Current Topics in Medicinal Chemistry 2002, 2, 817-832 817

GABAA Ligands and their Therapeutic Potentials

Bente Frølund,a,* Bjarke Ebert,b Uffe Kristiansen,c Tommy Liljeforsa and Povl Krogsgaard- Larsena aDepartments of Medicinal Chemistry and cPharmacology The Royal Danish School of Pharmacy, DK 2100 Copenhagen, Denmark and bDepartment of Molecular Pharmacology, H. Lundbeck A/S, 9 Ottiliavej, DK-2500 Valby, Denmark.

Abstract: The GABAA receptor system is implicated in a number of neurological diseases, making GABAA receptor ligands interesting as potential therapeutic agents. Only a few different classes of structures are currently known as ligands for the GABA recognition site on the GABAA receptor complex, reflecting the very strict structural requirements for GABAA receptor recognition and activation. Within the series of compounds showing agonist activity at the GABAA receptor site that have been developed, most of the ligands are structurally derived from the GABAA agonists , THIP or . Using recombinant GABAA receptors, functional selectivity has been shown for a number of compounds such as the GABAA agonists -4-acetic acid and THIP, showing highly subunit-dependent potency and maximal response. In the light of the interest in partial GABAA receptor agonists as potential therapeutics, structure-activity studies of a number of analogues of 4-PIOL, a low-efficacy partial GABAA agonist, have been performed. In this connection, a series of GABAA ligands has been developed showing pharmacological profiles from moderately potent low-efficacy partial GABAA agonist activity to potent and selective antagonist effect. Only little information about direct acting GABAA receptor agonists in clinical studies is available. Results from clinical studies on the effect of the GABAA agonist THIP on human sleep pattern shows that the functional consequences of a direct acting agonist are different from those seen after administration of GABAA receptor modulators.

Keywords. GABAA receptor, partial agonist, functional selectivity, THIP, 4-PIOL, IAA, muscimol.

INTRODUCTION bind to a group of GABA receptors which was insensitive to both BMC and [6, 8]. These receptors were named 4-Aminobutyric acid (GABA), the major inhibitory GABAC receptors or non-GABAA, non-GABAB (NANB) neurotransmitter in the central nervous system (CNS) receptors for GABA [9, 10]. operates through two different classes of receptors consisting of the ionotropic GABAA and GABAC receptors and the The pharmacological characterization of the GABAA metabotropic GABAB receptors. This receptor classification receptors was accelerated by the development of specific has been under development over a period of 40 years, GABAA agonists such as isoguvacine, 4,5,6,7-tetrahydro- starting in the early 1950s and is primarily based on the isoxazolo[5,4-c]pyridin-3-ol (THIP) and piperidin-4- selective pharmacology observed for the ligands depicted in sulphonic acid (P4S) [11, 12]. Following the observation Fig. (1). The identification of the alkaloid [1] that the binding site for the (BZDs) was and its quaternized analogue bicuculline methochloride associated with the GABAA receptors the exploration of the (BMC) [2] as competitive GABA antagonists in CNS GABAA receptors was dramatically intensified [13-17]. After tissues initiated the pharmacological characterization of cloning of a large number of GABAA subunits, this area of GABA receptors. The observation that the antispastic effect the pharmacology is developing rapidly. of baclofen, which was designed as a lipophilic GABA analogue, could not be antagonized by BMC led to the A number of compounds with selective action at the recognition of the GABAB receptors as a distinct class [3-5]. GABAB receptors, have been described [18, 19]. The number Subsequently, in connection with the design of of selective GABAC receptor ligands is still limited, making conformationally restricted analogues of GABA, cis- this receptor group the least characterized of the GABA aminobut-2-enoic acid (CACA) was developed [6]. CACA receptors [10]. and the structurally related analogue, cis-2-aminomethyl- cyclopropanecarboxylic acid (CAMP) [7], were shown to

THE GABAA RECEPTOR COMPLEX

*Address correspondence to this author at the Department of Medicinal The GABAA receptor is a member of the superfamily of Chemistry, The Royal Danish School of Pharmacy, 2 Universitetsparken, ligand-gated ion channels. The structure and function of this DK-2100 Copenhagen, Denmark; Tel: (+45) 35306495; Fax: (+45) 35306040; E-mail: [email protected] group of receptors display a high degree of complexity, of

1568-0266/02 $35.00+.00 © 2002 Bentham Science Publishers Ltd. 818 Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 Frølund et al.

OH

H2N OH OH O HO O OH OH P CH HN HN 3 O N O O O H2 N H N Cl 2 H2N

Isoguvacine THIP (R)-Baclofen CGP-35024 CACA CAMP

GABA GABAA GABAB C

OCH3 O OH HO O O CH3 O N P OH P P OH O CH H N H 3 2 O H CH3 O HN O H2N OH N

O N O Cl O O NH

BMC SR 95531 (R)- CGP-35348 TPMPA

Fig. (1). Schematic illustration of the different classes of GABA receptors and the structures of some key agonists (upper part) and antagonists (lower part) used for pharmacological characterization of these receptors. which only a few aspects of particular relevance to the development of GABAA receptor ligands will be mentioned.

The postsynaptic GABAA receptor is a receptor complex containing a considerable number of modulatory binding sites, as reflected by the structural diversity of compounds acting on GABAA receptors, including important drugs such as benzodiazepines, , and (Fig. 2) [10].

The heterogeneity of GABAA receptors in the brain is pronounced due to the large number of different GABAA receptor subunits. At least 17 different subunits (a 1–6, b1–4, g1–4, d, e, and p) have been identified [20, 21]. GABAA receptors are built up as pentameric assemblies of different families of receptor subunits, making the existence of a very large number of heteromeric GABAA receptors possible [22, 23]. The assembly, which in most receptors includes two a subunits, two b subunits and one g or d subunit, determines the pharmacology of the functional receptor [9, 21]. Site- directed mutagenesis studies have shown that the binding site for benzodiazepines is located at the interface between a and g subunits in the GABAA receptor complex, whereas Fig. (2). (A) Schematic model of the pentameric structure of the the binding site(s) for GABA, GABAA agonists and GABAA receptor complex and (B) a schematic illustration of the competitive antagonists are located at the interface between GABAA receptor complex indicating the chloride ion channel a and b subunits [24]. and additional binding sites.

The number of physiologically relevant GABAA that approximately twenty different subunit combination receptors, their subunit combinations, and their regional make up the major subtypes of GABAA receptors, with the distributions are far from being fully mapped out and studies a 1b2/3g2 receptor combination as the most abundant [27, in this area are in rapid progress [20, 25, 26]. It is estimated 28]. To further elucidate the physiological importance of the GABAA Receptor Ligands and their Therapeutic Potentials Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 819 various GABAA receptor subunits and their preferred although BMC has recently been shown to lack the combinations, GABAA ligands with subunit specific effects selectivity of bicuculline as a GABAA antagonist [33]. are needed as important tools. New structural classes of GABAA receptor antagonists This review will focus exclusively on ligands for the have been developed (Fig. 3). A series of arylaminopyrida- GABA recognition site on the GABAA receptor complex. zine analogues of GABA, notably SR 95531, show potent and selective competitive GABAA antagonist effects [34, 35]. These compounds bind tightly to GABAA receptor GABAA RECEPTOR LIGANDS sites, and tritiated SR 95531 is now used as a standard receptor ligand. Although SR 95531 and related compounds Only very few naturally occurring compounds have been containing a GABA structure element show convulsant effect identified as GABAA receptor agonists. Among these the after systemic administration [36], these zwitterionic most important one is the potent GABAA agonist compounds do not easily penetrate the BBB. Compound 3, muscimol, a constituent of the mushroom Amanita in which the GABA structure element has been replaced by a muscaria (Fig. 3) [29]. The histamine metabolite imidazole- unit (see Fig. 4), is the most potent GABAA 4-acetic acid (IAA) represents another structural class of antagonist in the arylaminopyridazine series [36]. The ligands and has been shown to be a relatively potent bicyclic 5-isoxazolol compound, iso-THAZ, derived from GABAA agonist and a GABAC antagonist [30]. IAA readily THIP, is a moderately potent GABAA antagonist [37]. penetrates the blood-brain barrier (BBB) and may play a role as a central and/or peripheral endogenous GABAA receptor ligand. Like the imidazole group of IAA, the structurally DEVELOPMENT OF GABAA RECEPTOR related isothiouronium element of (RS)-2-amino-2-thiazoline- AGONISTS 4-acetic acid (1) is effectively recognized by the GABAA receptors [31]. (1S,3S)-3-Aminocyclopentane-1-carboxylic In order to determine the structural requirements for acid (2), a GABA analogue containing a conformationally activation of the GABAA receptor a number of structural restricted carbon backbone, is also a specific GABAA modifications of the molecule of GABA have been made, agonist [32]. some of which are shown in Fig. (4). Conformational restriction of various parts of the molecule and isosteric The classical GABAA antagonist bicuculline and BMC replacements of the carboxyl group have led to a broad have played a key role in studies on GABAA receptors, spectrum of specific GABAA agonists. Some of these

