Design, Synthesis, and Evaluation of Anticonvulsant Activities of New Triazolopyrimidine Derivatives

Epilepsy, a severe brain disease affecting a large population, is treated mainly by antiepileptic drugs (AEDs). However, toxicity, intolerance, and low efficiency of the available AEDs have prompted the continual attempts in the discovery of new AEDs. In this study, we discovered a skeleton of triazolopyrimidine for the development of new AEDs. The design, synthesis, in vivo anticonvulsant activity evaluation of triazolopyrimidines (3a–3i and 6a–6e), and pyrazolopyrimidines (4a–4i) are reported. We found that most triazolopyrimidines showed anticonvulsive activity in the maximal electroshock (MES) and pentetrazol (PTZ)-induced seizure models. On the contrary, pyrazolopyrimidines (4a–4i) showed weak or no protective effects. Among the tested derivatives, compound 6d, holding a median effective dose (ED50) of 15.8 and 14.1 mg/kg against MES and PTZ-induced seizures, respectively, was found to be the most potent one. Moreover, the protection index (PI) value of 6d was significantly higher than that of the available AEDs such as valproate, carbamazepine, and diazepam. The antiepileptic efficacy of compound 6d was also observed in the 3-mercaptopropionic acid and bicuculline-induced seizure models. Antagonistic effects of flumazenil and 3-MP for the anticonvulsive activity of 6d and also the radioligand-binding assay confirmed the involvement of GABA receptors, at least benzodiazepine (BZD) receptor, in the anticonvulsant activity of compound 6d. The docking study of compounds 4e and 6d with GABAA receptor confirmed and explained their affinity to the BZD receptors.


INTRODUCTION
Epilepsy is a kind of common brain disease accompanied by repeated and unprovoked seizures. As one of the most common CNS disorders, it affects individuals of any age, any gender, and any ethnicity. About 3-4% of people in developed countries encountered this disease during their lifetime. For people in developing countries, this proportion is higher (Beghi and Hesdorffer, 2014). Epilepsy has a deleterious impact on society, economy, and specifically the physical and psychological well-being of patients (Perucca et al., 2018). According to the Global Burden of Disease Study (GBD) 2019, idiopathic epilepsy was ranked 20th in respect of disability-adjusted life years (DALYs) among 369 health conditions (GBD 2019 Diseases andInjuries Collaborators, 2020).
Antiepileptic drugs (AEDs) are the main method of epilepsy treatment. About two-thirds of patients with epilepsy can achieve seizure freedom. Since the 1990s, numbers of second-generation AEDs and third-generation AEDs (developed after 2000) have been introduced (Figure 1), expanding the opportunity for individual therapy for epilepsy patients. Some of them offer advantages in terms of favorable pharmacokinetics, fewer drug interactions, less adverse effects, and improved tolerability (Vidaurre and Herbst, 2019). Nevertheless, they did not considerably change the overall proportion of seizure-protected patients with epilepsy. Therefore, finding novel AEDs with high efficiency and low toxicity is still the focus of current AED therapy.
Triazole refers to a five-membered heterocycle with the molecular formula C 2 H 3 N 3 . It can be considered as an isostere of pyrazole, thiazole, imidazole, and oxazole. The electron richness and aromaticity of triazole enable it to freely bind with a wide range of biomacromolecules by interactions of pi-pi bonds, H bonds, and ion-dipole bonds. Therefore, it has been widely employed in various fields, especially in pharmaceutical research. Many drugs (such as triazolam, alprazolam, rufinamide, fluconazole, itraconazole, lamtidine, ribavirin, brassinazole, and sitagliptin) assembling the triazole moiety have already been used in clinics for the treatment of various diseases. Moreover, a terrific amount of triazolecontained derivatives have been designed and prepared to discover new anticonvulsants (Song and Deng, 2018), anti-Alzheimer (Xu et al., 2019), anticancer (Wen et al., 2020), antibacterial (Gao et al., 2019), antifungal (Peyton et al., 2015), and anti-inflammatory agents (Zhou and Wang, 2012).
Diazepam, a benzodiazepine drug, is often used to treat anxiety disorders, alcohol withdrawal syndrome, benzodiazepine withdrawal syndrome, seizures, and insomnia (Calcaterra and Barrow, 2014). The efficacy of diazepam comes from the activation of γ-aminobutyric acid (GABA) after it binds to benzodiazepine receptors (Riss et al., 2008). It is the top choice for the epileptic status to receive the intravenous or rectal administration of diazepam. Eighty percent of cases with different forms of epileptic status can be suppressed, and the effective proportion increases for generalized seizures (Loiseau, 1983). However, the side effects of diazepam including drowsiness, motor coordination disorders, and drug dependence limit its application in the clinics (Wu et al., 2016).
Based on these facts, a series of 7-substituted-[1,2,4]triazolo [1,5-α]pyrimidine derivatives (3a-3i) were designed and synthesized to develop new anticonvulsant agents in this study. The rationale behind this design refers to the similarity between the target compounds 3a-3i and diazepam ( Figure 2). The target compounds have the requirements for binding to benzodiazepine (BZD) receptors 1) an aromatic ring A, 2) a coplanar proton-accepting group B, and 3) a non-coplanar aromatic ring C. With an intention, the designed compounds act as diazepam to display anticonvulsant activity through  binding to BZD receptors. In addition, the triazole moiety in the compounds 3a-3i was replaced by pyrazole to give bioisosteric 4a-4i. These bioisosteres, without the proton-accepting group in a suitable position, were designed to enrich the structure-activity relationships (SARs) and verify the rationality of the aforementioned design. To obtain more SARs around the triazole derivatives, some triazole derivatives (6a-6e) with higher hydrophobicity were also prepared.

