New Anticonvulsant Agents

: The search for antiepileptic compounds with more selective activity and lower toxicity continues to be an area of intensive investigation in medicinal chemistry. This review describes new anticonvulsant agents representing various structures for which the precise mechanism of action is still not known. Many of the compounds presented in this review have been tested according to the procedure established by the Antiepileptic Drug Development Program of the Epilepsy Branch of the National Institute of Neurological Disorders and Stroke, National Institute of Health, USA. The newer agents include sulfonamides, amino acids, amides (analogs of g -vinyl GABA, N -benzylamides, 2,6-dimethylanilides, carboxyamides, hydroxyamides, alkanoamides); heterocyclic agents ((arylalkyl)imidazoles, pyrrolidin-2,5-diones, lactams, semi- thiosemicarbazones, thiadiazoles, quinazolin-4(3H)-ones, 2,5-disubstituted 1,2,4-thadiazoles, xanthones, derivatives of isatin) and enaminones. These new structural classes of compounds can prove useful for the design of future targets and development of new drugs.


INTRODUCTION
Epilepsy is a common neurological condition, affecting 0.5 to 1% of the population worldwide (45-100 million people) [1]. Conventional antiepileptic drugs (AEDs) phenobarbital, primidone, phenytoin, carbamazepine, ethosuximide and benzodiazepine, are widely used but exhibit an unfavorable site effect profile and failure to adequately control seizures. In the recent years several new drugs (oxcarbazepine, lamotrigine, topiramate, gabapentin, zonisamide, tiagabine, fosphenytoin, vigabatrin and felbamate) have been added to the list of therapeutic agents against epilepsy. However, there is a significant group of patients (up to 30%) who are resistant to the available antiepileptic drugs. The long-established AEDs control seizures in 50% of patients developing partial seizures and in 60-70% of those developing generalized seizures [2][3][4][5][6]. Hence, there is an urgent need to develop new AEDs [7].
The search for antiepileptic compounds with a more selective activity and lower toxicity continues to be an area of investigation in medicinal chemistry. A rational drug design process of a new anticonvulsant could be achieved in several ways [8,9]. The first strategy is the identification of new targets through better understanding of molecular mechanisms of epilepsy. Another way is to modify already existing drugs and formulations. AEDs belong to many different chemical classes of compounds, including: hydantoines, iminostilbenes, barbiturates, benzodiazepines, valproate, imides, oxazolidine-2,3-diones, sulfonamides and miscellaneous agents [10]. The efficacy of AEDs is due to the main activities which include interaction with ion channels or neurotransmitter systems [11][12][13][14][15][16][17]. Currently available AEDs can be broadly classified into four categories: (1) those whose main action relates to the inhibition of sustained repetitive firing, through blockage of voltage-dependent sodium channels and consequent inhibition of the release of excitatory neurotransmitters (phenytoin, carbamazepine, oxcarbazepine); (2) those which enhance GABA-ergic transmission (benzodiazepines, barbiturates, vigabatrin, tiagabine); (3) those stabilizing thalamic neurons through inhibition of T-type calcium channels (ethosuximide) and (4) those possessing a combination of the above actions, often coupled with additional mechanisms (valproic acid, gabapentin, lamotrigine, topiramate, zonisamide, felbamate). However, this classification has limited value because the majority of AEDs possess more than one mechanism of action, which may account for their efficacy, and it is also the fact that some of the clinically used drugs have not been linked with a specific site the brain, and the exact mechanisms of many AEDs remain unknown [18,19].
The new AEDs and anticonvulsant agents have been reviewed during last few years [20][21][22][23][24]. The chemical diversity and various mechanisms of action of anticonvulsants make it difficult to find a common way of identifying new drugs. Novel anticonvulsant agents are discovered through conventional screening and/or structure modification rather than a mechanism-driven design. Therefore, drug identification is usually conducted via in vivo screening tests, on the basis of seizure type rather than etiology.
This review presents new anticonvulsant agents representing various structures for which the precise mechanism of action is still not known. The newer agents include heterocyclic compounds, sulfonamides, amino acids, amides, enaminones and others. These new structural classes of compounds can be useful for the design of future targets and development of new drugs.

