Studies on the Anticonvulsant Activity and Influence on GABA-ergic Neurotransmission of 1,2,4-Triazole-3-thione-Based Compounds

The anticonvulsant activity of several 1,2,4-triazole-3-thione derivatives on mouse maximal electroshock-induced seizures was tested in this study. Characteristic features of all active compounds were rapid onset of action and long lasting effect. Structure-activity observations showed that the probability of obtaining compounds exerting anticonvulsant activity was much higher when at least one of the phenyl rings attached to 1,2,4-triazole nucleus had a substituent at the para position. The obtained results, moreover, permit us to conclude that despite the structural similarity of loreclezole (second-generation anticonvulsant drug) and the titled compounds, their anticonvulsant activity is achieved via completely different molecular mechanisms.


Anticonvulsant Activity
Compounds (1a-30a) were evaluated in vivo using the mouse maximal electroshock-induced seizure model-an experimental model of human generalized tonic-clonic seizures. Procedures involving animals and their care were conducted in accordance with current European Community and Polish legislation on animal experimentation. Additionally, all efforts were made to minimize animal suffering and to use only the number of animals necessary to produce reliable scientific data.
In the initial stage of the study, groups of eight mice were supplied with the compounds (administered intraperitoneally) in a dose of 300 mg/kg and then exposed to electrical impulse at fixed intervals (15,30,60, 120 min). The anticonvulsant effects of the compounds were considered important if at least 50% of the animals tested were protected against electroshock-induced seizures (Table 1). In the cases of compounds 3a and 14a, administered in a dose of 300 mg/kg, it was impossible to establish their protective effect in the maximal electroshock-induced seizure (MES) test, as the dose resulted in the occurrence of acute adverse effects. Once the doses of both compounds were decreased to lower concentrations, which did not induce neurotoxic effects, only Compound 3a retained its anticonvulsant activity. Nine of the 30 tested compounds (1a-30a) exhibited various levels of anticonvulsant activity. All active derivatives (3a, 4a, 6a, 10a, 12a, 13a, 15a, 16a, 22a) were characterized by a quick onset of action. A significant level of protection against maximal-electroshock induced seizures was observed as early as 15 min after their parenteral administration. From a preclinical point of view, the long-lasting anticonvulsant activity is also noteworthy. Apparently, the highest ability to remain in the central nervous system was presented by 13a and 16a, whose time to peak of the anticonvulsant action was observed at 120 min after their systemic administration.
In the second stage of the pharmacological research median effective doses (ED 50 ), median toxic doses (TD 50 ), and protective index (PI) values for compounds 3a, 4a, 6a, 10a, 12a, 13a, 15a, 16a, and 22a were determined ( Table 2). The most potent activity (approximately five and a half times higher than that of valproate) was exhibited by 5-(3-chlorophenyl)-4-(4-fluorophenyl)-2,4-dihydro-3H-1,2,4triazole-3-thione (3a) whose ED 50 equaled 35.2 mg/kg. Time-course and dose-response relationship of action of 3a revealed that the time to peak of anticonvulsant action was observed at 15 min after its i.p. administration. The level of toxicity (TD 50 ) of the compound, ranging from 136.7 to 201.0 mg/kg makes long-term use of this derivative impossible. However, due to the fast onset of action, low ED 50 value, as well as the favorable protection index value (PI = 3.9), Compound 3a should undergo further investigation for its potential use in interrupting status epilepticus. Among the remaining derivatives, only compounds 6a and 22a exhibited activity slightly weaker than that of valproate. The former compound, despite its less potent anticonvulsant activity, significantly (i.e., by 70%) elevated the threshold for electroconvulsions and intensified the anticonvulsant action of carbamazepine, phenobarbital, and valproate [12]. A very beneficial activity profile, expressed by low variability in the ED 50 values at different pretreatment times (15,30, 60, 120 min), was observed in the cases of 2-fluorophenyl (12a) and 2,4-difluorophenyl (13a) derivatives. The ED 50 values for these derivatives were 186.4-217.8 mg/kg and 80-108.9 mg/kg, respectively. It is also noteworthy that 5-(4-chlorophenyl)-4-(2-fluorophenyl)-2,4-dihydro-3H-1,2,4-triazole-3-thione (12a) was the best tolerated compound among all tested 1,2,4-triazole derivatives. Its median toxic doses (TD 50 ) ranged from 418.5 to 534.3 mg/kg. In most cases, the toxicological profile of the discussed compounds was similar to that of valproate (PI = 1.7-2.1). A more beneficial value of the protection index (PI = 4.5-5.9) was observed for 5-(3-chlorophenyl)-4-(4-methylphenyl)-2,4-dihydro-3H-1,2,4-triazole-3-thione (10a). Additionally, our previous studies