Instruments and Reagents
Melting points were determined in open capillary tubes and were uncorrected. 1 H NMR and 13 C NMR spectra were measured using an AV-300 spectrometer (Bruker, Switzerland). Highresolution mass spectra were recorded using a MALDI-TOF/ TOF mass spectrometer (Bruker Daltonik, Germany). The majority of chemicals were purchased from Aldrich Chemical Corporation (St Louis, United States).

Synthesis of
Considering compound 2a as an example: acetophenone (1.00 g, 8 mmol) and DMF-DMA (1.50 g, 13 mmol) were mixed and stirred at 100°C for about 10 h. The reaction was monitored by TLC (30% ethyl acetate in petroleum ether). After the reaction was completed and cooled to about 40°C, 10 ml of petroleum ether was added and stirred until it was completely dissolved. The mixture was placed in a fridge to maintain at 4°C overnight. The solid precipitated was filtered and dried to obtain compound 2a.  154.3, 140.5, 130.9, 128.1, 127.5, 92.2, 45.1, and 37.3 Considering compound 3a as an example: compound 2a (0.7 g, 4 mmol), 3-amino-1,2,4-triazole (0.34 g, 4 mmol), and 4 ml of glacial acetic acid were mixed and stirred at 120°C for about 7 h. The reaction was monitored by TLC (50% ethyl acetate in petroleum ether). After the reaction was completed, 40 ml of ice water was added to precipitate, which was filtered and recrystallized with 50% ethanol to obtain compound 3a. Compounds 3b-3i were prepared under the same conditions.
2.1.4 Synthesis of 7-Phenylpyrazolo[1,5-a]Pyrimidines (4a-4i) Considering compound 4a as an example: compound 2a (0.7 g, 4 mmol), 3-aminopyrazole (0.33 g, 4 mmol), and 4 ml of glacial acetic acid were mixed and stirred at 120°C for about 7 h. TLC (50% ethyl acetate in petroleum ether) was used to monitor the reaction process. After the reaction was completed, 40 ml of ice water was added to precipitate, which was filtered and recrystallized with 50% ethanol to obtain compound 4a. Compounds 4b-4i were prepared under the same conditions. 2.1.5 Synthesis of 4-([1,2,4]Triazolo[1,5-a] Pyrimidin-7-yl)Phenol (5) Compound 3i (1.13 g, 5 mmol), BBr 3 (5.00 g, 20 mmol), and 30 ml of dry CH 2 Cl 2 were mixed and stirred in an ice bath for about 6 h. TLC (66% ethyl acetate in petroleum ether) was used to monitor the reaction process. After it was completed, the reaction was quenched by ice water. The organic solvent evaporated under reduced pressure to leave the product dispersed in the water, which was filtered to give compound 5. Considering compound 6a as an example: compound 5 (0.42 g, 2 mmol), K 2 CO 3 (0.28 g, 2 mmol), bromopropane (0.30 g, 2.4 mmol), and 30 ml of acetonitrile were mixed and stirred at 90°C for about 8 h. TLC (50% ethyl acetate in petroleum ether) was used to monitor the reaction process. After the reaction was completed, the organic solvent was evaporated, and the residue was added to 30 ml of water and extracted by dichloromethane (3 ml × 30 ml). The combined dichloromethane was dried by MgSO 4 and evaporated in a vacuum to give a residue, which was purified by column chromatography (silica gel, petroleum ether/ EtOAc, 70:30) to provide compound 6a. Compounds 6b -6e were prepared under the same conditions using other haloalkanes.