A OH OH OH OH

O O O H2 N N O N NH S N

NH2 H2N

Muscimol IAA 1 2

B O OCH3 O O

N O CH3 H H OH O OH N N N O N N O S O O NH NH

Bicuculline SR 95531 3

OH HN

O N

Iso-THAZ

Fig. (3). Structures of (A) GABAA receptor agonists and (B) antagonists. 820 Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 Frølund et al. molecules have played a key role in the development of the represents the bioactive conformation of muscimol at the pharmacology of the GABAA receptor. GABAA receptor, and pharmacophore models have been proposed based on this assumption [44, 45]. Ab initio The fact that muscimol is a non-specific GABAA quantum chemical calculations on the conformational receptor agonist [38, 39], a substrate for the GABA- properties of muscimol indicate that in order for muscimol metabolizing enzyme, GABA transaminase [40], and to mimic the conformation of THIP, a very high moreover a neurotoxin, makes the compound therapeutically conformational energy (8.9 kcal/mol) is required [46]. Thus, less valuable. The 3-isoxazolol moiety of muscimol has, it is highly improbable that the bioactive conformation of however, proved to be very efficient as a bioisostere for the muscimol corresponds to that of THIP. carboxyl group in GABA. Therefore, muscimol has been extensively used as a lead structure in the search for potent The 3-isoxazolol carboxyl group isostere of muscimol and selective compounds acting at GABAA receptors as can be replaced by a 3-isothiazolol unit to give illustrated in Fig. (4). Further conformational restriction of thiomuscimol without significant loss of GABAA receptor the GABA structural element in muscimol has been achieved activity [38], whereas the corresponding sulfur analogue of by incorporating the amino group into a ring THIP, thio-THIP, displays more than 300 times lower leading to the bicyclic analogue, THIP, a specific GABAA affinity than THIP itself [47]. These results are in agreement agonist [11]. THIP has been shown to be devoid of the with the conformational studies mentioned above, and neurotoxic properties of muscimol and, in contrast to strongly indicate that muscimol and THIP bind to the muscimol, is metabolically stable. THIP shows non- GABAA receptor in different binding modes with respect to analgesic effect [41] and is in clinical trials as an atypical the 3-isoxazolol moiety. In light of the structural similarity agent [42]. In contrast to GABA and in spite of of THIP and thio-THIP the markedly different pharmacology their predominantly zwitterionic structure, muscimol and of these compounds suggests that the sulphur atom, in THIP are both capable of penetrating the BBB [43]. contrast to the sulphur atom in thiomuscimol, directly interferes with the receptor, where it appears to cause adverse The similarity in the pharmacological profile of the two steric or electronic interactions. compounds led to the assumption that the structure of THIP

GABAA AGONISTS GABAC ANTAGONISTS

OH OH OH OH OH H N HN 2 O HN HN H2N N N N N S O S N H GABA Thiomuscimol THIP Thio-THIP Aza-THIP

OH OH OH OH OH HN HN HN H N O O 2 HN N O N O N O

Muscimol Isonipecotic acid Isoguvacine Iso-THIP THAZ

INACTIVE

CH3 OH OH OH OH

HN HN HN N N N HN N CH3 O N N O

CH3 4-Me-THIP 1-Me-Aza-THIP 2-Me-Aza-THIP THIA

Fig. (4). Comparison of the structures of some GABAA agonists, GABAC antagonists and some inactive THIP analogues. GABAA Receptor Ligands and their Therapeutic Potentials Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 821 In support of these conclusions, docking of THIP into a corresponding saturated cyclic amino acid isonipecotic acid recently proposed pharmacophore model for GABAA as a GABAA agonist [11], DH-P4S is an order of magnitude agonists [48-50] indicates that on binding to the receptor, weaker than P4S as an agonist at the GABAA receptors [12, the 3-isoxazolol rings of muscimol and THIP do not overlap 56]. as displayed in Fig. (5). This pharmacophore model will be further discussed below. A series of isonipecotic acid analogues has been investigated electrophysiologically at heteromeric (a xb3g2) human GABAA receptors expressed in Xenopus oocytes [57]. Like GABA and isonipecotic acid, the respective seleninic acid isosteres, 4 and 5, and the phosphinic acid isostere, 6, of isonipecotic acid were characterized as GABAA agonists (Fig. 6). The agonist profile displayed by the bioisosteres was highly dependent on the receptor subunit combination as shown in Table 1. The agonist potencies were generally highest at a 6 containing receptors. Furthermore, the maximal response of the GABAA agonists relative to that of GABA is dependent on the type of a subunit present. The seleninic acid analogue, 4, of GABA is a partial agonist on all subunit combinations tested. Whereas P4S effectively activates a 2 and a 5 containing receptors it essentially blocks the a 4 and a 6 containing receptors. The seleninic acid and the phosphinic acid analogues of , 5 and 6, respectively, act as low-efficacy agonists at a 3 and a 6 containing receptors, but as competitive antagonists at other subunit combinations tested. Although the structural determinants for this apparent functional selectivity is yet not clarified, these findings open up the perspective for developing functionally selective compounds.

Fig. (5). Perspective drawing of muscimol and THIP zwitterions In contrast, binding studies have revealed that only small (upper part) and a superimposition of the two compounds and differences in binding affinities for agonists can be observed their interactions with an arginine residue (lower part). using recombinant GABAA receptors containing different a or b subunits [58]. This indicates that, whereas the GABAA Like thio-THIP, the structurally related THIP analogues, receptor mechanisms transducing binding into a aza-THIP, iso-THIP and THAZ (Fig. 4) are virtually inactive physiological response are highly dependent of the receptor at the GABAA receptors [51, 52], but disclose antagonistic subunit composition, the binding step of receptor ligand effects at the GABAC receptors [53]. Introduction of steric interaction may play only a minor role in the subunit bulk in different positions of the molecule of THIP and aza- dependency of GABA ligand potency. THIP, as shown in Fig. (4), leading to 4-Me-THIP [54], 1- Me-aza-THIP, 2-Me-aza-THIP [52] and the seven-membered Conversion of the phosphinic acid, 6, into the ring analogue, THIA [29, 55] has been detrimental to methylphosphinic acid, 7, the phosphonic acid, 8, and the agonist activity, resulting in inactive compounds. These hydroxy substituted phosphinic acid, 9, resulted in GABAA results emphasize the very strict structural requirements of antagonists [57]. As shown in Table 1, the potencies of GABAA receptors for recognition and activation. these antagonists are essentially independent of the subunit composition of the GABAA receptor. This is in line with previous findings for the competitive GABAA antagonist MONOCYCLIC GABAA RECEPTOR LIGANDS SR 95531 and bicuculline [58]. DERIVED FROM THIP Besides having effect on the GABAA receptors, the Using THIP as a lead, a series of specific phosphinic, phosphonic and the seleninic acid analogues, 4– monoheterocyclic GABAA agonists, including isoguvacine 9, all have been characterised in human GABAC r 1 receptors and isonipecotic acid, was developed [11, 52]. THIP and the expressed in oocytes as weak to potent GABAC antagonists, equally potent and specific amino acid analogue isoguvacine where the seleninic acid analogues, 4 and 5, were the most are now standard GABAA agonists. potent ones [59]. Furthermore, the seleninic acid analogue of GABA, 4, is a potent GABAB agonist [60]. As illustrated in Fig. (6), the carboxyl group of isoguvacine and isonipecotic acid has been replaced by a number of different isosteric groups leading to compounds PARTIAL GABA AGONISTS with marked variation in pharmacological profile. A Therapeutic use of full GABA agonists or antagonists As P4S, the unsaturated analogue, DH-P4S, is a A may be associated with severe side-effects. Full GABAA GABAA agonist. It is interesting to note that although agonists may induce desensitization of the target receptors, isoguvacine is an order of magnitude more potent than the which can lead to tolerance and subsequent withdrawal 822 Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 Frølund et al.