THE ANTICONVULSANT SCREENING PROJECT (ASP)
Many of the compounds presented in this review have been tested according to the procedure established by the Antiepileptic Drug Development (ADD) Program [25,26]. Since 1975, the Epilepsy Branch of the National Institute of Neurological Disorders and Stroke, National Institute of Health, through its Antiepileptic Drug Development (ADD) Program, has collaborated with the pharmaceutical industry in developing new therapeutic agents for the treatment of seizure disorders. In 1993 felbamate, the first new drug in nearly two decades, was approved for sale in the United States. Felbamate's development was a collaborative effort of both a pharmaceutical sponsor and the ADD Program. Two other drugs, topiramate and remacemide have been identified in the Branch's preclinical evaluation and are now in late clinical development. Several other drugs, losigamone, retigabina, soretolide and rufinamide ( Fig. 1) are now in early clinical development [27][28][29]. The ADD Program boasts of both preclinical and clinical trial components. The preclinical component consists of drug discovery and toxicology elements. The Anticonvulsant Screening Project uses for its initial screening procedure two major convulsant tests: maximal electroshock (MES) and subcutaneous pentylenetetrazole (scPTZ) as well as a toxicity screen (rotorod in mice, positional sense and gait in rats). The MES is a model for generalized tonic-clonic seizures. The behavioral and electrographic seizures generated in this model are consistent with the human disorder [30]. This model identifies those compounds which prevent the spread of seizures. The scPTZ seizure test is a model which primarily identifies compounds that raise seizure threshold. The behavioral seizure produced is not typical of absence epilepsy but clonic in nature. Like other rodent models of absence seizures, PTZ-induced seizures are potentiated by γaminobutyric acid (GABA) agonist. With some minor exceptions, the pharmacological profile of the scPTZ seizure model is consistent with the human condition [30,31]. All clinically active anticonvulsants have been found to be protective in at least one of these two tests.
Compounds that possess significant anticonvulsant activity in rats and in mice and do not exhibit substantial neurotoxicity or lethality are considered for the ADD Program's multiphase evaluation to establish a compound's pharmacodynamic/pharmacokinetic profile. During the past five years, over 4000 compounds have been evaluated by the ADD Program. Selected results of these evaluations have been published and are presented in this review.

Heterocyclic Analogs of Carabersat
Carabersat (SB-204269) was originally designed by modifying the structure of cromakalim (Fig. 2). This later drug represents a class of antihypertensive agents which act via the relaxation of vascular smooth muscle caused by the opening of ATP-sensitive potassium channels. The potassium channel activators which more readily penetrate the central nervous system (CNS) may have therapeutic potential in the treatment of epilepsy [11]. Replacement of the 2-pyrrolidinone group of cromakalim by the fluorobenzoylamino group has introduced anticonvulsant activity [32]. Carabersat, the benzopyran derivative, is a chemically novel AED with a novel mechanism of action and a stereospecific CNS binding site. It has potent oral anticonvulsant properties in a range of rat seizure models, with potency and efficacy equivalent to or better than carbamazepine and lamotrigine. Carabersat is currently being proposed for the treatment of epilepsy and migraine prophylaxis [33,34]. Subsequent exploration of structureactivity-relationships, using high-throughput screening of the novel SB-204269 binding site led to the identification of anticonvulsant active series of the isomeric tetrahydroisoquinolinyl (THIQ) benzamides 5-, 7-and 8-substituted [35,36]. Compound SB-270664, the 7-substituted THIQ was   3). In this series a gem dimethyl group is incorporated to prevent aromatization, and replacement of the THIQ benzo ring with pirydyl represents an attempt to reduce hydroxylation. In this new series, compound 1 was identified as new potential agent displaying excellent anticonvulsant activity and an encouraging pharmacokinetic profile in vivo. Compound 1 has high affinity at the [ 3 H]-SB-204269 binding site (pK i 8.7), suggesting a novel mechanism of action, comparable with carabersat. In vivo, the rat maximal electroshock threshold (MEST) test at 2 mg kg -1 p.o. compound 1 showed a good level of anticonvulsant activity.
In the rat supraMES model, analysis by 'ALLFITrd-quo compound 1 produced an ED 50 value of 3.9 mg kg -1 . This figure is comparable with that obtained with carabersat in the same model under identical conditions (ED 50 of 6.3 mg kg -1 ). Compound 1 has excellent aqueous solubility (>1 mg mL -1 ) and has been shown to have an encouraging pharmacokinetic profile and good in vivo activity in preclinical anticonvulsant models in rats [37].

