Hepatotoxicity and antioxidant activity of some new N,N′-disubstituted benzimidazole-2-thiones, radical scavenging mechanism and structure-activity relationship

2.1 Chemistry

The synthesis of the 1,3-disubstituted compounds was carried out as shown in Scheme 1.

Close

Scheme 1

Synthesis of the studied benzimidazole compounds. Reagents and conditions: (a) CS₂, KOH ethanol solution, refluxing; (b) methyl acrylate, DMF, refluxing; (c) hydrazine hydrate, ethanol solution, refluxing.

Export to PPT

The starting precursors 1–5 were synthesized by refluxing 4-substituted-1,2-diaminobenzenes, carbon disulfide, ethanol and potassium hydroxide according to the method described by us earlier (Mavrova et al., 2015). Compounds 1–5 may exist in thione and thiol form as depicted in Scheme 1 depending on the medium. In solid state, 1–5 were isolated as thiols (Supplementary Material).

The novel synthetic method for obtaining 5-substituted dimethyl 3,3′-(2-thioxo-1H-benzimidazole-1,3(2H)-diyl)dipropionates 6–10 utilizes the Michael addition of the starting compounds 1–5 to methyl acrylate. The reaction was carried out in DMF media, where 1–5 are predominantly in thione form due to the solvent polarity. Our studies showed that for the preparation of 1,3-disubstituted benzimidazole-2-thiones the process could be carried out at a molar ratio of the starting reactants 1:2 (5-substituted benzimidazol-2-thiones:methyl acrylate). The reaction optimization conditions could be viewed on the following Table 1.

Table 1 Reaction optimization of Michael addition of compounds 1–5 to methyl acrylate.

Entry	5-substituted benzimidazole-2-thione	Molar ratio to methyl acrylate	Temperature	Time (h)	Product (% yield) (%)
1.	1	1:11	100 °C	6	63
2.	1	1:11	Reflux	2	69
3.	1	1:6	Reflux	2	68
4.	1	1:2	Reflux	2	65
5.	1	1:2	100 °C	2	57
6.	2	1:11	Reflux	2	70
7.	2	1:2	Reflux	2	67
8.	3	1:11	Reflux	2	64
9.	3	1:2	Reflux	2	62
10.	4	1:11	Reflux	3	65
11.	4	1:2	Reflux	3	60
12.	5	1:11	Reflux	5	63
13.	5	1:2	Reflux	5	62

IR spectra, ¹H NMR and ¹³C NMR confirmed that the benzimidazoles 1–5 reacted in their thione form and not as the thiol subsequently leading to the formation of 1,3-disubstituted derivatives 6–10. In the ¹H NMR spectra, particularly meaningful are characteristic N—CH₂CH₂CO signals (triplets or double triplets) of the 1,3-substituted compounds shifted downfield due to the deshielding effect of both N-atoms and the CO groups. The chemical shift values for N—CH₂ for the esters varied in the range from 4.60 to 4.45 ppm while those for the CH₂—CO— groups were from 3.00 to 2.82 ppm depending on the solvent and the substituent in the benzimidazole ring.

The condensation in absolute ethanol medium of 6–10 with hydrazine hydrate in molar ratio 1:4 afforded hydrazides 11–15. The chemical structures of the compounds and their purity were confirmed by IR-, ¹H NMR and ¹³C NMR, DEPT, COSY and HSQC spectra and the results are presented in Section 4 and in the Supporting material. In the hydrazides group, analogously to the esters, the most typical are also the signals for the N—CH₂CH₂—CO-groups, as for the N—CH₂ groups the chemical shifts are in the range of 4.48 to 4.43 ppm and those for the CH₂—CO-group from 2.62 to 2.50 ppm. The labile NH protons are not characteristic and their chemical shifts depend on the water quantity in the solvent.

2.2 Pharmacology

2.2.1

2.2.1 Hepatotoxicity

Convenient well-controlled biological model systems with high drug-metabolizing capacities, which can be used in experimental toxicology, are isolated rat hepatocytes. This in vitro system is part of the recommended tests from the European Centre for the Validation of Alternative Methods (ECVAM). The main goal of ECVAM was to promote the acceptance of alternative methods, which are important for reducing, refining and replacing the use of laboratory animals (Blaauboer et al., 1994). The studied benzimidazole-2-thiones were tested on isolated hepatocytes from male Wistar rats by a method providing a higher amount of live and metabolically active cells.