GABAA AGONISTS / PARTIAL AGONISTS

OH OH Se H N H N 2 O 2 O

GABA 4

OH OH OH OH OH OH P O S O S O Se HN HN HN HN HN HN H O O O O O

Isonipecotic acid Isoguvacine P4S DH-P4S 5 6

GABAA ANTAGONISTS

OH OH OH P O P O P OH H HN HN HN OH CH3 O

7 8 9

Fig. (6). Structures of some GABA and isoguvacine analogues. symptoms, whereas GABAA antagonists are potential subunit dependence of the maximal response of THIP and anxiogenics and proconvulsants. This has focused interest on IAA are qualitatively the same as that of P4S. An example partial GABAA receptor agonists as potential therapeutics. illustrating the complete unpredictability of the functional consequences of modifications of subunit compositions Under clinical conditions where stimulation of the relates to the GABAA receptors in the cerebellar granule GABAA receptor system may be relevant partial agonists cells. Cerebellar granule cells contain a 1 b2 g2, a 6 b2 g2 and displaying a relatively high efficacy may be useful. The level a 1 a 6 b2g2 containing receptors. A comparison of the of efficacy needed may be dependent on the particular activity of P4S at these three different combinations reveals disease. The potent analgesic effects of THIP seem to that at a 1 b2 g2 and a 6b2 g2 containing receptors P4S is a indicate that this relatively high level of efficacy [61, 62] is partial agonist with a maximal response of 38 % and 15 %, close to optimal with respect to treatment of pain. respectively, of that of GABA, whereas at a 1a 6 b2 g2 containing receptors, the maximum response to P4S is 75% Analogously, very low-efficacy GABAA agonists of that of GABA and the potency is different from that at the showing predominantly antagonist profiles may have clinical other receptor combinations [64]. This result is not only interest in conditions where a reduction in GABA receptor A surprising, it is also in contrast to data for P4S at a 1 activity may be needed. Such compounds may produce /a /a a or a /a /a a containing receptors, where co- sufficient GABA receptor activity to avoid the above 5 1 5 1 3 1 3 A expression of two different a subunits resulted in an a 1-like mentioned side effects. maximum response and a significantly lower potency than at a a a a As observed for THIP, the heterocyclic GABA isosteres 1 / 5 or 1 / 3 containing receptors [65]. In general coexpression of different a subunits results in novel IAA and P4S show the characteristics of a partial GABAA agonist [61, 63]. As described for the analogues of pharmacology [66], underlining the complexity of the system. isoguvacine, it has been shown, using GABAA abg receptors cloned in oocytes, that the relative efficacy as well as the potency for these compounds is highly dependent on The non-fused THIP analogue, 5-(4-piperidyl)isoxazol-3- the receptor subunit combination (Table 1) [57]. As ol (4-PIOL) (Fig. 8), is a low affinity GABAA agonist, showing no significant stimulatory effects on BZD binding illustrated in Fig. (7), the pattern of GABAA receptor GABAA Receptor Ligands and their Therapeutic Potentials Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 823

Table 1. Agonist Activity and Antagonist Potencies as Function of GABAA Receptor Subunit Composition

a GABAA Agonists / Partial Agonists, EC50 in µM (Max in %)

a 1b3g2 a 2b3g2 a 3b3g2 a 4b3g2 a 5b3g2 a 6b3g2

GABA 80 (100) 40 (100) 28 (100) 50 (100) 3.0 (100) 1.5 (100)

4 1300 (30) 350 (47) 160 (33) 135 (16) 1300 (58) 213 (51)

Isoguvacine 160 (88) 50 (88) 55 (69) 180 (88) 75 (94) 20 (98)

Isonipecotic acid 620 (46) 420 (53) 370 (46) 3000 (104) 690 (57) 160 (83)

P4S 40 (21) 26 (96) 65 (75) 200 (7.2) 20 (92) 20 (15)

5 200 (3) 170 (2) 180 (28) 160 (0) 420 (3) 240 (13)

6 270 (5) 280 (17) 160 (29) 140 (0) 260 (17) 140 (35)

b GABAA Antagonist, K i in µM

7 310 600 300 340 420 460

8 3300 2100 3800 3100 2800 4800

9 1400 1500 4900 2100 1900 3000 a Agonists were characterized at human GABAA receptors containing axb3g2 (x=1–6), expressed in Xenopus oocytes. EC50 values are in µM, and dose/response curves were fitted. The maximun responses (Max), relative to that of GABA at the same oocyte, were calculated as described in [57]. Values presented are mean values. SEM values for maximum responses were less than 10% and relative SEM of pEC50 values were less than 10%. Modified version of table from [57]. b Antagonist potencies (Ki values in µM) were determined at human GABAA receptors containing axb3g2 (x=1–6), expressed in Xenopus oocytes by parallel shift of GABA dose/response curves in the presence of a fixed concentration of antagonist. Values presented are mean values. Relative SEM of pKi values were less than 10%. Modified version of table from [57].

Fig. (7). Subunit dependent efficacy of isoguvacine, THIP, P4S and IAA. Oocytes were injected with cDNA encoding for a xb3g2S and concentration response curves were constructed for the compounds. Maximum response at every receptor combination at every oocyte was determined using 3mM GABA [57].

[67]. On the other hand, 4-PIOL dose-dependently showing a predominant GABAA antagonist profile, being antagonized muscimol induced stimulation of BZD binding about 30-fold weaker than BMC as an antagonist at the in a manner similar to that of the GABAA antagonists BMC GABAA receptors. In cortical neurones, the agonist potency [68]. 4-PIOL has been characterised as a partial GABAA of 4-PIOL has been shown to be approximately equivalent to agonist using functional patch-clamp techniques on cultured that of isoguvacine and GABA, where the maximum cerebral cortical and hippocampal neurones [69, 70] as well response of 4-PIOL only was a small fraction of those as on recombinant GABAA receptors expressed in oocytes observed by administration of submaximal concentrations of [71]. isoguvacine [64]. Importantly, repeated administration of 4- PIOL did not cause significant desensitisation of the In cortical neurones, the action of 4-PIOL was compared GABAA receptors [64, 69]. The high potency found for 4- with those of the GABAA agonist isoguvacine and the PIOL in cortical neurones is in accordance with findings in GABAA antagonist BMC [70]. Based on these studies, it is recombinant GABAA receptors and is surprising in light of concluded that 4-PIOL is a low-efficacy partial agonist the low affinity for the GABAA receptors shown in rat brain 824 Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 Frølund et al. homogenate [70]. A possible explanation may be that the receptor binding affinity of 4-PIOL is determined on non- desensitized receptors, whereas the affinity of the highly efficacious agonists are measured on desensitized receptors [64].

Recently, 4-PIOL has been proposed to possess some subtype specific characteristics [73]. On recombinant GABAA receptors of the a 1b2g2 subtype expressed in HEK- 293 cells, 4-PIOL acted as a weak agonist, whereas it was devoid of activity in the a 6b2g2 receptor subtype. In order to determine the structural determinants for this unique pharmacological profile, a series of analogues of 4- PIOL, shown in Fig.(8), has been synthesized [70], where the 3-isoxazolol was replaced by a 3-isothiazolol or a 2- isoxazolin-5-one ring system to give thio-4-PIOL and iso-4- PIOL, respectively. The two isosterically derived analogues of 4-PIOL and the unsaturated analogues DH-4-PIOL and DH-thio-4-PIOL showed qualitatively similar effects on Fig. (9). Efficacy estimates for the partial GABAA agonists 4- cultured cortical neurones as 4-PIOL [70]. The relative PIOL and analogues normalized to the full agonist isoguvacine efficacies of these compounds as partial GABAA agonists (2 mM) measured on cultured cortical neurons using whole-cell range from levels below that of 4-PIOL to significantly patch-clamp techniques [70, 64]. Insert: Efficacy estimates for higher levels. The compounds were compared with the effect agonists and partial agonists normalized to the full agonist of 20 mM isoguvacine corresponding to approximately 18 % GABA (2 mM) [64]. of the maximal effect of isoguvacine. Assuming that isoguvacine is a full agonist, the analogues have maximum responses ranging from 1% to 10% of that of the maximal effect of isoguvacine, where the rank order is: DH-thio-4- 4-PIOL ANALOGUES AS GABAA ANTAGONISTS PIOL>thio-4-PIOL>4-PIOL>DH-4-PIOL>iso-4-PIOL (Fig. Introduction of alkyl groups into the 4-position of the 3- 9). isoxazolol ring of muscimol and THIP severely inhibits These results are in accordance with results from studies interaction with the GABAA receptor recognition site as of a number of GABA ligands, including 4-PIOL and thio- illustrated in Fig. (10). Thus, 4-Me-muscimol is 3–4 orders A of magnitude weaker than muscimol as an inhibitor of 4-PIOL, on human a 1b3g2S GABAA receptors cloned in oocytes [57, 58, 72]. Potencies and relative efficacies of GABAA receptor binding [55], whereas 4-Et-muscimol [55] these compounds are shown in Table 2. and 4-Me-THIP [54] are inactive. In contrast, the GABAA recognition site tolerates introduction of alkyl groups into the 4-position of the 3-isoxazolol ring of 4-PIOL [49]. These In spite of the heterogeneity of GABAA receptor sub- types present in cultured neurones, pharmacological actions structure-activity relationships indicate that the binding determined in these systems generally are in agreement with modes of the GABAA agonists, muscimol and THIP and in data from recombinant receptors expressed in oocytes. particular of the low-efficacy partial agonist 4-PIOL are different.