Animals
Male Kunming mice with weights from 18 to 22 g were used in all the pharmacological experiments. The mice were housed in the standard cages with 10 each, where a controlled temperature (25°C ± 2°C) with a 12 h light/dark cycle was maintained. All animals had free access to food and water. All experiments and procedures involving animals were carried out according to the Guide for the Care and Use of Laboratory Animals (National Academies Press, Washington, D.C., United States). Approval from the local medical ethical committee was obtained before the experiment.

Maximal Electroshock Seizure Test
In the MES test, mice were subjected to ear stimulation using a 0.2 s 60-Hz 50-mA alternating current to elicit seizures. Tonic hind limb extension was considered as the seizure occurring, and the vanishing of tonic hind limb extension indicates the protection against the MES-induced seizures. Mice were injected intraperitoneally with test compounds at doses of 300, 100, and 30 mg/kg to evaluate their general anticonvulsant activity. Half and 4 h intervals were chosen to test the anti-MES activity. Based on the rough effective dose obtained in the aforementioned assay, compounds effective at the dose of 30 mg/kg were subjected to the quantitative assay to determine their median effective dose (ED 50 ). A series of doses of the tested compounds were injected into the mice with 10 in one group. When not less than three doses were obtained at which the mice were protected in 10-90%, the respective ED 50 , and 95% confidence intervals were calculated using these data by probit analysis.

Subcutaneous Pentylenetetrazole Test
In the PTZ test, mice were subcutaneously administered with PTZ at the dose of 85 mg/kg. At this dose, mice can be elicited a threshold seizure characterized by an episode of clonic spasms and tonic spasms. After being treated with test compounds at doses of 300, 100, and 30 mg/kg, the mice were considered protected if a single episode of clonic spasms of more than 5 s duration was not obtained. The quantitative anticonvulsant activity (ED 50 ) in the PTZ model was determined and calculated with the same method described in the MES test.

Neurotoxicity Screening
The rotarod test was used to evaluate the neurotoxicity of the synthesized compounds. Before the formal experiments, mice were trained to climb on a rotating rod of diameter 3.2 cm with the rotation speed of 10 rpm. Test compounds in the three doses were intraperitoneally injected into the trained mice. After 30 min, they were put on the rotating rod to observe their ability to maintain equilibrium. The rough neurotoxic dose was obtained in which mice cannot maintain equilibrium on the rod for at least 1 min. The quantitative neurotoxic dose (TD 50 ) of the test compounds was measured and calculated with the same method described in the MES test.

Bicuculline Test
In the BIC-induced seizure test, groups of 10 mice were administered with 30 mg/kg of test samples or vehicle. Then, 30 min later, the minimum dose of BIC (5.4 mg/kg) that could completely induce seizures for mice was subcutaneously injected into the test mice. The mice were placed into separate cages each, and the number of clonic seizures, tonic seizures, and mortality was recorded for 30 min.

3-Mercaptopropionic Acid Test
In 3-MP-induced seizure test, groups of 10 mice were administered with 30 mg/kg of test samples or vehicle. After 30 min, the minimum dose of 3-MP (60 mg/kg) that could completely induce seizures for mice was intraperitoneally injected into the test mice. The mice were placed into separate cages each, and the number of clonic seizures, tonic seizures, and mortality was recorded for 30 min.

Docking
Molecular docking was performed by Discovery Studio (release 2019). The three-dimensional (3D) GABA A receptor binding with diazepam was obtained from Richter's publication (Richter et al., 2012). The hydrogen atoms of the GABA A receptor were added, and water and diazepam were removed. The 3D structures of test molecules were constructed by ChemDraw 16.0 software; then, they were transferred into the Discovery Studio platform to minimize energy. After the aforementioned preparation finished, the molecular docking was performed for the test molecules and GABA A . After the completion of molecular docking, the docking gave a batch of poses, which were ranked and selected by CDOCKER interaction energy. The lowest energy conformation of the ligand-receptor complex was evaluated, and the interaction type between GABA A receptor and test molecules was analyzed by Discovery Studio (release 2019).