Sulfonamides
Acetazolamide {N-(5-aminosulfonyl)-1,3,4-thiadiazol-2yl acetamide} and methazolamide are old members of this class, which have shown anticonvulsant activity (Fig. 4). These drugs are bifunctional 5-membered heterocycles, comprised of a sulfonamide and an amide as well as a 1,3,4thiazole nucleus, constituting potent carbonic anhydrase (CA) inhibitors. Several new sulfonamide CA inhibitors such as topiramate and zonisamide have been and are still used as antiepileptic drugs (Fig. 4). The anticonvulsant effects of these or related sulfonamides are probably due to CO 2 retention secondary to the inhibition of red cell and brain enzymes, but other mechanisms of action, such as blockade of sodium channels and kainite/AMPA receptors, as well as enhancement of GABA-ergic transmission, have also been hypothesized/proven for some of these drugs. Acetazolamide and methazolamide are still clinically used nowadays in some forms of epilepsy, but they are considered to belong to a minor class of antiepileptic agents. The recently developed drug topiramate is a very effective antiepileptic, and it also acts as a strong CA inhibitor with a potency similar to that of acetazolamide against the physiologically important isoenzyme CA II [38][39][40]. Acetazolamide, topiramate and zonisamide possess a sulfamate moiety, which is essential for their anhydrase inhibition. Several new carbonic anhydrase inhibitors derivatives of sulfonamide have been developed by Scozzafava, Supuran and coworkers [41][42][43][44][45][46]. A series of aromatic/heteroaromatic sulfonamides incorporating valproyl and other lipophylic moieties has been found to possess potent CA inhibitory properties as well as anticonvulsant in vivo effects. The hybride compound, valproyl derivative of acetazolamide (5-valproylamido-1,3,4thadiazole-2-sulfonamide, 2) was one of the best hCA I and hCA II (human cloned isoenzymes) inhibitors in the series and it exhibited very strong anticonvulsant properties in the MES test in mice [45]. Inhibition data of carbonic anhydrase K i (nM) hCA I and hCA II are as follows: for topiramate 250, 5; for compound 2 56, 6. Both drugs efficiently protected mice against seizures induced by electroshock at a dosage of 30 mg kg -1 ; the protection rate was 75-100% at 0.5. h and 25-100% at 3 h after drug administration. Investigation at a decreased dosage of 10 mg kg -1 of compound 2 showed a rate of protection in the range of 25-44% at 0.5 h and 87-100% at 3 h after administration. Several other 1,3,4thiadiazole-sulfonamide derivatives possessing potent CA inhibitory properties and substituted with various alkyl/ arylcarboxamido/sulfonamide/ureido moieties in the 5position have been investigated for their anticonvulsant effects in the same animal model. It was observed that some lipophilic derivatives, such as 5-benzoylamido-, 5toluenesulfonylamido-, 5-adamantylcarboxamido-, and 5-pivaloylamido1,3,4-thiadiazole-2-sulfonamide, show promising in vivo anticonvulsant properties and that these compounds may be considered as interesting leads for developing anticonvulsant or selective cerebrovasodilator drugs [46]. A new series of sulfonamides incorporating adamantyl moieties attached to the scaffolds of aromatic/heteroaromatic sulfonamides have shown good inhibitory potency against two human CA isozymes, compound 3 and 4 ( Fig. 5), inhibition data of carbonic anhydrase K i (nM) hCA I and hCA II are as follows: for compound 3 850, 10; for compound 4 77-, 12. The anticonvulsant activity of these two CA inhibitors against MES tests in mice revealed that after i.p. injection (30 mg kg -1 ), compounds 3 and 4 exhibit good protection against electrically-induced convulsions (>90%). Their ED 50 values were 3.5 and 2.6 mg kg -1 , respectively.