OH OH

N N O S HN HN

4-PIOL Thio-4-PIOL

OH OH OH

N N O O S N HN HN HN

DH-4-PIOL DH-Thio-4-PIOL Iso-4-PIOL

Fig. (8). Structures of the low-efficacy partial GABAA agonist, 4-PIOL, and some analogues. GABAA Receptor Ligands and their Therapeutic Potentials Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 825

Table 2. EC50 and Maximum Response for GABAA Agonists / in order to further investigate the steric tolerance of this position [50]. The results from these studies are exemplified Partial Agonists at a 1b3g2 GABA Receptors Expressed in Oocytes [57]

a a EC50 (mM) maximum response in %

Isoguvacine 160 88±3

THIP 350b 70±3b

P4S 40 21±1

IAA 310 24±8

Thio-4-PIOL 16c 8±4c

4-PIOL 100c 3±2c Fig. (11). A pharmacophore model for GABAA receptor agonists a Determined by voltage-clamp on human GABAA receptors containing a1b3g2 showing the proposed binding modes of muscimol (dark grey expressed in Xenopus oocytes. The maximum responses, relative to that of GABA bonds) and 4-PIOL (light grey bonds) and their interactions at the same oocyte, were calculated as described in [57] with two different conformations of an arginine residue. The b Taken from [58] tetrahedrons indicate positions of methyl groups in GABAA cEbert et al. unpublished agonists, which cause strong steric repulsions with the receptor. The dark grey spheres indicate sites to which the ammonium There is strong evidence that an arginine residue at the group in muscimol interacts via hydrogen bonds. GABAA receptor recognition site is directly involved in the binding of the anionic part of the receptor ligand [74-76]. in Fig. (12). Substitution of the 4-position with alkyl or Based on this observation, a hypothesis has been proposed benzyl groups, as for compounds 10–12, resulted in affinity concerning the binding modes of the bioactive and potencies comparable with those of 4-PIOL. conformations of muscimol and 4-PIOL [49] as illustrated in Interestingly, introduction of more bulky groups such as Fig. (11). In these binding modes the two 3-isoxazolol rings diphenylalkyl and naphthylalkyl groups, as exemplified by do not overlap. This means that the 4-position of the 3- the 1-naphthylmethyl, the 2-naphthylmethyl and the 3,3- isoxazolol ring in muscimol does not correspond to the 4- diphenylpropyl analogues (13–15) are not only tolerated but position in the 3-isoxazolol ring of 4-PIOL during resulted in a marked increase both in affinity and potency, as interaction of muscimol and 4-PIOL with the receptor shown in Fig. (12). recognition site. Using whole-cell patch-clamp techniques on cultured A number of analogues of 4-PIOL have been synthesized cerebral cortical neurones in the electrophysiological testing, with substituents in the 4-position of the 3-isoxazolol ring the pharmacology of the 4-PIOL analogues in the absence or

OH OH OH OH OH

H N N HN HN 2 H2N N H2N N N N O O O O O

Receptor binding Muscimol 4-Me-Muscimol 4-Et-Muscimol THIP 4-Me-THIP (IC50) 0.006 mM 26 mM >100 mM 0.13 mM >100 mM

OH OH OH

N N N O O O HN HN HN

Receptor binding 4-PIOL 10 11 (Ki) 9 mM 10 mM 6 mM

Fig. (10). GABAA agonist binding data (IC50 or Ki, (µM)) for muscimol, THIP, 4-PIOL and some alkylated analogues of these GABAA receptor ligands. 826 Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 Frølund et al.

Fig. (12). (A) Receptor binding affinity (Ki values from [H]muscimol binding on rat brain synaptic membranes) and in vitro electrophysiological activity (IC50 values from whole-cell patch clamp recording from cerebral cortical neurons) for a number of 4- PIOL analogues. (B) Effect of the partial agonists or antagonists on the response to 20µM isoguvacine using whole-cell patch-clamp recordings from cultured cerebral cortical neurons. 20µM isoguvacine and varying concentrations of antagonist/partial agonist were applied simultaneously to the cells. The response to 20µM isoguvacine alone has been set to 100%, and the other responses are expressed relative to this [50]. in the presence of the specific GABAA receptor agonist The antagonistic effects of the very potent 4-PIOL isoguvacine was studied [50]. The results demonstrated that analogues, 13 and 15, were evaluated in mice after the structural modifications led to a change in the intracerebroventricular (i.c.v) administration [50]. Both pharmacological profile of the compounds from moderately compounds showed potent convulsant activity with ED50 = potent low-efficacy partial GABAA receptor agonist activity 0.024 µmol/kg and ED50 = 0.21 µmol/kg, respectively. to potent and selective antagonist effect. The 2- naphthylmethyl and the 3,3-diphenylpropyl analogues, 13 and 15, respectively, showed an antagonist potency comparable with or markedly higher than that of the standard GABAA antagonist SR 95531. Only the methyl and ethyl analogues, 10 and 11, retained detectable ability to induce an agonist effect shown on cultured cerebral neurones [50] or recombinant receptors [71]. In the study on recombinant GABAA receptors, the weak responses of 10 and 11 were potentiated by simultaneously administered and inhibited by the competitive antagonist SR 95531 [71].