Prediction of Drug Likeliness Parameters
Computational calculations were carried out to predict the drug likeliness properties of target compounds. The polar surface area (TPSA), number of rotatable bonds (n-ROTB), miLogP, number of hydrogen donor (HBD) and acceptor (HBA), and violations of Lipinski's rule of five were calculated by using Molinspiration Property Calculator (Molinspiration Cheminformatics, 2021), which is a free online property calculating toolkit. Absorption (% ABS) was calculated using an equation % ABS = 109-(0.345 * PSA) (Zhao et al., 2002). LogBB, a ratio of the steady-state concentrations of a compound in the brain to that in the blood [i.e., log (C brain /C blood )], was calculated using Clark's equation as follows: logBB = −0.0148 × PSA + 0.152 × clogP + 0.139 (Clark, 1999).

Radioligand Receptor Binding Assay
The radioligand receptor binding studies were performed according to the procedure described by Ahmadi et al. (2014). Selected compounds were evaluated for their affinity to the GABA A BZD receptor by the radioligand receptor binding assay. Male SD rats with weights ranging from 220 to 260 g were sacrificed after a spinal dislocation. The cerebral tissues except the brainstem was immediately removed and homogenized for 30 s in 10 times the volume of ice-cold Tris-HCl buffer (50 mM, pH 7.4). The homogenates were centrifuged at 1,000 g for 10 min. The resulting supernatant was centrifuged at 30,000 g for 20 min. The precipitate was washed three times with ice-cold buffer by resuspension and re-centrifugation. All of the centrifugation procedures were performed at 4°C. The obtained membrane preparation was maintained under −4 Fahrenheit until it was used in 20 days. The protein concentration was estimated by the Bradford method. After the membrane preparation, saturation and competition experiments were conducted to obtain the inhibition of radioligand-specific binding (IC 50 ) and affinity (K i ) of the test compounds and diazepam. In the saturation experiment, 200 μg of membrane protein was incubated with nine different concentrations (0.25-25 nM) of [ 3 H]-flumazenil at 30°C for 35 min. The total binding (TB) (receptor + radioligand), nonspecific binding (NSB) (receptor + radioligand + excess diazepam), and specific binding (SB) (TB-NSB) were measured at the nine radioligand concentrations. The receptor-binding affinity (K d ) of [ 3 H]-flumazenil and the maximum binding capacity of BZD receptors (Bmax) were calculated from the saturation experiment data. In the competition experiment, 200 μg of membrane protein in Tris-HCl buffer (50 mM, pH 7.4) was incubated with 0.5 nM [ 3 H]flumazenil and different concentrations (50 pm-5 mM) of test compounds at 30°C for 35 min. After incubation, the assay was terminated by an ice bath. The activity of each tube was measured by a liquid scintillation counter after centrifugation and wash treatment. The competition curve was achieved by plotting the %SB of radioligand vs. the concentrations of test compounds. %SB = 100 (BC-NSB)/(TB-NSB). BC represents the binding in the presence of the test compound. IC 50 inhibiting the [ 3 H]-flumazenil-specific binding and affinity for BZD receptors (K i ) of test compounds were calculated according to the Cheng-Prusoff equation. All the experiments were carried out in triplicates.