demonstrated pharmacologically beneficial interactions of 10a with valproate that were pharmacokinetic in nature [13]. Sub-protective doses (i.e., the doses that per se did not affect the seizure threshold) of compound 10a intensified the anticonvulsant action of valproate as the total brain concentration of this drug was elevated by 52%. Table 1. Time-course of anticonvulsant effects of compounds (1a-30a) against maximal electroshock-induced seizure (MES)-induced seizures in mice. Data is presented as the number of animals protected against maximal electroshock (MES)-induced seizures out of eight animals per group. The MES test was performed at various pretreatment times (15,30,60, 120 min) after systemic administration of the investigated compounds in a fixed dose of 300 mg/kg. p-number of animals protected against MES-induced seizures; t-number of animals tested. "*" -Substance administered at a dose of 300 mg/kg produced acute neurotoxic effects in mice and it was impossible to determine its anticonvulsant action in the mouse MES model.      Based on the data presented in Tables 1 and 2, it may be assumed that the anticonvulsant activity of 1,2,4-triazole derivatives depends on the structure of substituents at positions N2, N4 and C5 of the triazole ring. In the group of monosubstituted derivatives of 4-aryl-5-(3-chlorophenyl)-2,4-dihydro-3H-1,2,4-triazole-3-thione (1a-11a), the most advantageous position for the substituent seemed to be the para position. Simultaneously, the electronic nature of the substituents was much less important as regards the effect on the anticonvulsant potency (since derivatives with both electron-withdrawing and electron-donating substituents were active). Moreover, the ED 50 values for para substituted derivatives (3a, 4a, 6a, 7a) allow the observation of the negative correlation between the anticonvulsant activity and the size of halogen substituent. Following this correlation, the 4-fluorophenyl derivative (3a) exhibited the most potent activity, whereas the 4-iodophenyl derivative (7a) was inactive. This may suggest that a large iodine atom constitutes a steric hindrance, which prevents the 4-halogenophenyl fragment from binding to appropriate molecular targets. Comparison of the activity of 2-fluorophenyl (1a vs. 12a) and 2-bromophenyl (5a vs. 15a) derivatives is more evidence confirming that the probability of obtaining compounds exerting anticonvulsant activity is much higher when at least one of the phenyl rings attached to the 1,2,4-triazole nucleus has a substituent at para position. Exchanging 3-chlorophenyl moiety (1a, 5a) for a 4-chlorophenyl group (12a, 15a) resulted in the emergence of anticonvulsant activity in the cases of 2-substituted derivatives.
In order to study the effect of the presence of a substituent at position N2, four active 1,2,4-triazoles (6a, 10a, 15a, 16a) were transformed into their appropriate N2-aminomethyl derivatives (19a-30a). Among the compounds obtained this way, only compound 22a exhibited a weak anticonvulsant activity, much less potent than that of its predecessor (10a). The remaining N2-aminomethyl derivatives were inactive. According to Waterbeemd and co-workers, the essential factors affecting the drug delivery across the blood-brain barrier (BBB) are molecular weight (M W ) and polar surface area (PSA) [20]. Based on a study of marketed central nervous system (CNS)-and non-CNS drugs, it was established that the transport of a drug to the brain is possible when (T)PSA < 90Å, while M W < 450. Among the Mannich bases (19a-30a), only compound 22a met these requirements (Table 3). Therefore we speculate that the most probable cause of the loss of anticonvulsant activity, once the hydrogen atom is substituted with an aminomethyl moiety, is the loss of the ability of such compounds to penetrate across the blood brain barrier (BBB). Apart from the BBB, another important biological barrier the permeability of which (or lack thereof) affects drug efficiency, is the intestinal epithelium, which modulates drug absorption after oral administration. Most antiepileptic drugs are administered orally, with the exception of drugs interrupting status epilepticus. Therefore, newly synthesized compounds with antiepileptic potential should be efficiently absorbed from the digestive system. It has been demonstrated experimentally that intestinal absorption of drugs is significantly correlated with the (T)PSA parameter. Palm et al., have proven, based on Caco-2 cell studies, that the drugs with a (T)PSA below 60 are completely absorbed in the intestine [21]. In the cases of active compounds (3a, 4a, 6a, 10a, 12a, 13a, 15a, 16a, and 22a), the PSA values ranged from 33.62 to 35.23 Å 2 and the level of intestinal absorption according to the algorithm described by Zhao et al., [22] was in the range of 96.8%-97.4% (Table 3). Using these calculations as a theoretical model it is highly likely that 1,2,4-triazole-based compounds could be considered as promising drug candidates, although further experiments should confirm their applicability in clinical settings.