Derivatives of Amino Acids
Amino acids that are functionalized at both the N-and Cterminal are proven potent anticonvulsant agents . In the recent years, Kohn and coworkers have reported on the anticonvulsant activity of a series of functionalized amino acids [47][48][49][50][51][52][53][54][55][56][57][58][59][60][61] (FAA) 5 ( Fig. 6). A structure activity relationships study of over 250 compounds has yielded 12 compounds with anticonvulsant activity in rodents that is equal to or greater than phenytoin according to the MESseizure test. N-Benzyl-2-acetamidopropionamide 9 was the parent compound in this series [55]. These investigations have enabled the selection of (R)-N-benzyl-2-acetamido-3methoxy-propionamide 10 as the lead compound ( Table 1). Compound 10 has now entered phase II clinical trials for the treatment of epilepsy and neuropatic pain. Extensive SAR study of FAA 5 revealed that the N-terminus is an important FAA structural unit. In the initial design of FAA, the Nterminal amine was protected as an amide to provide compounds with increased lipophilicity. Subsequent studies demonstrated the importance of the acetamido unit (R 1 =C(O)CH 3 ) for potent anticonvulsant activity and showed that either a decrease (i.e., R 1 = C(O)H) or increases (i.e., 3 ) in the size of this moiety led to reduced activity. Furthermore, when the acetamido (CH 3 C(O)NH) unit in 5 was replaced with methyl, methoxy, hydroxyl, acetoxy, or halogen, the obtained compounds exhibited diminished anticonvulsant activity. Structural modifications also included replacing the C(2) unit with the corresponding N(2) group giving the structurallyrelated semicarbazide derivatives 6 [57]. Evaluation of aza analogues 6 of functionalized amino acids in both mice (i.p.) and rats (p.o.) showed that the compounds exhibited significant anticonvulsant activities but in most cases at levels lower than their amino acid counterparts. Comparison of a selected series of semicarbazides 6 with their FAA counterparts 5 showed that replacing the tetrahedral C(2) carbon in 5 with a trivalent N led to a reduction in pharmacological activity in most cases upon administration to mice (i.p.). It was found that oral administration of the N(2)-substituted semicarbazides to rats led to improved anticonvulsant activities. Of the investigated compounds, 1acetyl-4-benzyl-2-(thiazol-2-yl)-semicarbazide 11 displayed moderate-excellent activity in mice (MES i.p. ED 50 = 22 mg kg -1 , PI = 5.4) and excellent activity in rats (MES p.o. ED 50 = 6.2 mg kg -1 , Tox TD 50 > 250) which exceeded that of phenytoin ( Table 1).
Conformationally restricted analogues of anticonvulsant functionalized amino acids have also been investigated [56]. Four peptidomimetic compounds of parent FAA 9 such as 1,5-disubstituted tetrazole 12, 3-substitued 1-benzylpyrrolidin-2-one 13, proline 14, and (thio)hydantoins 15, 16 as well as peptidomimetic FAA derivatives have been evaluated (Fig. 7). No improvement in pharmacological activity was observed upon conformational constraint, however new important information on the SAR of FAAs was obtained. It was shown that FAAs(1)-alkylation did not reduce anticonvulsant activity while N(3)-alkylation led to appreciable activity loss. These studies also revealed that derivatives of hydantoin 15 and thiohydantoin 16, upon p.o.  blockade of Na + channels (35% blockage at -60mV), while the two FAAs (9 and (R)-10) had no significant effect on peak current at 100 µM. Two prototypical antiepileptic Na + channel blockers, phenytoin and lamotrigine, provided 48 and 53 % blockage at -60mV, respectively. These findings indicate that 15 expresses its anticonvulsant activity in part by acting on the Na + channel. The divergent SAR and the different electrophysiological findings for (thio)hydantoins and FAAs provided evidence that these two classes of compounds function by different mechanisms.
It was determined that FAK exhibit excellent anticonvulsant activities that approach those observed for their FAA   Conversion of the acetamido unit in 5 to an amino moiety provided amino acid amides (AAA, 8) that are likely to have increased water solubility compared with their FAA 5 counterparts [61]. It was demonstrated that AAA 8 are potent anticonvulsants, and that the terminal amino group is prone to metabolic change. Among the investigated compounds, 20 displayed in MES test (i.p. mice) a protection level of ED 50 = 36 mg kg -1 , PI = 2.0, (p.o. rats) ED 50 = 7.1 mg kg -1 , PI = 33. The AAA anticonvulsant activity was neither strongly influenced by the C(2) substituent nor by the degree of terminal amine substitution (Fig. 9). The mechanism of action of FAA remains elusive. Some amino acid derivatives have also been studied by Paruszewski and coworkers [62][63][64][65][66][67]. Amides of Nsubstituted natural and anatural amino acids containing benzylamide, 4-fluorobenzylamide, 4-methoxybenzylamide, 2-furfurylamide, phenylethylamide, 3-pirydylmethylamide, buthylamide, isobuthylamide, usoamylamide moiety as well as esters have been synthesized and evaluated according to the procedure of the ASP of NINDS. Among the tested compounds, benzylamide derivatives of β-Ala Recent studies have demonstrated that some picoline and nicotinic acid benzylamides substituted on the phenyl ring also possess anticonvulsant properties [65][66][67]. Of these, the most active was the picolinic acid fluorobenzylamide (Pic-FBZA 25). ED 50 of the most effective amide 25 was 14.7 mg kg -1 (MES), >50 mg kg -1 (scPTZ) and PI < 3.4 against MES (rats, i.p.) [66].
SAR studies of alanine derivatives suggested that the structure of this amino acid, especially of fragment N-C α -C ' , is responsible for its action [62]. It was also found that with an α-amino acid structure, and N-acyl-or N-alkyl group, a small substituent at C-α or an aromatic amide substituent appear to be the most useful for anticonvulsant activity [64]. The importance of the benzylamide group of anticonvulsant active amino acids was confirmed both by Kohn's and Paruszewski's research groups.