The effect of the compounds on the functional-metabolic status of the cells was assessed by monitoring the cell viability and changes in lactate dehydrogenase, glutathione and malondialdehyde levels. Lactate dehydrogenase (LDH) is a soluble cytoplasmic enzyme released into extracellular space when the plasma membrane is damaged. Therefore, in the present study the intracellular LDH leakage was used as an indicator of cell membrane integrity and cell viability of hepatocytes. It is well known that glutathione (GSH) is not only a reductant and one of the important antioxidants in the cells but also plays a mediator role in many physiologic processes, such as proliferation and apoptosis (Montserrat et al., 2009). Depletion of hepatic GSH was monitored as an indicator of mitochondria-associated apoptosis and oxidative stress. In light of the crucial role of oxidative stress in liver diseases, the oxidative stress in the isolated hepatocytes was monitored also by the production of malondialdehyde (MDA), which is the most abundant individual aldehyde resulting from the lipid peroxidation in biological systems. Increased MDA levels reveal elevated lipid peroxidation caused by overproduction of ROS. All the observed effects were compared to a control group of non-treated hepatocytes.

The studied compounds affected the functional-metabolic status of isolated rat hepatocytes to a different extent (Table 2).

Table 2 Effects of the tested compounds in concentration of 250 μM on the parameters characterizing the functional-metabolic status of isolated rat hepatocytes.

Group	Cell viability (%)	LDH (μmol/min/106 cell)	GSH (nmol/106 cell)	MDA (nmol/106 cell)
Control	85 ± 1.0	0.117 ± 0.004	19 ± 1.5	0.049 ± 0.01
6	49 ± 3.6**	0.366 ± 0.03**	11 ± 1.2**	0.125 ± 0.003*
7	53 ± 3.1**	0.338 ± 0.02**	12 ± 1.5**	0.121 ± 0.01*
8	47 ± 2.6**	0.452 ± 0.01**	11 ± 3.1**	0.128 ± 0.003*
9	45 ± 1.0**	0.507 ± 0.01***	9 ± 1.5***	0.220 ± 0.02***
10	63 ± 3.1*	0.225 ± 0.01*	16 ± 2.1	0.110 ± 0.01*
11	50 ± 6.0**	0.260 ± 0.04*	12 ± 1.8**	0.111 ± 0.01*
12	66 ± 4.5*	0.222 ± 0.01*	12 ± 1.5*	0.107 ± 0.01*
13	48 ± 4.5**	0.434 ± 0.01**	11 ± 1.2**	0.148 ± 0.03**
14	45 ± 1.5**	0.524 ± 0.01***	8 ± 2.1***	0.239 ± 0.01***
15	67 ± 2.6*	0.205 ± 0.01*	16 ± 2.5	0.104 ± 0.01*

* P < 0.05.

** P < 0.01.

*** P < 0.001 vs control (non-treated hepatocytes).

As can be seen ester 10 (1,3-bis(methoxycarbonylethyl)-5-benzoyl-1,3-dihydro-2H-benzimidazol-2-thione) decreased cell viability only by 27% compared to non-treated hepatocytes, while 8 and 9 had much higher toxicity and induced statistically significant decrease in cell viability by 47% and 45%, respectively. The reduction in cell viability was related to membrane damage as evidenced by the detected LDH leakage. The cell membrane integrity was best preserved by incubation with compound 10 (a statistically significant increase in LDH leakage by 92%), which had no effect on the GSH levels and elevated the MDA production to the least extent (124%) among the esters. Incubation with chloro-derivative 9 induced the greatest LDH leakage among the ester derivatives – 334%, and the highest oxidative stress according to the MDA production (349% increase) and GSH depletion (53% decrease).

Among the hydrazides, compound 15 had the lowest hepatotoxic effects, while hydrazide 14 revealed the highest cytotoxicity (Table 1). Derivative 15 induced a statistically significant decrease in cell viability by only 21% and showed no toxic effect on the GSH level. When considering the effects of compound 15 it should be pointed out that it exhibited the lowest levels of LDH leakage and MDA production – 75% and 112% respectively, not only among the hydrazides, but within the group of all tested benzimidazole-2-thiones.

The reduced cell viability (47%) resulting from incubation with the chloro-substituted hydrazide 14 was accompanied by the highest detected LDH leakage (348%), GSH depletion (58%) and MDA production (388%). Obviously, this compound induced the highest oxidative stress in the isolated rat hepatocytes resulting in membrane damage and cell death.

The differences in the toxicity are due to the presence of different substituents in 5-th position of the benzimidazole ring. In the ester group the toxicity increased in direction 10 (benzoyl group) < 7 (methyl group) < 6 (unsubstituted) < 8 (nitro group) < 9 (chloro group). The same relationship was observed in the hydrazides group.