Radioligand Binding Assay
Epileptic seizures are caused by disruption of the balance between excitatory and inhibitory systems in the central nervous system. The most important neurotransmitter inhibiting neuronal processes is γ-aminobutyric acid (GABA). Among three types of GABA receptors (i.e., GABA A , GABA B , GABA C ), the main role in preventing seizures is played by the GABA A receptors [23,24]. Antiepileptic drugs acting directly on GABA A receptors, enhancing the synthesis or inhibiting the metabolism of GABA, as well as blocking the reuptake of GABA from the synaptic cleft, cause an increase in the GABA-ergic neurotransmission [25]. Our studies conducted to date [12,13,26,27], involving the design and synthesis of 1,2,4-triazole derivatives endowed with anticonvulsant activity, have led to promising results, which in turn has encouraged us to attempt the elucidation of their mechanism of action. A representative group of the synthesized 1,2,4-triazole-based compounds (including active and inactive ones) was tested with the aim to establish whether the anticonvulsant activity of these derivatives was a result of their interactions with GABA A receptor complex and/or their affinity to benzodiazepine (BDZ) binding sites. The degree of specific binding of the compounds to GABA A receptors and BDZ-binding sites was low and ranged from 2% to 24% (Table 4). This forced us to exclude the possibility of a direct involvement of GABA A receptors in the occurrence of the anticonvulsant effect. Similarly, the possibility of allosteric modulation of the function of this receptor via the interactions of the investigated 1,2,4-triazole derivatives with BDZ-binding sites should be excluded.
Activation of the GABA A receptor may be due to binding of various chemical substances to the allosteric sites on the receptor surface [28]. Barbiturates and benzodiazepines are the best-known allosteric modulators of GABA A receptors. However, a similar mechanism of action has also been described for loreclezole -a second-generation antiepileptic drug. Yet, the binding sites for this drug are distinct from those of the barbiturates and benzodiazepines. Loreclezole enhances GABA A -mediated chlorine currents via specific modulatory sites present on the β 2 and β 3 subunits of the GABA A receptor [29]. Moreover, Wingrove and co-workers put forward a hypothesis that the activity of loreclezole is dependent on the interaction between its 1,2,4-triazole moiety and the amide group of asparagine (Asn-289), which is located on the β 2 subunit of the GABA A receptor [11]. This fact prompted us to check, whether the obtained 1,2,4-triazole derivatives may act via a mechanism analogous to that of loreclezole. For this study, two derivatives (6a and 10a) were selected, as they are known to significantly elevate the seizure threshold and enhance the protective activity of classic antiepileptic drugs [12,13]. In the radioligand binding assay, the affinity of GABA administered alone or in combination with 1,2,4-triazole derivatives (6a, 10a), was compared. The IC 50 and K i values for GABA administered alone were 406.5 ± 42.0 nM and 96.6 ± 9.9 nM, respectively ( Table 5). The addition of 6a and 10a did not lead to an increase in the affinity of GABA to GABA A receptors. The obtained results permit us to conclude that despite the structural similarity of loreclezole and compounds (1a-30a), their anticonvulsant activities are achieved via a completely different molecular mechanism.

Chemistry
All reagents were purchased from Alfa-Aesar and Sigma-Aldrich and used without further purification. Melting points were determined by using Fischer-Johns apparatus (Sanyo, Japan) and are uncorrected. The 1 H-NMR spectra were recorded on a Bruker Avance instrument using DMSO-d 6 or CDCl 3 as solvents and tetramethylsilane (TMS) as an internal standard. Chemical shifts are expressed as δ (ppm). The infrared (IR) spectra were recorded in KBr using a Perkin-Elmer 1725X Fourier transform infrared (FTIR) spectrometer. The purity of the compounds was checked by thin-layer chromatography (TLC) on plates precoated with silica gel Si 60 F 254 , produced by Merck Co. (Darmstadt, Germany). The spots were detected by exposure to UV-lamp at λ = 254 nm. Elemental analyses were performed on AMZ 851 CHX analyzer and the results were within ±0.4% of the theoretical value. Molecular properties (i.e., M W , TPSA, and logP) of the compounds were calculated using Molinspiration online tool [30].

General Information
Adult male Swiss mice (weighing 22-26 g) that were kept in colony cages with free access to food and tap water, housed under standardized housing conditions, were used. After seven days of adaptation to laboratory conditions, the animals were randomly assigned to experimental groups each comprised of eight mice. Each mouse was used only once. Procedures involving animals and their care were conducted in accordance with current European Community and Polish legislation on animal experimentation. The experimental protocols and procedures described in this manuscript were approved by the First Local Ethics Committee at the Medical University in Lublin and the Second Local Ethics Committee at the University of Life Sciences in Lublin and complied with the European Communities Council Directive of 24 November 1986 (86/609/EEC).