Analogs of γ-vinyl GABA
Vigabatrin (γ-vinyl GABA) is being proposed as an anticonvulsant agent with a mechanism reportedly based on an inhibitor of GABA aminotransferase [17]. Lee and coworkers have developed several analogs of vigabatrin as potential dual acting prodrugs which were covalently coupled with an amide bond of vigabatrin and GABA mimetic substances such as GABA, γ-vinyl GABA, valproic acid, isonipecotic acid, nipecotic acid and 2-pyrrolidinone [73,74]. Most of these compounds have shown moderate anticonvulsive activities. Among them, compounds 30 and 31 displayed the most potent anticonvulsive activity and a broader spectrum when compared to vigabatrin (Fig. 11).

N-Benzylamides of γ-hydroxybutyric Acid
Derivatives of α-substituted γ-amino-, γ-phthalimido-, γacetoxy-and γ-hydroxy butyric acid, such as acids, esters and amides, have been investigated as new potent anticonvulsants [75][76][77][78][79][80][81][82][83][84]. It has been shown that α-substitutes Nbenzylamides of γ-hydroxybutyric acid (GHB) are the most potent compounds in this group, possessing anticonvulsant activities in MES (i.p. in mice) screens. The most potent anticonvulsants were α-(benzylamino)-γ-hydroxybutyric acid N-benzylamide 32 and N-(2-chlorobenzylamide) 33 (Fig. 12); their ED 50 being respectively 63.0 and 54.0 mg kg -1 . In the MES screen, these compounds were less active than the commonly used anticonvulsants carbamazepine and phenytoin, but possessed higher activity than sodium valproate [79]. Biochemical tests have indicated that the active amides act as allosteric modulators of the γ-aminobutyric acid, GABA A complex, and have an affinity to voltage-sensitive calcium channel receptors.  SAR studies have enabled the definition of structural elements responsible for the anticonvulsant activity of these several series of compounds [83,84]. These features are as follows: the presence of the N-benzylamide fragment, a hydrophobic unit (aryl ring) as a distal binding site and a group which could act as an H-bond donor. It was also concluded that a hydroxyl group was necessary for MES activity, and more lipophylic compounds showed better anticonvulsant properties [82].