The ester 10 and the hydrazide 15, which revealed the lowest cytotoxic effects on isolated rat hepatocytes, were examined for possible antioxidant activity in tert-butyl hydroperoxide-induced oxidative stress. Effects were compared to those of Quercetin – a well-known antioxidant (Fig. 1).

Close

Figure 1

Effects of 10, 15 and Quercetin (250 μM) on tert-butyl hydroperoxide (75 μM)-induced oxidative stress on parameters, characterizing the functional-metabolic status of isolated rat hepatocytes. ^***P < 0.001 – vs control; ⁺⁺P < 0.01, ⁺⁺⁺P < 0.001 – vs tert-butyl hydroperoxide ^***P < 0.001 – vs control group (non-treated hepatocytes); ⁺⁺P < 0.01, ⁺⁺⁺P < 0.001 – vs tert-butyl hydroperoxide.

Export to PPT

Administered alone, in concentration 75 μM, tert-butyl hydroperoxide showed statistically significant cytotoxic effects on isolated rat hepatocytes – decreased cell viability and GSH level by 78% and 79% respectively, and increased LDH leakage and MDA production by 458% and 476% respectively, compared to non-treated hepatocytes. The antioxidant activity of compounds 10 and 15 was estimated based on the comparison of the above-mentioned parameters with the hepatocytes treated with the tert-butyl hydroperoxide toxic agent in combination with 10 and 15. The ester derivative 10 preserved cell viability and GSH level with 189% and 225% respectively, while 15 preserved these parameters with 205% and 250%. Concerning LDH leakage and MDA level, 10 decreased them by 58% and 41% respectively, while 15 – by 60% and 43%. Quercetin in this model preserved cell viability and GSH level with 242% and 300% respectively, and decreased LDH leakage and MDA production by 61% and 48% respectively.

In tert-butyl hydroperoxide (t-BuOOH) induced oxidative stress, ester 10 and hydrazide 15 revealed statistically significant cytoprotective and antioxidant effects similar to those of quercetin (Fig. 1).

The metabolism of tert-BuOOH to free radicals passes through several steps. In microsomal suspension (in the absence of NADPH), tert-BuOOH undergoes one-electron oxidation to a peroxyl radical (reaction (1)), whereas in the presence of NADPH – one-electron reduction to an alkoxyl radical (reaction (2)). The toxic agent undergoes β-scission to methyl radical (reaction (3)) in isolated mitochondria and intact cells. All these radicals cause lipid peroxidation (Ollinger and Brunk, 1995; O’Donell and Burkit, 1994).

(reaction1)

$${({\text{CH}}_3)}_3\text{CCOOH}\to {({\text{CH}}_3)}_3{\text{CCOO}}^{\text{•}}+{\text{e}}^{-}+{\text{H}}^{+}$$

(reaction2)

$${({\text{CH}}_3)}_3\text{CCOOH}+{\text{e}}^{-}\to {({\text{CH}}_3)}_3{\text{CCO}}^{\text{•}}+{\text{OH}}^{-}$$

(reaction3)

$${({\text{CH}}_3)}_3{\text{CCO}}^{\text{•}}\to {({\text{CH}}_3)}_2\text{CCO}+{}^{\text{•}}{\text{CH}}_3$$

The cytoprotective effects of compounds 10 and 15 on t-BuOOH-induced oxidative stress can be attributed to their activity as scavengers of ROS and influence on the metabolism of tert-butyl hydroperoxide. As a result, both compounds were able to reduce the lipid peroxidation, upregulate the GSH levels and protect the cell membrane integrity in the rat hepatocytes.

2.3 Single-Crystal X-ray diffraction analysis

For better understanding of the structure and its influence on the biological properties, structural characterization of the studied compounds was performed by X-ray diffraction analysis and DFT methods. Two of the ester derivatives, 6 and 10, gave crystals suitable for X-ray diffraction analysis and enabled crystal structure determination. Compound 6 crystallized in the triclinic space group P-1, (No. 2) while compound 10 crystallized in the monoclinic space group P2₁/n (No. 14), both with one molecule per asymmetric unit (Figs. 2a and 2b, Table 5 in Section 4).

Close

Figure 2a

ORTEP drawing of (a) 6 and (b) 10 (atomic displacement ellipsoids are at 50% probability), the hydrogen atoms are shown as small spheres of arbitrary radii.

Export to PPT

Close

Figure 2b

Overlay of the molecules of 6 (in blue/dark) and 10.