Evaluation of the Anticonvulsant Activity and Neurotoxicity
The anticonvulsant activities of targets were evaluated using seizure models induced by MES and PTZ. The neurotoxicity was evaluated via the rotarod test. Structurally, target molecules could be categorized into three groups: 3a-3i (containing the triazole moiety), 4a-4i (containing the pyrazole moiety), and 6a-6e (containing the triazole moiety and higher hydrophobicity than the series of 3a-3i). In each series, substituents on the benzene ring at position 7 were altered to optimize the anticonvulsant activity.
Initially, all the target compounds were screened at three dosages (30, 100, and 300 mg/kg). As exhibited in Table 1, compounds 3a, 3b, 3c, 3e, 3h, and 3i from the first series of compounds displayed medium-to-weak anticonvulsant activity at 100 or 300 mg/kg at 0.5 h in the MES test. None in the second series of compounds (4a-4i) displayed anticonvulsant activity even at the highest dose of 300 mg/kg in the MES test. On the contrary, all compounds in the series of 6a-6e showed anticonvulsant activity in three dosages. Compounds 6c, 6d, and 6e displayed the best protective effect at the dosage of 30 mg/kg and remained active at 100 or 300 mg/kg after 4 h.
The performance of designed compounds in the PTZ model was generally similar to that in the MES test. Compounds 3a, 3b, 3c, 3e, 3h, and 3i from the first series of compounds displayed protection at the dosage of 100 or 300 mg/kg in the PTZ model at 0.5 h. Compounds 4e, 4h, and 4i in the second series exhibited anticonvulsant activity at the highest dose of 300 mg/kg at 0.5 h intervals. Compounds 6a and 6b protected the mice against PTZinduced seizure at 300 mg/kg at 0.5 h interval. Compounds 6c, 6d, and 6e showed protection for mice at the dose of 30 mg/kg at 0.5 h interval.
None of the target molecules except 6d and 6e exhibited neurotoxicity at the maximum dosage of 300 mg/kg in the rotarod test. Some mice injected with 6d or 6e in 300 mg/kg could not maintain balance on a rotating rod. In addition, no compounds were observed to have side effects in common with antiepileptics, such as sedation, hypnosis, and anxiety (with features of bradykinesia or running) at the dosage of 100 mg/kg.
According to the visible antiepileptic activity of compounds 6c, 6d, and 6e indicated in the initial trials, the quantitative evaluation trials were conducted to obtain the ED 50 and TD 50 values. The quantitative data of compounds 6c, 6d, and 6e and Frontiers in Chemistry | www.frontiersin.org June 2022 | Volume 10 | Article 925281 those of the positive drugs valproate, carbamazepine, and diazepam, are listed in Table 2. Compound 6d with an ED 50 value of 15.8 mg/kg was found to have the strongest effect against MES-induced seizures in mice. Meanwhile, compounds 6c and 6e displayed ED 50 values of 25.3 and 38.0 mg/kg, respectively. All compounds tested were more potent than valproate but less active than that of carbamazepine and diazepam in the MES-induced seizure model. In the PTZ model, compounds 6c, 6d, and 6e gave ED 50 values of 23.7, 14.1, and 28.4 mg/kg, respectively. They were more active than valproate and carbamazepine but less active than that of diazepam in the PTZ-induced seizure model. As for the neurotoxicity, all synthesized compounds, including the standard drug valproate, showed neurotoxicity at very high dosages (TD 50 ≥ 317 mg/kg). However, the TD 50 values for the positive controls carbamazepine and diazepam were 44.0 and 3.3 mg/kg, respectively, being remarkably less than those of the synthesized compounds. Based on the obtained results from MES, PTZ, and rotarod tests, the protective index (PI = TD 50 /ED 50 ) was calculated and is presented in Table 2. The results revealed that the designed compounds were able to display good anticonvulsant activity with a high safety profile. Although the anti-MES and anti-PTZ activities of the compounds 6c-6e were lower than those of diazepam, their PI values were significantly higher than those of the latter.

Antiepileptic Activity Evaluation of Compound 6d in Other Seizure Models
To further confirm the antiepileptic activity, two seizure models induced by 3-MP and BIC were used. 3-MP, a competitive inhibitor of glutamate decarboxylase (GAD), can decrease the level of GABA in the brain by inhibiting its synthesis (Crick et al., 2014;Enrique et al., 2017). As shown in Table 3, 3-MP induced convulsions in 100% of the mice at 60 mg/kg. 3-MP and 6d combination therapy showed complete inhibition on mice's clonic seizures, tonic seizures, and death in comparison with the results following treatment of 3-MP alone. BIC could induce seizures in mice as a competitive GABA A receptor antagonist (Velisek, 2006). BIC caused seizures in 100% of the mice at 5.4 mg/kg. As indicated in Table 4, carbamazepine significantly inhibited tonic seizure and death. Similarly, compound 6d inhibited tonic seizure, death (both from 100% to 0%, p < 0.001 for seizures and death), and clonic seizure (from 100% to 10%, p < 0.05) significantly. The superiority of compound 6d in inhibiting clonic seizures, tonic seizures, and death in seizure models caused by 3-MP and BIC, compared to carbamazepine, further identified the antiepileptic effect of 6d.