2,6-dimethylanilides, Carboxamides
The activity of several 2-piperidinecarboxyamides in the MES test in mice has been reported [85,86]. Receptor binding studies indicate that these amides demonstrated weak binding affinity at the phencyclidine (PCP) site on the N-methyl-D-aspartate (NMDA) receptor complex; however, a correlation between affinity and seizure protection in the MES test was not observed. Using N-(2,6-dimethyl)phenyl-2-piperidinecarboxyamide 35 and N(α-methylbenzyl)-2piperidinecarboxamide 36 as structural leads (Fig. 13), a variety of analogues have been synthesized and evaluated for anticonvulsant activity in the MES test in mice [87]. The following modifications led to an increase in MES activity: replacement of the piperidine ring with pyridine and movement of the carboxamide group to the 4-position, then opening the piperidine ring. The 2,6-dimethylanilides were the most potent compounds in the MES test in each group of compounds (Fig. 14). The 4-pyridinecarboxamide 37 (ED 50 Fig. (13). Structures of 2-piperidinecarboxamides with activity against MES in mice.

Hydroxyamides
Brown and coworkers have evaluated a series of novel hydroxyamides [88,89]. Anticonvulsant testing of these compounds revealed the lead, 3,3,3-trifluoro-2-hydroxy-2phenyl-propionamide 39, to have potent anticonvulsant activity. Anticonvulsant evaluation of compound 39 administered i.p. in mice demonstrated complete (3/3 mice) protection at a dosage of 100 mg kg -1 up to 4 h when challenged with MES. Compound 39 also demonstrated effectiveness against scPTZ (5/5 mice protected) for 0.5 h at the same dosage. In rats p.o. test compound 39 protected: the MES ED 50 value was found to be 9.9 mg kg -1 , the scPTZ ED 50 was 34 mg kg -1 and the TD 50 noted at 100 mg kg -1 , yielding a therapeutic index of 10 for the MES model and 2.6 for the scPTZ model. In this new series of compounds, two, 40 and 41, were the most active. Analogue 41's MES ED 50 was also determined to be 62.4 mg kg -1 but has notably longer-lasting effects (up to 6 h) and a TD 50 over 100 mg kg -1 . Patch clamp electrophysiology studies demonstrated significant tonic blockade of T-type calcium current by compounds 39-41 at 1mM. Furthermore, compounds 39 and 40 induced a significant use-dependent blockade of T-type calcium current. These results suggest that the mechanism of anticonvulsant activity may include blockade of T-type calcium currents. Compounds 40 and 41 are methylated versions of compound 39 at the amide and alcohol, respectively (Fig. 15). Summing up, compound 39 is an active orally available anticonvulsant with similar activity to phenytoin, and its methylated alcohol and amide have shown similar activity. Fig. (15). Derivatives of hydroxyamides.

HETEROCYCLIC AGENTS (Arylalkyl)imidazoles
One of the structurally distinct classes of antiepileptic drugs is the (arylalkyl)imidazoles. Denzimol (+/-)-N-[β−[4(β-phenylethyl)phenyl]-β-hydroxyethyl]imidazole and nafimidone (1-[2-naphthoylmethyl)imidazole are examples of a class of anticonvulsants; the (arylalkyl)imidazoles (Fig.  17) [10]. SAR studies show that anticonvulsant properties of this group are associated with the presence of a small oxygen functional group (such as carbonyl, ethylene dioxy, methoxy, acyloxy and hydroxyl substituents) in the alkylene bridge in addition to an imidazole ring and a lipophilic aryl portion facilitating penetration of the blood barrier. The introduction of oxime and oxime ether groups to the alkylene bridge of (arylalkyl)imidazole as a small oxygen functional group led to new compounds, which displayed various levels of  Fig. (17). Structures of (arylalkyl)imidazoles antiepileptic drugs. activity [91]. Nafimidone oxime did not exhibit any anticonvulsant activity. O-Alkylation of nafimidone oxime, i.e. addition of an oxime ether functional group to the alkylene bridge of nafimidone, resulted in new compounds possessing anticonvulsant properties. The size of the alkyl moiety on the oxime group appears to be important for these properties. The O-alkyl substituted compounds (44)(45) were found to be more active than the O-arylalkyl substituted compounds (Fig. 18). Anticonvulsant activity against MESinduced convulsion showed that compound 44 E (trans) isomer, was the most active with ED 50