These structure-activity studies seem to support the proposed hypothesis concerning the distinct binding mode of 4-PIOL, implying that the 4-position in 4-PIOL does not correspond to the 4-position in muscimol (Fig. 11). Thus, a cavity of considerable binding capacity seems to exist at the 4-PIOL recognition site of the GABAA receptor. Molecular modeling studies of the two high affinity compounds containing a 2-naphthylmethyl and a 3,3- diphenylpropyl substituent, 13 and 15, and the less active 1- naphthylmethyl analogue, 14, indicate that this proposed binding cavity may be exploited in two directions (Fig. 13). In both of these positions an aromatic ring seems to be Fig. (13). Proposed bioactive conformations for some high highly favourable for the receptor affinity[50]. affinity 4-substituted analogues of 4-PIOL. GABAA Receptor Ligands and their Therapeutic Potentials Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 827 Based on the reported high correlation between binding Quite interestingly, a compound like Tiagabin which acts data obtained in cell lines expressing human GABAA by increasing synaptic GABA via a blockade of primarily receptors [58] and rat brain homogenate [70], it has been the GAT-1 GABA transporter, is now widely used as an concluded, that the inter species homology in GABAA analgesic, illustrating that the hyper-polarizing effects of receptor sequence and structure is very high. Also, when Tiagabin has significant clinical impact. comparing functional antagonist potencies obtained from cloned human receptors with binding affinities determined in Although intellectually very tempting, only very little binding studies using rat brain homogenates, a high information on direct acting GABAA receptor agonists in correlation is observed. It must be expected that the clinical studies has been obtained. The structural heterogeneity of the GABA receptors in rat brain requirements for GABAA receptor agonists are very distinct homogenate is high. It is therefore concluded, in agreement and specific, and a general consensus on the risk of using with earlier observations for GABAA antagonists [57], that GABAA receptor agonists, which may activate every the potencies of these antagonists are essentially independent GABAA receptor subtype in the CNS, has existed. Since of the subunit composition of the receptors studied. amino acid residues contributing to the binding site appear to be conserved within the a and b subunit families [79], As mentioned earlier the GABA binding site in the the consequence is that the actual binding site is conserved GABAA receptor is assumed to be located at the interface independent of the receptor configuration. Therefore, receptor between a and b subunits [24]. It has been speculated that affinity appears to be independent of receptor subunit the GABAA antagonists bind to and stabilize a distinct combination [58]. However, the functional consequences of inactive receptor conformation. In case of the 4-arylalkyl receptor binding are highly dependent on the receptor substituted 4-PIOL analogues it may be speculated that the assembly [57, 58, 65]. Thus, as observed for large cavity accommodating the 4-substituent is located in benzodiazepines, only functional characterisation of ligands the space between these subunits. GABAA receptors belong may provide a basis for understanding the in vivo to the same superfamily as the nicotinic acetylcholine pharmacological consequences of drug administration. As receptors. It has been proposed that the mechanism for described above functional selectivity is obtainable for a ligand induced channel opening in nicotinic acetylcholine number of compounds like the GABA agonists IAA and receptors involves rotations of the subunits in the ligand THIP. IAA and THIP display a highly subunit dependent binding domain [77, 78]. Assuming that the GABAA potency and maximal response resulting in functional receptors utilize a similar mechanism for channel opening, selectivity. At some combinations, the compounds may act large substituents may interfere with the channel opening as agonists and, at the same time, as antagonists or low resulting in antagonistic effects of the compounds [50]. efficacy partial agonists at other combinations. Our understanding of the determinants of this functional selectivity is still very limited, and development of these CLINICAL POTENTIAL OF DIRECT ACTING compounds with predefined receptor subtype selectivity GABAA RECEPTOR AGONISTS profile is still based on trial and error. An important aspect of this problem is the selection of subunit combinations GABAA receptor dysfunction has been implied in a relevant for the prediction of in vivo activity. number of CNS diseases including sleep disturbances, anxiety, premenstrual syndrome, muscle spasms, epilepsy, An example of the consequences of variations of subunits Alzheimer’s disease, chronic pain and schizophrenia. is the effect of a substitution of the d subunit for a g2S Although the hypothesized dysfunction of the GABA A subunit in a receptor complex containing a 4b3. The a 4 receptor system is well described for all the above mentioned containing receptors exist predominantly in the thalamic area diseases and thus the basis for a potential intervention with [80]. Recent studies [81] have shown that some of these GABAA receptor ligands may be well founded, little is receptors may be located extrasynaptically, making them known about the underlying mechanisms precipitating the potentially interesting drug targets. When a 4b3g2S receptors dysfunction. Treatment of the mentioned diseases with are compared with a 4 b3d containing receptors in Xenopus GABAA receptor ligands is therefore primarily symptom oocytes, THIP is approximately 10 times more potent at d alleviation with the implicit assumption that the CNS after a containing receptors than at g2S containing receptors. certain time with the perturbation of the homeostasis tends Furthermore, at d containing receptors, THIP appears to have to regress towards the normal equilibrium. There are, a completely different pharmacological profile than at other however, significant deviations from this tendency, since receptor combinations [82]. Thus, despite the fact that the schizophrenia and epilepsy once precipitated are life long GABAA binding site is located at the interface between a diseases, although with significant temporal fluctuations in and b subunits, the pharmacological profile of the receptor is severity. In the case of epilepsy, several attempts have been determined by the interaction of all subunits present [65]. made to develop GABAergic compounds as novel drugs. Much evidence now suggests that although GABAA receptor THIP appears to act as a partial agonist at some receptor agonists, direct and indirect acting, may act as subunit combinations and at the same time as an agonist anticonvulsive compounds, the antiepileptic activity in the with higher efficacy than GABA at other receptor sense of inhibiting epilepsy and the progression of epilepsy, combinations. This functional selectivity obtained in vitro is very limited. Thus, compounds as , barbiturates, has been demonstrated for most GABAA receptor agonists, the GABA uptake inhibitor Tiagabin and THIP may be but our understanding of the mechanism(s) underlying the active as anticonvulsives but at the same time only possess translation of this effect into activity of the CNS in an intact limited effects when used as monotherapies. organism is still very limited. Furthermore, since the 828 Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 Frølund et al.

GABAA binding site is conserved within known subunit GABA concentrations are different from those seen with compositions, and since pharmacophore modelling of GABAA receptor modulators. GABAA receptor agonist interactions still is based on affinity data, development of subtype selective agonists is hampered by our lack of methods for prediction of activity. CONCLUSION When it becomes possible to model functional activity including potency and maximal response, will it be possible Molecular biology studies have revealed a high degree of to predict subtype selectivity. structural heterogeniety of the GABAA receptors. Development of subtype selective or specific compounds is GABAergic compounds acting at the site, the of key importance for the understanding of the physiological site, or the site of the GABAA and pathological roles of different receptor subtypes and may receptor complex have been used as for years [83]. lead to valuable therapeutic agents. Studies along these lines Most of these hypnotics interact with all of the different have so far been complicated by the lack of information subtypes of receptors. However, as a 4 and d containing about the topography of the recognition site(s) of the receptors do not bind benzodiazepines, effect of GABAA receptor complex. In the absence of a crystal benzodiazepines at these receptor subtypes is absent [84]. It structure of the GABAA receptor complex, 3D- is generally agreed that the effect of benzodiazepines pharmacophore models, based on the analysis of known is mediated primarily via a 1 containing receptors [85, 86] receptor ligands, have been useful tools in the design of new whereas side effects related to amnesia may be mediated by ligands. Recombinant receptors play a key role in the a 5, primarily located in the hippocampal region [87]. pharmacological characterization of ligands for the GABAA However, short-acting a 1 selective compounds like receptor. However, problems regarding subunit composition zaleplone, and do produce memory of native receptors are far from being fully elucidated. impairment and hangover effects [88], suggesting that even Furthermore, the subtypes of receptors involved in different an a 1 selective compound with a very short half life may disorders are still largely unknown. Although the functional produce side effects. The reason for this side effect profile consequences of modifications of subunit compositions so may well be a consequence of the high degree of GABA far is unpredictable, it has been shown that functional receptor activation caused by the positive GABAA receptor selectivity is obtainable for a number of GABAA agonists. modulator. The massive activation of GABA receptors will An important aspect is the selection of subunit combinations influence several other systems, ultimately resulting in a relevant for the prediction of in vivo activity. general acute modification of the overall function of the CNS. Electroencephalographic (EEG) measurements during Neither full GABAA agonists nor antagonists may be sleep support this hypothesis. A normal sleep pattern useful therapeutic agents. Whereas the high-efficacy partial involves a complex variation of different degrees of sleep, GABAA agonist THIP shows very potent non-opioid ranging from light sleep via deeper sleep stages to the analgesic effects in the human clinic, it seems likely that dream-associated rapid eye movement (REM) stage of sleep. partial GABAA agonists showing lower levels of efficacy Present understanding of sleep quality and the relation to such as thio-4-PIOL or related 4-PIOL analogues, may have EEG patterns is still limited, however, not only the duration therapeutic interest in certain CNS disorders. Information of REM sleep but also the transitions between the different about the basic mechanism of receptor-ligand interaction sleep stages are important [88]. The effect of benzodiazepines resulting in partial agonism is, however, still not available, at the sleep micro architecture is not limited to the onset of making design of new partial GABAA agonists on a rational REM sleep, which is delayed. Thus, the number of basis difficult. transitions between different sleep stages is significantly reduced – probably as a consequence of a general depression of the excitatory systems [89]. In contrast to the ACKNOWLEDGEMENTS observations with benzodiazepines, barbiturates and neurosteroids, a series of studies have shown that this This work was supported by grants from the Lundbeck perturbation of the sleep micro architecture may not arise Foundation, the Danish Medical Research council, the with compounds acting directly at the GABAA recognition Centre for Drug Design and Transport, the NeuroScience site [89, 90]. In this respect, THIP is an interesting PharmaBiotec Centre and the Danish State Biotechnology compound. THIP is currently developed as and Programme. appears to improve the quality of sleep, as measured using behavioural parameters. The compound shows no effect on the onset of REM sleep, as measured using EEG [89, 90]. REFERENCES Furthermore, studies in elderly patients, suffering from significant reduction in non REM sleep, showed that [1] Curtis, D.R.; Duggan, A.W; Felix, D.; Johnston, G.A.R. . Nature Gaboxadol was able to normalize the sleep pattern [91]. GABA, Bicuculline and Central Inhibition 1970, 226, 1222-1224. Following dosing with Gaboxadol, patients did not experience the hangovers or impaired attention as reported [2] Johnston, G.A.R.; Beart, P.M.; Curtis, D.R.; Game, C.J.A.; for benzodiazepines [92]. Similar results have been obtained McCulloch, R.M.; MacLachlan, R.M. Bicuculline with muscimol and with the GABA uptake inhibitor Methochloride as a GABA Antagonist. Nature 1972, Tiagabine [93], strongly suggesting that the functional 240, 219-220. consequences of a direct acting agonist or enhanced synaptic GABAA Receptor Ligands and their Therapeutic Potentials Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 829 [3] Curtis, D.R.; Game, C.J.A.; Johnston, G.A.R.; McCulloch, [18] Froestl, W.; Mickel, S.J.; von Sprecher, G.; Diel, P.J.; R.M. Central Effects of b-(p-Chlorophenyl)-g- Hall, R.G.; Maier, L.; Strub, D.; Melillo, V.; Baumann, aminobutyric Acid. Brain Res. 1974, 70, 493-499. P.A.; Bernasconi, R.; Gentsch, C.; Hauser, K.; Jaekel, J.; Karlsson, G.; Klebs, K.; Maltre, L.; Mondadori, C.; Olpe, [4] Bowery, N.G., Classification of GABA Receptors. In The H.R.; Walmeier, P.C.; Bittiger, H. Phosphinic Acid GABA Receptors, Enna, S.J.; Ed.; 1983; The Humana Analogues of GABA. 2. Selective, Orally Active GABAB Press: Clifton. p. 177-213. Antagonists. J. Med. Chem. 1995, 38, 3313-3331.