Study on the Possible Mechanism of Action
As is known, diazepam belongs to BZDs. The anticonvulsant activity of diazepam can be reversed by the BZD-receptor antagonist flumazenil (Kim et al., 2020). To understand the possible mechanism of the anticonvulsant activity of the most potent compound 6d, we checked if the protection activity of the compound 6d in the MES and PTZ tests could be antagonized by flumazenil. As shown in Table 5, administration of compound 6d at 30 mg/kg alone gave full protection to mice against the MES and PTZ-induced seizures, while pretreatment of flumazenil completely reversed the anti-MES and anti-PTZ activities of compound 6d. These results suggested that the anticonvulsant action of compound 6d might be related to BZD receptors.
Thiosemicarbazide (TSC), a competitive inhibitor of the GABA synthesis enzyme, can decrease the GABA level in the brain by inhibiting the synthesis of GABA (Collins, 1973). The influence of TSC on the antiseizure activity of 6d was investigated to further confirm that the GABAergic system was involved in the SCHEME 1 | Synthetical route of target compounds 3a-3i and 4a-4i. SCHEME 2 | Synthetical route of target compounds 6a-6e.
Frontiers in Chemistry | www.frontiersin.org June 2022 | Volume 10 | Article 925281 8 antiseizure mechanism of 6d. As displayed in Table 5, the normal mice were completely protected in the MES model at 30 mg/kg, while the effects vanished in the treated mice. The anticonvulsant activity of 6d was completely reversed after the pretreatment with TSC at the dosage of 25 mg/kg for 3 days in the MES and PTZ models. Based on this finding, it could be concluded that compound 6d acts its antiepileptic activity by regulating the GABA function in the brain.
Frontiers in Chemistry | www.frontiersin.org June 2022 | Volume 10 | Article 925281 9 The radioligand receptor-binding study is a very powerful tool in the study of receptors and their ligands. To verify the hypothesis that these compounds are binding and affecting the GABA A receptor chloride ionophore complex via the BZP receptor, the affinity of compound 6d to the BZD receptor was evaluated using [ 3 H] flumazenil as the specific ligand. The results showed that compound 6d and diazepam have an affinity to the BZD receptor in nanomolar concentrations (IC 50 = 8.4 and 1.3 nM and K i values = 3.9 and 0.61 nM, respectively). This result further confirmed that the anticonvulsant activity of 6d was mediated by BZD receptors ( Table 6).

Molecular Docking Study
To explore the interaction pattern of the synthesized compounds to the BZD-binding site of the GABA A receptor, the docking of active compounds 3e and 6d with the BZD-binding site of the GABA A receptor (α1β2γ2) was conducted by Discovery Studio 4.0 Client. Interactions of the diazepam and compounds 3e and 6d with amino acids of BZD-binding pocket of the GABA A receptor are illustrated and shown in Figure 3. As shown in Figure 3, the most important residues in the binding mode of diazepam are Ser204, Thr206, Tyr209, His101, and Tyr159. These results are consistent with those reported by Ernst and Foroumadi (Richter et al., 2012;Mohammadi-Khanaposhtani et al., 2016). Interactions of compound 3e and the GABA A receptor were identical with those of diazepam. Critical amino acid residues SER204, Thr206, Tyr209, His101, and Tyr159 were all involved, but the amino acid residue Thr206 formed a hydrogen bonding with the N atom in position 3. The benzene ring in compound 3e was responsible for π-π interaction with His101 and Tyr159. The binding model of compound 6d was different from diazepam and compound 3e. Amino acid residues Phe77, His101, Arg132, and Tyr 209 were involved in the interactions between compound 6d and the GABA A receptor. It is worth mentioning that the N atom in position 3 was also a critical one that was involved in hydrogenbonding interaction with His 101.