Export to PPT

The majority of the bond lengths and angles of 6 and 10 are comparable to those of similar structures (Jishkariani et al., 2013; Mohamed et al., 2009; Hahn et al., 2005; Mairesse et al., 1984; Ito et al., 1987; Kutzke et al., 1996) found in the CSD (Table 3).

The benzimidazole-2-thione and benzoyl rings are nearly planar with rmsd of 0.007, 0.006 and 0.004 Å in 6 and 10 respectively. The steric repulsion of the ortho H atoms of the benzimidazole-2-thione and benzoyl moieties in 10 leads to a twist of the rings as typically observed in benzoyl substituted aromatic compounds (Cox et al., 2008). The angle of 50.88(7)° between the mean planes of the two rings corresponds to that in 4,4′-bis(diethylamino)benzophenone, 3,4-dihydroxybenzophenone and 3-hydroxybenzophenone. The benzoyl oxygen (016) is either below or above the mean planes of the two ring systems −0.676(4) and 0.539(3) Å.

Interestingly, the N-alkyl chains are in trans position (facing the opposite sides of the plane formed by the benzimidazole fragment) in 6, while they are cis (facing the same side of the plane) in 10. While no typical hydrogen bonding interactions could be detected in the crystal structures of 6 and 10, a multitude of weak interactions can be identified (Table 2). Among those, two intramolecular C—H⋯S and C—H⋯O (or π⋯O) interactions (Fig. 3) certainly play an essential role for the stabilization of the molecular geometry and the resulting conformation.

Close

Figure 3

Observed weak interactions in 6 and 10 stabilizing the molecular geometry, S1⋯H9A: 2.784 Å (6) and 2.876 Å (10); O14⋯H6: 2.738 Å (6) and 2.601 Å (10).

Export to PPT

From the intermolecular weak interactions contributing to the crystal structure stabilization of 6 one should note the C13H13B⋯O10 producing chain propagation along c. The chains are linked by C5H5⋯O11 to form pseudo-layers parallel to bc (Fig. S1a). The pseudo-layers are stacked and connected by CH⋯S and CH⋯O and thus the crystal structure is stabilized by a complex network of weak interactions (Fig. S1b). Similar to 6, in 10 the crystal structure stabilization is also achieved through a complex network of weak interactions (Fig. S2). However, due to the benzoyl fragment some additional CH3⋯π and CH⋯O interactions are observed.

2.4 Molecular geometry

As evidenced by the X-ray analysis, the 1,3-disubstituted benzimidazole-2-thiones have flexible N-alkyl chains able to occupy different positions. In order to find the preferred geometry of 1,3-disubstituted benzimidazole-2-thiones in isolated state, a large number of probable conformations were constructed and optimized at B3LYP/6-311++G^∗∗ level of theory. For every structure, the stationary points found on the molecular potential energy hypersurfaces were characterized using standard analytical harmonic vibrational analysis. The absence of imaginary frequencies, as well as of negative eigenvalues of the second-derivative matrix, confirmed that the stationary points correspond to minima on the potential energy hypersurface. The benzimidazole-2-thione fragment is characterized by a planar structure, while the alkyl chains attached to N1 and N3 produce several extended and folded conformations and mutual orientations. It was found that bent conformation of the alkyl chain with the carbonyl ester group pointing to the benzimidazole moiety is the most favorable disposition of the N-substituents in isolated state. For the simplest ester and hydrazide derivatives (6 and 11) with an unsubstituted benzimidazole fragment, two stereoisomers are possible according to the mutual orientations of the N-alkyl chains (Scheme 2).

Close

Scheme 2

Possible stereoisomers of 6.

Export to PPT

The stereoisomer with N-substituents in trans conformation is more favorable than that with N-alkyl chains in cis conformation. The conformational interconversion is related to a strong change in polarity, from 1.62 D for the trans form to 3.21 D for the cis form of compound 6. The predicted isolated-state trans conformation of 6 is very similar to the one found in crystal state (Fig. S3a), with rmsd value of 0.2427 Å.

In all other studied derivatives the presence of a substituent at 5-position of the benzimidazole fragment creates two distinct faces of the plane of symmetry, which results in four possible stereoisomers (Scheme S1). In the case of benzoyl substituted derivatives (ester 10 and hydrazide 15), the number of possible stereoisomers is increased by rotation of the benzoyl group. For each of the above mentioned stereoisomers there are four possible conformations of the benzoyl group (Scheme S2). The energy differences (0.05–0.40 kJ·mol⁻¹) between the stereoisomers resulting from the different orientation of the N-alkyl chains indicated that the side chains of the studied benzimidazole-2-thione derivatives are extremely flexible and the conversion of one form into another has very low energy requirements. Rotation of the benzoyl group also gave rise to stereoisomers with small energy differences (0.05–2.36 kJ·mol⁻¹). Comparing the total energies of the possible conformers of 10 and 15, it was established that trans conformation is the preferred one for both molecules in isolated state (Scheme 3).

