h only for the samples at 20 μM concentration (Figure 6a).
Log k' values (k' = (t r -t 0 )/t 0 ) and purity determinations were carried out using a Phenomenex Gemini C18 4.6 × 150 mm, with 3 μm size particles, built on a Waters double pump HPLC system in isocratic conditions. Injection volumes were 10 μL, flow rate was 0.5 mL/min, and detection was performed with UV (λ = 254 and 370 nm). Samples were prepared by dissolving 0.1 mg/mL of the solute in 10 % v/v DMSO and 90 % v/v methanol. Retention times (t r ) were measured at least from three separate injections, and dead time (t 0 ) was the retention time of deuterated methanol. The mobile phase was filtered through a Supelco Nylon-66 membrane 0.45 μm (Merck Life Science Srl, Milan, Italy) before use. For each reference compound, the average t r of three consecutive injections of 10 μL of sample was used to calculate the log k' values. The eluent consisted of five different mixtures of methanol and PBS buffer 10 mM at pH 7.4, with methanol/buffer ratios ranging from 80 : 20 to 60 : 40 v/v.

Enzyme inhibition
Inhibition assays of hChEs and hMAOs (all from Sigma Aldrich) were performed by using already published protocols. [24] All compounds were assayed at 10 μM concentration and, for those showing inhibition > 60 %, IC 50 values was calculated by testing seven concentrations in the range 30-0.01 μM. Briefly, the classical spectrophotometric Ellman's test (for ChEs) and the fluorimetric detection of 4-hydroxyquinoline (for MAOs) were adapted to a plate reader procedure with 96-well microtiter plates (Greiner Bio-ChemMedChem Full Papers doi.org/10.1002/cmdc.202100209