Prediction of Drug-Like Parameters
Drug-like property is an important basis for evaluating the potential of drug development. The drug-like properties of targeted compounds, such as molecular weight (MW), octanol/ water partition coefficients (ClogP), the number of H-bond donors (HBD), the number of H-bond acceptors including (HBA), the number of rotatable bonds (ROTB), and the polar surface area (tPSA), were predicted by the computational study. The PSA of a molecule effectively represents the portion of its surface belonging to polar atoms, such as oxygen, nitrogen, and    attached hydrogen, and is a descriptor related to the passive molecular transport of a molecule through membranes. From the PSA, intestinal absorption (%ABS) was calculated according to the algorithm described by Zhao et al. (2002). It has been proven by Palm et al. (1996) that drugs with a PSA value below 60 were completely absorbed in the intestine based on Caco-2 cell studies.
The PSA values of most target compounds were all below 60 except for compounds 3g and 4g, and the level of intestinal absorption (%ABS) was bigger than 90% except compounds 3g and 4g. The following conclusions can be drawn that these compounds might display good transport properties in the intestines.
The "rule of five" is a set of simple molecular descriptors, which was established by Lipinski, which states that most "drug-  like" molecules have common parameters, including LogP ≤ 5, Mw ≤ 500, HAB ≤ 10, HBD ≤ 5, and ROTB ≤ 10 (Lipinski et al., 2001). Molecules that violate one of the aforementioned rules may have bioavailability problems. It can be seen from Table 7 that all synthetic compounds meet all parameters of Lipinski and have good drug-likeliness. Good blood-brain barrier (BBB) permeability is necessary for CNS drugs (Fu, 2018) To predict the BBB penetration, logBB of the compounds was calculated according to the algorithm described by Clark (1999), which refers to the ratio of the steady-state concentrations of compounds in the brain and blood [i.e., log (C brain /C blood )]. As reported, compounds with logBB < −1.0 are poorly distributed in the brain (Abraham et al., 1997). The logBB values of synthesized compounds are calculated and presented in Table 6. The measured logBB values for compounds 3a-3i, 4a-4i, and 6a-6e were ranging from −0.95 to 0.23. The most active compound 6d had logBB of −0.05, being the acceptable value for a CNS-active compound.