Close

Scheme 3

Optimized geometry of the most stable stereoisomer of 10 (a) and 15 (b).

Export to PPT

The angles between the mean planes of the benzimidazole-2-thione and benzoyl moieties in the optimized structures are 50° (10) and 54.8° (15), respectively.

The calculation results point out that the isolated-state structure of 10 differs from the crystal-state one (Fig. S3b). As is well-known, the crystal packing is affected by the formation of classical and non-classical hydrogen bonding, ring–ring (π⋯π) interactions, X—H⋯π interactions and other non-bonded intermolecular contacts influence. Thus considering the insignificant energy differences between the cis and trans conformers, the low energy requirements for the rotation of the benzoyl group and the additional factors affecting the crystal structure, the presence of different conformations in isolated state and in the crystal is not surprising. The crystal structure of 10 and the theoretically calculated geometry of the corresponding conformer are in an excellent agreement, with rmsd value of 0.3731 Å (Fig. S3c).

2.5 SAR

Based on the small energy differences between the different conformations, it was assumed that 10 might easily convert into its lowest-energy form in biologically relevant liquid medium. Therefore, the computational study on the possible mechanism of antioxidant action was carried out by using the global energy minimum conformations of 10 and 15. UB3LYP/6-311++G^∗∗ level of theory was applied in order to enable comparison with calculated reaction enthalpies of other well-known and structurally related radical scavengers.

A good antioxidant should be able to easily react with a wide variety of free radicals. This includes the HO^•, which is the most reactive among the ROS with little selectivity toward the various possible sites of attack (Samuni et al., 1983); the less reactive HOO^• capable of diffusing into remote cellular locations (Marnett, 1987) and responsible for the initiation of lipid peroxidation (Aikens and Dix, 1991); the alkoxyl radicals LO^• which are formed from the reduction of peroxides and are less reactive than HO^•, but significantly more reactive than the peroxyl radicals (León-Carmona and Galano, 2011); and the least reactive organic peroxyl radicals i.e. LOO^• (Huie and Neta, 2002).

The studied benzimidazole-2-thiones partially resemble the structure of melatonin and its N-substituted derivatives (Phiphatwatcharaded et al., 2014) due to the presence of a flat conjugated aromatic heterocycle substituted at 5-position, as in melatonin, and side chains of approximately the same length containing carbonyl ester or amide function (Scheme 4).

Close

Scheme 4

Chemical structure of benzimidazole-2-thiones, melatonin and its N-substituted derivatives.

Export to PPT

Melatonin is known as a versatile antioxidant able to scavenge various free radicals (ROS and RNS) (Shirinzadeh et al., 2010). Its N-substituted derivatives exhibit in vitro antioxidant and anti-inflammatory activities. Hydrogen atom transfer (HAT) and single electron transfer (SET) were identified, from experimental data, as the main mechanisms for the reactions of melatonin, its N-substituted derivatives (Johns and Platts, 2014) and other indole derivatives (tryptophan and N-methylindole) with different free radicals (Solar et al., 1991): $$\text{R—H}+{\text{LOO}}^{\text{•}}\to {\text{R}}^{\text{•}}+\text{LOOH}(\text{HAT})$$ $$\text{R—H}+{\text{LOO}}^{\text{•}}\to {\text{RH}}^{+\text{•}}+{\text{LOO}}^{-}(\text{SET})/{\text{RH}}^{+\text{•}}+{\text{LOO}}^{-}\to {\text{R}}^{\text{•}}+\text{LOOH}(\text{PT})$$

The SET mechanism for direct radical scavenging of melatonin is the most favorable mechanism in aqueous solution, while in nonpolar aprotic medium HAT is prevailing (Galano, 2011). Therefore, based on the structure similarities, these mechanisms are also expected to contribute to the overall free radical–scavenging activity of the benzimidazole-2-thiones studied by us.