Maximal Electroshock Seizure Test
The investigated compounds (1a-30a) were suspended in a 1% solution of Tween 80 in distilled water and administered intraperitoneally (i.p.) as a single injection, in a volume of 5 mL/kg body weight. Magnesium valproate, used as a reference drug, was directly dissolved in distilled water. Fresh drug solutions were prepared on each day of experimentation and administered at 15, 30, 60 and 120 min. before the initiation of maximal electroconvulsions. Electroconvulsions were produced by a current (0.2 s stimulus duration; 500 V, 50 Hz, fixed current intensity of 25 mA) delivered via ear-clip electrodes by a Rodent Shocker generator (constant-current stimulator Type 221, Hugo Sachs Elektronik, Freiburg, Germany). The criterion for the occurrence of seizure activity was the tonic hind limb extension. The animals were administered with a constant dose of 300 mg/kg of each of the examined compounds and were subjected to MES-induced seizures. The anticonvulsant activity of the active compounds administered i.p. at various pretreatment times was determined as their median effective doses (ED 50 values in mg/kg) in the MES-induced seizure test in mice. The animals were administered with different doses of the tested compounds so as to obtain a variable percentage of protection against maximal electroconvulsions, allowing the construction of dose-response relationship lines for each examined compound administered i.p. at various pretreatment times, according to Litchfield and Wilcoxon [33]. Each ED 50 value represents the dose of the studied compounds that is required to protect 50% of the animals tested against MES-induced seizures.

Chimney Test
The chimney test of Boissier et al. [34] was used to quantify the acute adverse-effect potential of  compounds 3a, 4a, 6a, 10a, 12a, 13a, 15a, 16a, 22a, and valproate on motor performance in mice. In this test, the animals had to climb backwards up a plastic tube (3 cm inner diameter, 30 cm length), and motor performance impairment was indicated by the inability of the mice to climb backward up the transparent tube within 60 s. The acute adverse effects of the 1,2,4-triazole derivatives and valproate administered alone were expressed as their median toxic doses (TD 50 ), representing the doses, at which the investigated compounds impaired motor coordination in 50% of the animals tested in the chimney test. To evaluate each TD 50 value, at least four groups of animals (each group consisted of eight mice) injected with various doses of the appropriate compound were challenged with the chimney test. A dose-response relationship line was calculated on the basis of the percentage of mice showing motor deficits by means of the log-probit method according to Litchfield and Wilcoxon [33].

Protective Index (PI)
The protective index for the investigated compounds was calculated by dividing a TD 50 value, as determined in the chimney test, by the respective ED 50 value, as determined in the MES test. The protective index is considered as an index of the margin of safety and tolerability between anticonvulsant doses and doses of the compounds exerting acute adverse effects e.g., sedation, motor coordination impairment, ataxia or other neurotoxic manifestations [35].

Radioligand Binding Assay
Binding experiments were conducted in 96-well microplates in a total volume of 200 μL of appropriate buffer. The reaction mixture included 50 μL solution of test compound, 50 μL of radioligand, and 150 μL the tissue suspension. Specific assay conditions for each receptor are shown in the table below (Table 6). Rat cerebral cortex was used for each receptor. The radioactivity was measured in MicroBeta TriLux 1450-liquid scintillation counter (PerkinElmer,Waltham, MA, USA). Radioligand binding data were analyzed using iterative curve fitting routines (GraphPAD/Prism, Version 3.0, San Diego, CA, USA). K i values were calculated from the Cheng and Prusoff equation [36]. The concentrations of analyzed compound GABA (γ-aminobutyric acid) ranged from 1 × 10 −11 to 1 × 10 −3 M. The percentage of specific binding was calculated for analyzed compounds in the screening test, which was determined in a concentration of 10 −6 M.

Conclusions
In conclusion, a series of 1,2,4-triazole-based compounds has been synthesized and their anticonvulsant and neurotoxic effects have been evaluated. Nine out of 30 tested derivatives exhibited various level of anticonvulsant activity. Median effective dose (ED 50 ) for the most potent compound, i.e., 5-(3-chlorophenyl)-4-(4-fluorophenyl)-2,4-dihydro-3H-1,2,4-triazole-3-thione, equaled 35.2 mg/kg. Based on the radioligand binding assay results, the hypothesis that the anticonvulsant activity of the mentioned compounds results from the direct or allosteric modulation of GABA A receptors should be excluded.