DISCUSSION
Diazepam, a classic benzodiazepine drug, is widely used to treat epilepsy, especially the epileptic status (Calcaterra and Barrow, 2014). However, the side effects of diazepam including drowsiness, motor coordination disorders, and drug dependence limit its application in the clinics (Wu et al., 2016). To obtain new anticonvulsants, which are highly safe, triazole derivatives 3a-3i were designed and synthesized based on the pharmacophoric features of benzodiazepine agonists. The MES and PTZ models, as the most popular and widely used animal seizure models, were used in our anticonvulsant activity screen. As reported, the MES test is considered to be a predictor of the possible therapeutic effect on generalized tonic-clonic seizures, while the PTZ test represents an effective model for selecting candidates for human generalized myoclonic seizures (Loscher, 2011).
The preliminary study showed that some triazole derivatives synthesized displayed anticonvulsant activity in both seizure models. To verify whether the triazole moiety is necessary for the anticonvulsant activity, a series of pyrazole derivatives (4a-4i), as their bioisosterics, were prepared, and their anticonvulsant activity was screened. It was disappointing, but as expected, pyrazole derivatives (4a-4i) did not display any anticonvulsant activity in the MES test even at the highest dose. This suggested that triazole plays a critical role in anticonvulsant activity.
Considering that the BBB permeability is necessary for the compounds to work in vivo, we prepared additional compounds (6a-6e) with higher hydrophobicity. It is gratifying that these compounds (6a-6e) with higher hydrophobicity showed better performance in the seizure models. Compound 6d, whose ED 50 values are 15.8 and 14.1 mg/kg against MES-and PTZ-induced seizures, respectively, in mice, was the most promising one in this study. Hydrophobicity is an important parameter for the anticonvulsant activity of triazole derivatives (Song and Deng, 2018;Aggarwal and Sumran, 2020). Many triazole derivatives holding a suitable aryl or alkyl group gave a better anticonvulsant activity in their analogs (Deng et al., 2014a;Deng et al., 2014b). From Table 1, it was noticeable that the hydrophobicity of compounds 6a-6e affects their anticonvulsant activity. For compounds 6a-6d, with the extension of the alkyl chain, the hydrophobicity of the compounds increased gradually, and the activity increased accordingly. The ClogP values of 6a, 6b, 6c, 6d, and 6e were calculated by ChemDraw 16.0 software as 1.56, 2.09, 2.62, 3.15, and 2.27, respectively. The bigger hydrophobicity of 6d gives it higher BBB permeability and higher in vivo anticonvulsant activity.
The comparison of the PTZ and MES test results obtained showed that all active molecules possessed better performance in the PTZ model than in the MES model. Some studies indicated that PTZ diminishes the GABAergic function (Macdonald and Barker, 1977), through competitive antagonism to the BZD receptor (Rehavi et al., 1982). Correspondingly, drugs that enhance GABA A receptor neurotransmission, such as BZDs (Shafie et al., 2020;Navidpour et al., 2021), can inhibit seizures induced by PTZ. PTZ, 3-MP, and BIC also induced seizures through GABA A receptor neurotransmission. The ability of compound 6d to block PTZ, 3-MP, and BIC-induced seizures may be attributed to its modulatory effect on GABA A receptor neurotransmission.
The anticonvulsant effect of compound 6d against MES-and PTZ-induced seizures was completely reversed by the BZD-receptor antagonist flumazenil, which suggested that BZD receptors might be involved in the anticonvulsant effect of compound 6d. The radioligand-binding assay further confirmed that the anticonvulsant activity of compound 6d can be attributed to the direct modulation of the BZD receptor. This result is consistent with our initial design and also supported by the docking study. Interactions of compound 3e and GABA A receptor were identical with those of diazepam. The amino acid residue Thr206 formed a hydrogen bond with the N atom in position 3, which was missed in the inactive derivatives 4a-4i. The binding model of compound 6d was different from the diazepam and compound 3e. But the N atom in position 3 was also a critical atom, which was involved in a hydrogen bonding interaction with His101 in the interactions of compound 6e and the GABA A receptor.
BZDs represent one of the important classes of anticonvulsant agents with excellent activity against generalized tonic-clonic and partial seizures. It has been clarified that it works by binding to the GABA A receptors. This binding increases the frequency of chloride ion channel opening, which facilitates the inhibitory effects of GABA. However, certain side effects are associated with the short-and long-term use of benzodiazepines, which includes sedation, confusion, drowsiness, ataxia, and memory impairment (Schmitz, 2016). Since chlordiazepoxide was found in 1960, a large number of BZD drugs have been found and approved, among which clobazam, clonazepam, nordazepam, nimetazepam, lorazepam, and diazepam are used in the treatment of epilepsy. All of them shared the common skeleton, that is, the benzodiazepine moiety, and gave the common SARs (Riss et al., 2008). In the last decade, new BZDs were designed and reported to find novel anticonvulsant agents with less adverse effects (Arora et al., 2020;Maramai et al., 2020). In 2017, Shao et al. (2018)reported a series of 7-(benzylamino)-1H-benzo[b] [1,4]diazepine-2,4(3H,5H)-dione derivatives. It was found that the most active compound was substituted by 4-fluoro, which holds an ED 50 value of 36.5 mg/kg in the MES test and 68.2 mg/kg in the scPTZ test. Compared to that, compound 6d reported here exhibited better anti-MES and anti-PTZ activities. More important is that the structural skeleton jumped out of the bondage of benzodiazepine, which provides a new skeleton for the research of antiepileptic drugs based on the BZD-binding site of the GABA receptor.

CONCLUSION
In conclusion, the design, synthesis, and in vivo anticonvulsant activity of a series of triazole derivatives (3a-3i), pyrazolecontained bioisosterics (4a-4i), and analogs with higher hydrophobicity (6a-6e) were reported. More than half in the series of 3a-3i and all in the series of 6a-6e showed anticonvulsant activity in both MES-and PTZ-induced seizure models. On the contrary, pyrazole-containing bioisosterics (4a-4i) showed weak or no protective effects in both models. The triazole moiety played an important role in their anticonvulsant activity. This supposition was supported by the docking study. Among the synthesized compounds, the triazole derivative 6d was the most potent compound in the MES and PTZ tests. Although the anti-MES and anti-PTZ activities of the compound 6d were lower than those of diazepam, their PI values were significantly higher than those of the latter. Moreover, the experimental investigation of the compound 6d's mechanism of action suggested that compound 6d works as an anticonvulsant agent via regulating the GABA function such as the activation of the BZD receptor. Prominent anticonvulsive performance and low neurotoxicity, combined with excellent drug-like properties of the triazolopyrimidines reported in this work, indicate their potential in the development of new anticonvulsant agents.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material; further inquiries can be directed to the corresponding authors.