The first mechanism to be considered was HAT. Lipid peroxidation inhibition is known to proceed by this mechanism and involves all aforementioned oxygen-based radicals (Klein et al., 2007; Frankel, 1998; Burton et al., 1985; ​De Heer et al., 2000). For effective lipid peroxidation inhibition, the radical formed by the antioxidant agent should be less reactive and more stable than the attacked LOO^• i.e. the antioxidant should have lower bond dissociation enthalpy (BDE) than the free lipid radicals. The lower the BDE value, the higher the radical scavenging capacity of the antioxidant agent. For the accurate estimation of the antioxidant capacity via HAT by computational methods it is necessary to compare it to a lipid BDE value, representative for the chosen level of theory. In our study the lipid peroxyl and alkoxyl radicals were illustrated by the radicals of (Z)-4-hydroperoxyhex-2-ene, (Z)-4-hydroxyhex-2-ene, methyl hydroperoxide and methanol in gas-phase (Scheme 5).

Close

Scheme 5

BDE values of model lipid radicals calculated at B3LYP/6-311++G^** level of theory.

Export to PPT

Therefore, in order to scavenge the less reactive lipid peroxyl radicals, the antioxidant agents are expected to show BDE values below 330 kJ·mol⁻¹, while for scavenging of lipid alkoxyl radicals BDEs below 400 kJ·mol⁻¹ are sufficient. The BDE values of CH₃COO^• and CH₃O^• show only small deviation from the two larger radicals, which is an indication that the reactivity of the lipid radicals is sufficiently well described even by the small methoxyl and methyl peroxyl radicals. Thus CH₃COO^• and CH₃O^• can be used in the following study of radical scavenging as reliable representation of the lipid radicals.

α-Tocopherol, which is an effective chain-breaking antioxidant, has BDE of 327 kJ/mol according to the calculations in gas-phase at B3LYP/6-311++G^∗∗ level of theory (Klein et al., 2007). The O—H BDE values of quercetin range from 305 to 398 kJ·mol⁻¹ (at the same computational scheme) as has been reported (Vagánek et al., 2012). Melatonin is able to donate a hydrogen atom from the methylene group next to the indolyl fragment with C—H BDE value of 319 kJ·mol⁻¹ and from the indolyl N—H with BDE value of 322 kJ·mol⁻¹ (calculated at B3LYP/6-31G(d,p) level of theory) (Velkov et al., 2009).

For the ester 10, hydrogen atom abstraction would be possible from the alkyl groups in the side chains - the C—H bonds next to the benzimidazole N-atoms (sites 1, 1′, Fig. 4) or the C—H bonds next to the ester carbonyl groups (sites 2, 2′, Fig. 4). After full geometry optimization of the resulting radical species, the BDEs related to each of these sites were calculated and compared. As could be seen from the values depicted in Fig. 4 the lipid radicals will attack preferentially the hydrogen atom of the C—H bond next to the carbonyl group. The corresponding BDE values are approximately 15 kJ·mol⁻¹ lower than the BDE related to sites 2 and 2′. Obviously the proximity of the carbonyl groups and the lengthening of the C—H bonds due to the intramolecular C—H⋯S interactions considerably facilitate their cleavage. In comparison with the previously investigated 1,3-disubstituted-benzimidazol-2-imines (Mavrova et al., 2015) the methylene groups next to the N-atoms from the benzimidazole-2-thione ring are significantly less prone to donate a hydrogen atom.

Close

Figure 4

Hydrogen atom abstraction at different sites and corresponding B3LYP/6-311++G^** bond dissociation enthalpies of 10 in kJ·mol⁻¹.

Export to PPT

Taking into account the estimated lipid BDE values, it can be concluded that 10 would be an efficient radical scavenger of lipid alkoxyl radicals but nor peroxyl radicals. Therefore, the protective effect of the ester benzimidazole-2-thiones against lipid peroxidation should be exerted by scavenging the highly reactive HO^•, which initiates the degradation process, and the alkoxyl radicals LO^• formed from the reduction of lipid peroxides.

For the hydrazide derivatives, there are two more sites for hydrogen atom abstraction from each N-alkyl chain – from the amide N—H bond (site 3) and the amino N—H bonds (site 4) (Fig. 5). The BDE values for site 1 and 2 in 15 are comparable to those in 10. However, the N—H bonds in 15 show remarkably lower BDE values giving strong advantage to 15 as a radical scavenger as compared to 10, where only C—H bonds could be cleaved. The most favorable site for the hydrogen atom abstraction is the amide N—H bond (site 3) with BDE of 308 kJ·mol⁻¹, close to the calculated BDE value for 4′—OH group of quercetin (Vagánek et al., 2012). The second preferred one is site 4 with BDE of 345 kJ·mol⁻¹.