[5] Bowery, N.G.; Hill, D.R.; Hudson, A.L.; Doble, A.; [19] Froestl, W.; Mickel, S.J.; Hall, R.G.; von Sprecher, G.; Middlemiss, D.N.; Shaw, J.; Turnbull, M. (-)-Baclofen Strub, D.; Baumann, P.A.; Brugger, F.; Gentsch, C.; Decreases Neurotransmitter Release in the Mammalian Jaekel, J.; Olpe, H.R.; Rihs, G.; Vassout, A.; Waldmeier, CNS by an Action at a Novel GABA Receptor. Nature P.C.; Bittiger, H. Phosphinic Acid Analogues of GABA. 283 1980, , 92-94. 1. New Potent and Selective GABAB Agonists. J. Med. Chem. 1995, 38, 3297-312. [6] Johnston, G.A.R.; Curtis, D.R.; Beart, P.M.; Game, C.J.A.; McCulloch, R.M.; Twitchin, B. Cis- and Trans-4- [20] Sieghart, W.; Fuchs, K.; Tretter, V.; Ebert, V.; Jechlinger, aminocrotonic Acid as GABA Analogues of Restricted M.; Höger, H.; Adamiker, D. Structure and Subunit Conformation. J.Neurochem. 1975, 24, 157-160. Composition of GABAA Receptors. Neurochem.Int. 1999, 34, 379-385. [7] Allan, R.D.; Curtis, D.R.; Headly, P.M.; Johnston, G.A.R.; Lodge, D.; Twitchin, B. The Synthesis and Activity of [21] Mehta, A.K.; Ticku, M.K. An Update on GABAA Cis- and Trans-2-(aminomethyl)cyclopropanecarboxylic Receptors. Brain Res.Rev. 1999, 29, 196-217. Acid as Conformationally Restricted Analogues of GABA. J. Neurochem. 1980, 34, 652-656. [22] Tretter, V.; Ehya, N.; Fuchs, K.; Sieghart, W. Stoichiometry and Assembly of a Recombinant GABA [8] Drew, C.A.; Johnston, G.A.R.; Weatherby, R.P. A Receptor Subtype. J. Neurosci. 1997, 17, 2728-2737. Bicuculline Insensitive GABA Receptors: Studies on the Binding of (-)-Baclofen to Rat Cerebellar Membranes. [23] Nayeem, N.; Green, T.P.; Martin, I.L.; Barnard, E.A. Neurosci. Lett. 1984, 52, 317–321. Quaternary Structure of the Native GABAA Receptor [9] Barnard, E.A.; Skolnick, P.; Olsen, R.W.; Möhler, H.; Determined by Electron Microscopic Image Analysis. Sieghart, W.; Biggio, G.; Braestrup, C.; Bateson, A.N.; J.Neurochem. 1994, 62, 815-818. Langer, S.Z. International Union of Pharmacology. XV. [24] Smith, G.B.; Olsen, R.W. Functional Domains of GABAA Subtypes of Gamma-Aminobutyric AcidA Receptors: Classification on the Basis of Subunit Structure and Receptors. Trends Pharmacol.Sci. 1995, 16, 162-168. Receptor Function. Pharmacol.Rev. 1998, 50, 291-313. [25] Olsen, R.W.; Macdonald, R.L. GABAA receptor complex: [10] Chebib, M.; Johnston, G.A.R. GABA-Activated Ligand Structure and function. Glutamate and GABA Receptors Gated Ion Channels: Medicinal Chemistry and and Transporters: Structure, Function and Molecular Biology. J. Med. Chem. 2000, 43, 1427-47. Pharmacology, Egebjerg, J.; Schousboe, A.; Krogsgaard- Larsen. P. Eds.; 2002, Taylor & Francis: New York. 202– [11] Krogsgaard-Larsen, P.; Johnston, G.A.R.; Lodge, D.; 235. Curtis, D.R. A New Class of GABA Agonist. Nature 1977, 268, 53-55. [26] Möhler, H.; Knoflach, F.; Paysan, J.; Motejlek, K.; Benke, D.; Lüscher, B.; Fritschy, J.M. Heterogeneity of GABAA- [12] Krogsgaard-Larsen, P.; Snowman, A.; Lummis, S.C.; Receptors: Cell-Specific Expression, Pharmacology, and Olsen, R.W. Characterization of the Binding of the Regulation. Neurochem. Res. 1995, 20, 631–636. GABA Agonist [3H]Piperidine-4-sulfonic Acid to Bovine Brain Synaptic Membranes. J.Neurochem. 1981, [27] Benke, D.; Mertens, S.; Trzeciak, A.; Gillessen, D.; 37, 401-409. Möhler, H.; GABAA Receptors Display Association of g2-Subunit with a 1- and b2/3-Subunits. J. Biol. Chem. [13] Rudolph, U.; Crestani, F.; Mohler, H. GABAA Receptor 1991, 266, 4478-4483. Subtypes: Dissecting their Pharmacological Functions. Trends Pharmacol. Sci. 2001, 22, 188-194. [28] McKernan, R.M.; Whiting, P.J. Which GABAA-Receptor Subtypes Really Occur in the Brain? Trends Neurosci. [14] Haefely, W.; Kulcsar, A.; Möhler, H.; Pieri, L.; Polc, P.; 1996, 19, 139-143. Schaffner, S. R. Possible Involvement of GABA in the Central Actions of Benzodiazepines. Adv. Biochem. [29] Krogsgaard-Larsen, P.; Johnston, G.A.R.; Curtis, D.R.; Psychopharmacol. 1975, 14, 131–151. Game, C.J.A.; McCulloch, R.M. Structure and Biological Activity of a Series of Conformationally Restricted [15] Möhler, H.; Okada, T. Benzodiazepine Receptors: Analogs of GABA. J.Neurochem. 1975, 25, 803-809. Demonstration in the Central Nervous System. Science 1977, 198, 849-851. [30] Tunnicliff, G., Pharmacology and Function of Imidazole 4-Acetic Acid in Brain. Gen. Pharmac. 1998, 31, 503– [16] Squires, R.F.; Braestrup, C. Benzodiazepine Receptors in 509. Rat Brain. Nature 1977, 266, 732-734. [31] Campbell, M.M.; Mickel, S.J.; Singh, G. (±)-2-Amino-2- [17] Sieghart, W., Structure and Pharmacology of g- thiazoline-4-ethanoic Acid; A Novel, Specific GABAA Aminobutyric AcidA Receptor Subtypes. Pharmacol. Receptor Agonist. Bioorg.Med.Chem.Lett. 1991, 1, 247- Rev. 2000, 47, 181-234. 248. 830 Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 Frølund et al.