Pyrrolidin-2,5-diones
Ethosuximide, a derivative of pyrrolidin-2,5-dione, belong to a group of old antiepileptic drugs and is still used in the treatment of epilepsy.
SAR studies in this group of pyrrolidin-2,5-dione derivatives have led to a conclusion that the following structural elements are required for anticonvulsant activity: an aromatic ring at the 3-position of pyrrolidine-2,5-dione moiety and a 4-arylpiperazine fragment with selected substituents at the phenyl ring. The introduction of a spirocyclopentyl ring at the 3-position of pyrrolidine-2,5dione did not enhance the anticonvulsant activity.
Derivatives of (aryloxy)aryl semicarbazones displayed greater protection in the MES test than the scPTZ screen [103]. Quantification of approximately one-third of over 100 compounds tested revealed an average protection index of approximately 9. Following oral administration to rats, a number of compounds displayed significant potencies in the MES screen (ED 50 value of 1-5 mg kg -1 ) accompanied by a very high protection index. Later studies enabled the selection, in the series of aryloxyaryl semicarbazones, of a lead molecule 52 (Fig. 21). ED 50 values of compound 52 in the MES and scPTZ screens (i.p. in mice) were 48.6 and 94.1 mg kg -1 , with PI figures of 4.20 and 2.17 respectively [104]. Recently, compound 53 [4-(6-chlorobenzothiazol-2-yl)-1-(3isatinimino)thiosemicarbazone] has also shown strong activity in MES seizures and scPTZ screens [111]. Knowing that isatine derivatives possess anticonvulsant properties [113,114], compound 53 was designed as hybrid molecule incorporating a thiosemicarbazone fragment and an isatine molecule. Compound 53 was more potent than valproate against MES and scPTZ tests and was more effective than ethosuximide against MES-induced seizures, with an ED 50 of 17.86 and 6.07 mg kg -1 i.p. in mice and rat p.o., respectively. Compound 53 exhibited a PI value of 8.77 in the rat p.o.
identification in the MES screen. In the preliminary hippocampal-kindling screen in rats (i.p.), compound 53 showed weak ability to block the expression of fully kindled seizures. The potency and spectrum of activity of these 6chlorobenzothiazolyl thiosemicarbazones were comparable to those of standard drugs, and represent a structurally novel class for subsequent molecular modifications.

Thiadiazoles and Quinazolin-4(3H)-ones
New thiadiazolyl and thiazolidinonyl quinazolin-4(3H)ones have been synthesized and screened for their anticonvulsant activity comparing with the standard AEDs [115]. These hybrid molecules comprise two fragments, quinazolinone and thiazolidinone, whose derivatives have been found to show anticonvulsant properties [116]. Out of the 30 new hybrid compounds, the most active was 54 (Fig.  22). Compound 54 at three graded doses of 7.5, 15 and 30 mg kg -1 (i.p. in mice) in MES models was found to possess more potency than phenytoin and to be equipotent to lamotrigine at all the three dosage levels. In the scPTZ model, compound 54 exhibited better anticonvulsant activity than sodium valproate at all the three tested dosages (18.5, 37 and 74 mg kg -1 ). SAR studies have indicated that compounds having a 3-amino-2-methyl-6-bromoquinazolin-4(3H)-aryl moiety showed more protection than compounds with a 3-amino-2-methyl-quinazolin-4(3H)-aryl moiety.
Another group of new hybrid molecules represent derivatives of 1,3,4-thiadiazoles has been obtained by adding two heterocyclic nuclei possessing anticonvulsant activity, namely barbituric acid and quinazolinone (Fig. 22) [117]. Of the compounds studied, the most active one (55) displayed activity against MES and scPTZ seizure test in mice (i.p.) that was more potent than the standard drug. Compound 55 at three graded dosages of 17.5, 25 and 50 mg kg -1 has been found to incur 20, 40 and 90% protection in the MES model and 20, 50 and 90% protection against seizure in the scPTZ model (respectively).
In a series of alkanolamides and alkanoamines [120,121] [121]. Recent studies of some central effects of chiral xanthone derivatives demonstrated differing potencies of enantiomers and a racemic mixture of compound 65 [123].