Close

Figure 5

Hydrogen atom abstraction at different sites and corresponding B3LYP/6-311++G^** bond dissociation enthalpies of 15 in kJ·mol⁻¹

Export to PPT

As the BDE value for site 3 is lower than that of the lipid peroxyl radicals, 15 would be able to trap lipid peroxyl, as well as alkoxyl radicals, and inhibit directly the lipid peroxidation process. This makes the hydrazide benzimidazole-2-thiones more efficient lipid peroxidation inhibitors than the ester derivatives via HAT mechanism.

In a polar aqueous medium, ionization mechanisms of antioxidant action, such as SET, are more likely to occur. SET is characterized by the ionization potential (IP) of the studied molecules and a lower IP implies an easier electron release. In order to estimate the probability 10 and 15 to react via the SET mechanism and compare their reactivity via this route, we optimized the geometry of their radical cations at B3LYP/6-311++G^∗∗ level of theory and calculated the corresponding IP values.

In gas state the IP value of 10 is found to be 691 kJ·mol⁻¹ i.e. comparable to the value calculated for quercetin (698 kJ·mol⁻¹) by Vagánek et al. according to the same computational scheme (Vagánek et al., 2012). The hydrazide 15 showed IP of 682 kJ·mol⁻¹. In water, the solvation of the electron and the positively charged radical species lowers the IP values. For this reason, the IP and BDE values of 10 and 15 were calculated at IEF-PCM B3LYP/6-311++G^∗∗ level of theory and compared. The ester derivative 10 showed IP of 329 kJ·mol⁻¹ in water, dropping below its water BDE (site 1) value 374 kJ·mol⁻¹ and below the water IP of phenol - 346 kJ·mol⁻¹. Hydrazide 15 showed water IP of 328 kJ·mol⁻¹, while the water BDE value for hydrogen atom transfer from site 3 was 320 kJ·mol⁻¹. Subsequently the SET mechanism of radical scavenging becomes preferred in water medium for 10, and competitive to HAT for 15.

The ability of a system to donate one electron can be estimated also by calculating the relaxed ionization potential (rIP) according to the approach suggested (Gazquez et al., 2007). Conformation changes in the side chain of melatonin are reported to have little impact on rIP (Johns and Platts, 2014), as the removal of an electron to the neutral molecule affects only the flat conjugated indole moiety. Thus rIP of 6.85 eV was computed at B3LYP/6-31+G^∗ level of theory for melatonin in gas phase (Johns and Platts, 2014), which is in good accordance with the value estimated by photoelectron measurements (Cannington and Ham, 1983). The rIP values for 10 and 15 (computed in gas phase at B3LYP/6-311++G^∗∗ level of theory) are 7.14 and 7.04 eV, which suggest a good capacity for the benzimidazole-2-thiones to donate one electron and scavenge free radicals via the SET mechanism.

Therefore, the studied benzimidazole-2-thiones are expected to easily transfer an electron to the lipid radicals and form radical cations. After the transfer of the free electron, the melatonin can undergo molecular rearrangements and produce several metabolites, which in turn act as radical scavengers (Johns and Platts, 2014). In the present case several routes of radical scavenging also exist (Scheme 6).

Close

Scheme 6

Hypothetical mechanism of antioxidant action of benzimidazole-2-thiones via SET.

Export to PPT

The electron transfer might be immediately followed by proton transfer and complimented by the formation of a cyclic intermediate of the benzimidazole-2-thiones. The latter would be a neutral radical able to scavenge CH₃COO^• and CH₃O^• (adduct formation) and inhibit the lipid peroxidation. On the other hand, due to the substantial spin density localized over the S-atom in the radical cation, it is anticipated that another possible way would be to bind a lipid radical at this site and form a cation adduct.

A good antioxidant should be able to cross physiologic barriers and to be quickly transported into the cells. For this reason, the solubility of the studied compounds in lipids and water was assessed by calculating log P, molecular size, flexibility and the presence of hydrogen-donor and acceptors with a Molinspiration tool (Table 4) (Molinspiration Cheminformatics, 2015).

The m_i log P values of the benzimidazole-2-thione esters 6–10 and hydrazides 11–15 range between 0 and 2.7 outlining various possible biological sites where they can act as antioxidant agents. The lipophilicity of the esters 6–10 is slightly higher and they can be applied as lipid peroxidation protectors. As more hydrophilic compounds, the hydrazides 11–15 may protect against ROS damage in the aqueous medium along with the traditional protectors vitamin C, lipoic acid and aqueous antioxidant enzymes.