[32] Allan, R.D.; Johnston, G.A.R.; Twitchin, B. Synthesis of Mechanisms Operative at the GABA Receptor. J. Analogues of GABA. III. All Four Stereoisomers of 3- Neurosci. Res. 1995, 42, 666-673. Aminocyclopentanecarboxylic Acid and a Stereochemical Correlation with Amidinomycin. [46] Brehm, L.; Frydenvang, K.;. Hansen, L.M.; Norrby, P.-O.; Austr.J.Chem. 1979, 32, 2517-2521. Krogsgaard-Larsen, P.; Liljefors, T. Structural Features of Muscimol, a Potent GABAA Receptor Agonist. Crystal [33] Seutin, V.; Johnson, S.W. Recent Advances in the Structure and Quantum Chemical Ab initio Calculations. Pharmacology of Quaternary Salts of Bicuculline. Struct.Chem. 1997, 8, 443-451. Trends.Pharmacol.Sci. 1999, 20, 268-270. [47] Krogsgaard-Larsen, P.; Mikkelsen, H.; Jacobsen, P.; [34] Wermuth, C.G.; Bizière, K. Pyridazinyl-GABA Falch, E.; Curtis, D.R.; Peet, M.J.; Leah, J.D. 4,5,6,7- Derivatives: A New Class of Synthetic GABAA Tetrahydroisothiazolo[5,4-c]pyridin-3-ol and Related Antagonists. Trends Pharmacol.Sci. 1986, 7, 421-424. Analogues of THIP. Synthesis and Biological Activity. J.Med.Chem. 1983, 26, 895-900. [35] Wermuth, C.G.; Bourguignon, J.-J.; Schlewer, G.; Gies, J.-P.; Schoenfelder, A.; Melikian, A.; Bouchet, M.-J.; [48] Tagmose, L., A Pharmacophore Model for GABAA Chantreux, D.; Molimard, J.-C.; Heaulme, M.; Chambon, Receptor Agonists. Ph.D. Thesis, Royal Danish School J.-P.; Bizière, K. Synthesis and Structure-Activity of Pharmacy, Copenhagen, 2000. Relationships of a Series of Aminopyridazine Derivatives of g-Aminobutyric Acid Acting as Selective [49] Frølund, B., Tagmose, L.; Liljefors, T.; Stensbøl, T.B.; GABA-A Antagonists. J.Med.Chem. 1987, 30, 239-249. Engblom, C.; Kristiansen, U.; Krogsgaard-Larsen, P. A Novel Class of Potent 3-Isoxazolol GABAA [36] Melikian, A.; Schlewer, G.; Chambon, J.-P.; Wermuth, Antagonists: Design, Synthesis, and Pharmacology. C.G. Condensation of Muscimol or Thiomuscimol with J.Med.Chem. 2000, 43, 4930–4933. Aminopyridazines Yields GABA-A Antagonists. J.Med.Chem. 1992, 35, 4092-4097. [50] Frølund, B.; Jørgensen, A.T.; Tagmose, L.; Stensbøl, T.B.; Vestergaard, H.T.; Engblom, C.; Kristiansen, U.; [37] Arnt, J.; Krogsgaard-Larsen, P. GABA Agonists and Sanchez, C.; Krogsgaard-Larsen, P.; Liljefors, T. A Novel Potential Antagonists Related to Muscimol. Brain Res. Class of Potent 4-Arylalkyl Substituted 3-Isoxazolol 1979, 177, 395-400. GABAA Antagonists: Synthesis, Pharmacology and Molecular Modeling. J.Med.Chem. 2002, in press. [38] Krogsgaard-Larsen, P.; Hjeds, H.; Curtis, D.R.; Lodge, D.; Johnston, G.A.R.; Dihydromuscimol, Thiomuscimol and [51] Krogsgaard-Larsen, P.; Hjeds, H. Organic Hydroxylamine Related Heterocyclic Compounds as GABA Analogues. Derivatives. X. Structural Analogs of g-aminobutyric J.Neurochem. 1979, 32, 1717-1724. acid (GABA) of the Isoxazole Enol-betaine Type. Synthesis of 5,6,7,8-Tetrahydro-4H-isoxazolo[4,5- [39] Krogsgaard-Larsen, P.; Brehm, L.; Schaumburg, K. d]azepin-3-ol Zwitterion and 4,5,6,7-Tetrahydroisoxa- Muscimol, a Psychoactive Constituent of Amanita zolo[4,5-c]pyridin-3-ol Zwitterion. Acta Chem.Scand., Muscaria, as a Medicinal Chemical Model Structure. Ser.B 1974, 28, 533-538. Acta Chem.Scand., Ser.B 1981, 311, 0-24. [52] Krogsgaard-Larsen, P.; Roldskov-Christiansen, T. GABA [40] Fowler, L.J.; Lovell, D.H.; John, R.A. Reaction of Agonists. Synthesis and Structure-Activity Studies on Muscimol with 4-Aminobutyrate Aminotransferase. J. Analogs of Isoguvacine and THIP. Eur.J.Med.Chem.- Neurochem. 1983, 41, 1751-1754. Chim.Ther. 1979, 14, 157-164.

[41] Kendall, D.A.; Browner, M.; Enna, S.J. Comparison of the [53] Krehan, D.; Frølund, B.; Krogsgaard-Larsen, P.; Antinociceptive Effect GABA Agonists: Evidence for a Johnston, G.A.R.; Chebib, M.; Aza-THIP and Related Cholinergic Involvement. J. Pharmacol. Exp. Ther. Analogues of THIP as GABAC Receptor Antagonists. 1982, 220, 482-487. Bioorg.Med.Chem. Lett. 2002, submitted. [42] Lancel, M.; Langebartels, A. g-aminobutyric Acid(A) [54] Haefliger, W.; Révész, L.; Maurer, R.; Römer, D.; Büscher, (GABA ) Agonist 4,5,6,7-Tetrahydroisoxazolo[4,5- A H.-H. Analgesic GABA Agonists. Synthesis and c]pyridin-3-ol Persistently Increases Sleep Maintenance Structure-Activity Studies on Analogues and and Intensity During Chronic Administration to Rats. J. Derivatives of Muscimol and THIP. Eur.J.Med.Chem.- Pharmacol. Exp. Ther. 2000, 293, 1084-1090. Chim.Ther. 1984, 19, 149–156. [43] Krogsgaard-Larsen, P.; Falch, E.; Hjeds, H. Heterocyclic Analogues of GABA: Chemistry, Molecular [55] Krogsgaard-Larsen, P.; Johnston, G.A.R. Structure- Pharmacology and Therapeutic Aspects. Prog.Med.Chem. Activity Studies on the Inhibition of GABA Binding to 1985, 22, 67-120. Rat Brain Membranes by Muscimol and Related Compounds. J.Neurochem. 1978, 30, 1377-1382. [44] Buur, J.R.B.; Hjeds, H.; Krogsgaard-Larsen, P.; F.S. J¢rgensen, Conformational Analysis and Molecular [56] Falch, E., P. Jacobsen,; Krogsgaard-Larsen, P.; Curtis, D.R. GABA-Mimetic Activity and Effects on Diazepam Modelling of a Partial GABAA Agonist and a Antagonist Related to the GABAA Agonist, THIP. Drug Binding of Aminosulphonic Acids Structurally Related Des. Delivery 1993, 10, 213-229. to Piperidine-4-sulphonic Acid. J.Neurochem. 1985, 44, 68-75. [45] Galvez-Ruano, E.; Aprison, M.H.; Robertson, D.H.; Lipkowitz, K.B. Identifying Agonistic and Antagonistic [57] Ebert, B.; Mortensen, M.; Thompson, S.A.; Kehler, J.; Wafford, K.A.; Krogsgaard-Larsen, P. Bioisosteric GABAA Receptor Ligands and their Therapeutic Potentials Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 831 Determinants for Subtype Selectivity of Ligands for [70] Frølund, B.; Kristiansen, U.; Brehm, L.; Hansen, A.B.; Heteromeric GABAA Receptors. Bioorg. Med. Chem. Lett. Krogsgaard-Larsen, P.; Falch, E. Partial GABAA Receptor 2001, 11, 1573-1577. Agonists. Synthesis and In Vitro Pharmacology of a Series of Nonannulated Analogs of 4,5,6,7-Tetrahydro- [58] Ebert, B.; Thompson, S.A.; Suonatsou, K.; McKernan, R.; isoxazolo[5,4-c]pyridin-3-ol (THIP). J.Med.Chem. 1995, Krogsgaard-Larsen, P.; Wafford, K.A. Differences in 38, 3287-3296. Agonist/Antagonist Binding Affinity and Receptor Transduction using Recombinant Human g-Amino- [71] Mortensen, M.; Frølund, B.; Jørgensen, A.T.; Liljefors, butyric Acid type A Receptors. Mol.Pharmacol. 1997, T.; Krogsgaard-Larsen, P.; Ebert, B. Activity of Novel 4- 52, 1150-1156. PIOL Analogues at Human a 1b2g2S GABAA Receptors - Correlation with Hydrophobicity. Eur.J.Pharmacol. [59] Krehan, D.; Frølund, B.; Kehler, J.; Johnston, G.A.R.; 2002, submitted. Krogsgaard-Larsen, P.; Chebib, M. Seleninic, Phosphinic and Phosphonic Acid Analogues of Isonipecotic Acid as [72] Ebert, B., Storustovu, S.I.; Mortensen, M.; Frølund, B. GABAC Receptor Antagonists. Eur. J. Pharmacol. 2002, Characterisation of GABAA Receptor Ligands in the Rat submitted. Cortical Wedge Preparation - Evidence for Action at Extra Synaptic Receptors. British J. Pharmacol. 2002, [60] Stuhr-Hansen, N.; Ebert, B.; Krogsgaard-Larsen, P.; submitted. Kehler, J. Synthesis and Pharmacology of Seleninic Acid Analogues of the Inhibitory Neurotransmitter g- [73] Rabe, H.; Picard, R.; Uusi-Oukari, M.; Hevers, W.; Aminobutyric Acid. Org. Lett. 2000, 2, 7-9. Luddens, H.; Korpi, E.R. Coupling Between Agonist and Chloride Ionophore Sites of the GABAA Receptor: [61] Maksay, G., Thermodynamics of g-Aminobutyric Acid Agonist/Antagonist Efficacy of 4-PIOL. Eur. J Type A Receptor Binding Differentiate Agonists from Pharmacol. 2000, 409, 233-242. Antagonists. Mol. Pharmacol. 1994. 46, 386-390. [74] Hartvig, L.; Lükensmejer, B.; Liljefors, T.; [62] Krogsgaard-Larsen, P.; Hjeds, H.; Falch, E.; J¢rgensen, Dekermendjian, K. Two Conserved Arginines in the F.S.; Nielsen, L. Recent Advances in GABA Agonists, Extracellular N-terminal Domain of the GABAA Receptor Antagonists and Uptake Inhibitors: Structure-Activity a 5 Subunit are Crucial for Receptor Function. Relationships and Therapeutic Potential. Adv.Drug Res. J.Neurochem. 2000, 75, 1746-1753. 1988, 17, 381-456. [75] Westh-Hansen, S.E.; Witt, M.R.; Dekermendjian, K.; [63] Braestrup, C.; Nielsen, M.; Krogsgaard-Larsen, P.; Falch, Liljefors, T.; Rasmussen, P.B.; Nielsen, M. Arginine E. Partial Agonists for Brain GABA/Benzodiazepine Residue 120 of the Human GABAA Receptor alpha1, Receptor Complex. Nature 1979, 280, 331-333. Subunit is Essential for GABA Binding and Chloride Ion Current Gating. NeuroReport 1999, 10, 2417-2421. [64] Hansen, S.L.; Ebert, B.; Fjalland, B.; Kristiansen, U. Effects of GABAA Receptor Partial Agonists in Primary [76] Boileau, A.J.; Evers, A.R.; Davis, A.F.; Czajkowski, C. Cultures of Cerebellar Granule Neurons and Cerebral Mapping the Agonist Binding Site of the GABAA Cortical Neurons Reflect Different Receptor Subunit Receptor: Evidence for a b-Strand. J.Neurosci. 1999, 19, Compositions. Br. J. Pharmacol. 2001, 133, 539-549. 4847-4854.