Enaminones
Scott and coworkers have synthesized a series of enaminones bearing the aniline, benzylamine moieties and various aromatic heterocycles such as the pyridine, phenothiazines, and currently the isoxazole nucleus [127][128][129][130][131][132][133][134]. The prototype compound in the aniline series of enaminone was methyl 4-[(4-chlorophenyl)amino]-6-methyl-2-oxo-3-cyclohexene-1-carboxylate (67). Their ethyl analogue (68) was found to be the most promising enaminone in the aniline series in both rat and mouse models during preliminary pharmacological evaluation (Fig. 26). In MES model seizures, in mice, i.p. administration of 68 resulted in an ED 50 value of 16.7 mg kg -1 and TD 50 of 110.7 mg kg -1 , PI = 6.6 and an ED 50 value of 3.0 mg kg -1 and TD 50 > 250 mg kg -1 , PI = 83.3 p.o. administration in rats.  Compound 68 was more potent than phenytoin, however it was proposed that the enaminones possessed an anticonvulsant profile similar to phenytoin. Compound 68 was then evaluated for sodium channel binding activity. It provided a statistically significant block (P< 0.05) at -60mV, however the blockade was not strong enough to conclude that this was its principal mechanism of action. A biotransformation pathway for enaminones was also investigated [134]. The pharmacokinetic evaluation of 68 indicated that the carboalkoxy substituent converted in a two step reaction involving deesterification and decarboxylation to a second active enaminone metabolite 69 (Fig. 27). The 5methyl ketone (69)   The anticonvulsant activity of a series of enaminonederived isoxazoles provided a potent, orally active class of compounds [131]. Two series of enaminones possessing the isomeric 5-methyl-substituted isoxazoles and 3-methylsubstituted isoxazoles revealed both similarities and differences with regard to anticonvulsant activity (Fig. 28). The most potent anti-MES were three compounds in the 3amino series: ethyl ester 70 (ethyl 4-(5-methyl-3-isoxazolyl) amino]-6-methyl-2-oxo-3-cyclohexene-1-carboxylate), ED 50 68.9 mg kg -1 , p.o. in rats TD 50 > 500 mg kg -1 , PI > 49.6; methyl ester (71), ED 50 68.9 mg kg -1 i.p. in mice, TD 50 > 500 mg kg -1 , PI > 7.3, and tert-buthyl ester (72), ED 50 28.1 mg kg -1 p.o. in rats, TD 50 > 500 mg kg -1 , PI > 17.8. Sodium channel binding studies, as well as evaluations against pentylenetetrazol, bicuculline, and picrotoxin on izoxazole 70 were all negative, leading to an unknown mechanism of action. X-ray diffraction patterns of a representative of the 3amino series of isoxazoles displayed the existence of intramolecular hydrogen bonding of nitrogen to the vinylic proton in the cyclohexane ring, providing a pseudo-three ring  structure which was also detected previously in vinylic benzamides. According to the authors, the activity-inactivity of the 3-amino and 5-amino, isoxazoles may be due to the orientation of the pseudo-three ring system.

SUMMARY
The presented review summarizes ongoing medicinal chemistry investigations in search for new anticonvulsant compounds. Many of the compounds presented in this review have been evaluated by the ADD Program. Their anticonvulsant activity has been confirmed through in vivo screening tests, although for many compounds the precise mechanism of action is still not known. Some of the newer anticonvulsant agents represent structural modifications of pre-existing compounds, while others have been developed with the specific objective of modifying targets. These new agents belong to several different chemical classes. Some of them represent compounds bearing five-membered or other heterocyclic rings in their structure; additionally, numerous studies have demonstrated that derivatives of amino acids can function as potential new anticonvulsant agents. The most common structural elements of many active compounds are an amide bond (particularly a benzylamide group) and the presence of at least one aryl unit. New data has also confirmed that the lipophilicity of new active molecules is an important factor affecting their anticonvulsant potency. These new agents can be used for the design of future targets and development of new drugs. The discovery of a number of active leads may also ultimately help elucidate the mechanism of action of these new anticonvulsants.