Overall the calculated TPSA, molecular weight, number of rotatable bonds and number of hydrogen-bond acceptors and donors indicate good solubility and permeability of the studied compounds. Having in mind their relatively low hepatotoxicity, it could be concluded that they possess favorable properties for application as radical scavengers and oxidative stress inhibitors.

4.1 Synthesis of compounds 1–5

The starting 4-substituted-1,2-diaminobenzenes were commercially available. The 5-substituted benzimidazole-2-thiols 1–5 were synthesized as previously reported (Mavrova et al., 2005a,b).

4.2 General procedure for preparation of compounds 6–10

To a solution of 5-substituted-2-mercaptobenzimidazole (0.01 mol) in DMF (20 ml) methyl acrylate (0.02 mol) was added and the mixture was boiled under reflux for 2–5 h. The reaction was monitored by TLC and after finishing the reaction the DMF was distilled off under vacuum and a small portion of methanol was added to the residue. It was left over night. The crude crystals were recrystallized from methanol.

4.3 General procedure for preparation of compounds 11–15

Hydrazine hydrate (0.04 mol) and 0.01 mol of the corresponding 1,3-disubstituted-2-thione benzimidazole esters 6–10 were refluxed in absolute ethanol for 2 h. After completion of reaction the mixture was cooled and the obtained crystals were filtered off and recrystallized with ethanol.

4.4 Hepatotoxicity assay

4.5 X-ray experimental

Single crystals (colorless blocks with approximate dimensions 0.25 × 0.22 × 0.2 mm³) of 6 and 10 were obtained by slow evaporation from methanol. A transparent and crack free crystal was carefully selected and mounted on a glass capillary. Diffraction data were collected at room temperature by ω-scan technique, on an Agilent Diffraction SuperNova Dual four-circle diffractometer equipped with Atlas CCD detector using mirror-monochromatized MoKα radiation from a micro-focus source (λ = 0.7107 Å). The determination of cell parameters, data integration, scaling and absorption correction was carried out using the CrysAlis Pro program package (CrysAlis PRO, 2011). The structures were solved by direct methods (SHELXS-2014) (Sheldrick, 2008) and refined by full-matrix least-square procedures on F² (SHELXL-2014). The heavy atoms (C, N, O and S) were positioned from difference Fourier maps while hydrogen atoms were placed at idealized positions. The non-hydrogen atoms were refined anisotropically while the hydrogen atoms were constrained to ride on their parent atom with U_iso(H) values of 1.2U_eq (C). A summary of the fundamental crystal and refinement data is provided in Table 5. Crystallographic data (excluding structure factors) for the structural analysis have been deposited with the Cambridge Crystallographic Data Centre, Nos. CCDC-1496636 (6) and CCDC-1496637 (10). Copies of this information may be obtained free of charge from The Director, CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK. Fax: +44 (1223)336-033, e-mail: deposit@ccdc.cam.ac.uk, or www.ccdc.cam.ac.uk.

4.6 Computational details

All theoretical calculations were performed using the Gaussian 09 package (Frisch et al., 2009) of programs. Geometry and vibrational frequencies of species studied were performed by an analytical gradient technique without any symmetry constraint. All the results were obtained using the density functional theory (DFT), employing the B3LYP (Becke’s three-parameter non-local exchange (Becke, 1993; Lee et al., 1988 correlation potentials). Natural bond orbitals (NBO) analysis (Reed et al., 1988; Carpenter and Weinhold, 1988; Weinhold and Carpenter, 1988) has been performed to characterize the delocalization of electron density within the molecule.

The equations used for calculation of dissociation enthalpy (BDE) and ionization potential (IP) of the studied compounds are given below: $$\text{BDE}=\text{H}({\text{R}}^{\text{•}})+\text{H}({\text{H}}^{\text{•}})-\text{H}(\text{R—H})$$ $$\text{IP}=\text{H}({\text{R}}^{\text{•}+})+\text{H}({\text{e}}^{-})-\text{H}(\text{R—H})$$

The enthalpy of the hydrogen atom, H(H) was obtained by the same method and basis set. All reaction enthalpies were calculated at 298 K. The enthalpies of proton H(H⁺), and electron, H(e⁻), were taken from the literature: 6.197 kJ/mol and 3.145 kJ/mol, respectively (Klein et al., 2007). Solvation enthalpies of proton H(H+), electron, H(e⁻), in water, determined using IEF-PCM DFT/B3LYP/6-311++G^∗∗ calculations, were used as reported (Rimarčík et al., 2010). The lipid radical was modeled for comparison according to the same computational scheme.

Relaxed IP was calculated as the difference between the energy of the optimized radical cation and the optimized neutral molecule. (Johns and Platts, 2014).