[65] Ebert, B.; Wafford, K.A.; Whiting, P.J.; Krogsgaard- [77] Unwin, N., Nicotinic Acetylcholine Receptor at 9 Å Larsen, P.; Kemp, J.A. Molecular Pharmacology of g- Resolution. J. Mol. Biol. 1993, 229, 1101-24. Aminobutyric Acid Type A Receptor Agonists and Partial Agonists in Oocytes Injected with Different a , b [78] Unwin, N., Acetylcholine Receptor Channel Imaged in and g Receptor Subunit Combinations. Mol.Pharmacol. the Open State. Nature 1995, 373, 37-43. 1994, 46, 957-963. [79] Amin, J.; Weiss, D.S. GABAA Receptor Needs Two [66] Verdoorn, T.A., Formation of Heteromeric g-Amino- Homologous Domains of the b-Subunit for Activation butyric Acid Type A Receptors Containing Two by GABA but Not by . Nature 1993, 366, Different a Subunits. Mol. Pharmacol. 1994, 45, 475- 565-569. 480. [80] Sur, C.; Farrar, S.J.; Kerby, J.; Whiting, P.J.; Atack, J.R.; [67] Byberg, J.R.; Labouta, I.M.; Falch, E.; Hjeds, H.; McKernan, R.M. Preferential Coassembly of a 4 and d Krogsgaard-Larsen, P.; Curtis, D.R.; Gynther, B.D. Subunits of the g-Aminobutyric AcidA Receptor in Rat Synthesis and Biological Activity of a GABA-A Agonist Thalamus. Mol. Pharmacol. 1999, 56, 110-115. Which Has No Effect on Benzodiazepine Binding and Structurally Related Glycine Antagonists. Drug [81] Mody, I., Synaptic Activation of GABA-Rs', Presentation Des.Delivery 1987, 1, 261-274. at GABA2000 meeting 23 to 29 July. Cairns, Australia. Organisers Peter Gage and Istvan Mody. 2000. [68] Falch, E.; Larsson, O.M.; Schousboe, A.; Krogsgaard- Larsen, P. GABA-A Agonists and GABA Uptake [82] Wafford, K.A., Properties of a Stable Cell Line Inhibitors: Structure-Activity Relationships. Drug Expressing Human a 4b3d GABAA Receptors. Dev.Res. 1990, 21, 169-188. Presentation at the GABA2000 meeting 23 to 29 July. Cairns, Australia. Organisers Peter Gage and Istvan [69] Kristiansen, U.; Lambert, J.D.C.; Falch, E.; Krogsgaard- Mody. 2000. Larsen, P.; Electrophysiological Studies of the GABAA Receptor Ligand, 4-PIOL, on Cultured Hippocampal Neurones. Br.J.Pharmacol. 1991, 104, 85-90. 832 Current Topics in Medicinal Chemistry, 2002, Vol. 2, No. 8 Frølund et al.

[83] Haefely, W.E., Pharmacology of the Benzodiazepine [88] Landolt, H.P.; Gillin, J.C. GABAA a 1 Receptors: Receptor. Eur. Arch. Psychiatry Neurol. Sci. 1989, 238, Involvement in Sleep Regulation and Potential of 294-301. Selective Agonists in the Treatment of Insomnia. CNS Drugs 2000, 13, 185–199. [84] Wafford, K.A.; Thompson, S.A.; Thomas, D.; Sikela, J.; Wilcox, A.S.; Whiting, P.J. Functional Characterization [89] Lancel, M.; Cronlein, T.A.; Faulhaber, J. Role of GABAA of Human g-Aminobutyric AcidA Receptors Containing Receptors in Sleep Regulation. Differential Effects of the a 4 Subunit. Mol. Pharmacol. 1996, 50, 670-678. Muscimol and on Sleep in Rats. Neuropsychopharmacology 1996, 15, 63-74. [85] McKernan, R.M.; Rosahl, T.W.; Reynolds, D.S.; Sur, C.; Wafford, K.A.; Atack, J.R.; Farrar, S.; Myers, J.; Cook, G.; [90] Lancel, M.; Faulhaber, J. The GABAA Agonist THIP Ferris, P.; Garrett, L.; Bristow, L.; Marshall, G.; Macaulay, (Gaboxadol) Increases Non-REM Sleep and Enhances A.; Brown, N.; Howell, O.; Moore, K.W.; Carling, R.W.; Delta Activity in the Rat. NeuroReport 1996, 7, 2241– Street, J.L.; Castro, J.L.; Ragan, C.I.; Dawson, G.R.; 2245. Whiting, P.J. Sedative but Not Anxiolytic Properties of Benzodiazepines are Mediated by the GABAA Receptor [91] Lancel, M.; Wetter, T.C.; Steiger, A.; Mathias, S. Effect of a 1 Subtype. Nat. Neurosci. 2000, 3, 587-592. the GABAA Agonist Gaboxadol on Nocturnal Sleep and Hormone Secretion in Healthy Elderly Subjects. Am. J. [86] Rudolph, U.; Crestani, F.; Benke, D.; Brunig, I.; Benson, Physiol. Endocrinol. Metab. 2001, 281, E130-137. J.A.; Fritschy, J.M.; Martin, J.R.; Bluethmann, H.; Mohler, H. Benzodiazepine Actions Mediated by [92] Faulhaber, J.; Steiger, A.; Lancel, M. The GABAA Agonist Specific g-Aminobutyric AcidA Receptor Subtypes. THIP Produces Slow Wave Sleep and Reduces Spindling Nature 1999, 401, 796-800. Activity in NREM Sleep in Humans. Psychopharmacology 1997, 130, 285-291. [87] Collison, N., Cognition Enhancing Effects of a GABAA a 5 Selective Inverse Agonist. Presentation at [93] Lancel, M.; Faulhaber, J.; Deisz, R.A. Effect of the GABA GABA2000 meeting 23 to 29 July. Cairns, Australia. Uptake Inhibitor Tiagabine on Sleep and EEG Power Organisers Peter Gage and Istvan Mody. 2000. Spectra in the Rat. Br. J. Pharmacol. 1998, 123, 1471- 1477.