Functional Characterization of the 1,5-Benzodiazepine Clobazam and Its Major Active Metabolite N-Desmethylclobazam at Human GABAA Receptors Expressed in Xenopus laevis Oocytes

The 1,5-benzodiazepine clobazam is indicated for the adjunctive treatment of seizures associated with Lennox-Gastaut syndrome in patients 2 years of age or older in the United States, and for treatment of anxiety and various forms of epilepsy elsewhere. Clobazam has been reported to exhibit different in vivo adverse effects and addiction liability profile than the classic 1,4-benzodiazepines. In this study, it was investigated whether the in vitro pharmacological properties of clobazam and its major active metabolite N-desmethylclobazam could explain some of these clinical differences. The functional properties of the two 1,5-benzodiazepines were characterized at the human γ-aminobutyric acid type A receptor (GABAAR) subtypes α1β2γ2S, α2β2γ2S, α3β2γ2S, α5β2γ2S and α6β2δ expressed in Xenopus laevis oocytes by use of two-electrode voltage-clamp electrophysiology and compared to those exhibited by the 1,4-benzodiazepine clonazepam. All three compounds potentiated GABA EC20-evoked responses through the α1,2,3,5β2γ2S GABAARs in a reversible and concentration-dependent manner, with each displaying similar EC50 values at the four subtypes. Furthermore, the degrees of potentiation of the GABA EC20 currents through the four receptors mediated by saturating modulator concentrations did not differ substantially for any of the three benzodiazepines. The three compounds were substantially less potent (200-3900 fold) as positive allosteric modulators at the α6β2δ GABAAR than at the α1,2,3,5β2γ2S receptors. Interestingly, however, clobazam and especially N-desmethylclobazam were highly efficacious potentiators of α6β2δ receptor signaling. Although this activity component is unlikely to contribute to the in vivo effects of clobazam/N-desmethylclobazam, the 1,5-benzodiazepine could constitute an interesting lead for novel modulators targeting this low-affinity binding site in GABAARs. In conclusion, the non-selective modulation exerted by clobazam, N-desmethylclobazam and clonazepam at the α1β2γ2S, α2β2γ2S, α3β2γ2S and α5β2γ2S GABAARs indicate that the observed clinical differences between clobazam and 1,4-benzodiazepines are likely to arise from factors other than their respective pharmacological properties at the GABAARs as investigated here.

The signaling through the GABA A R is susceptible to modulation by numerous ligands acting through different allosteric sites in the receptor complex, and many of these ligands are used to treat human pathologies [12][13][14][15][16]. Delineation of the molecular modes of action of these modulators at the receptors has provided considerable insight into the signal transduction mechanism of the GABA A R as well as the molecular compositions of its allosteric sites. The allosteric modulation of the α 1,2,3,5 βγ GABA A Rs exerted by benzodiazepines is predominantly mediated through a high-affinity binding site located in the extracellular α (+) /γ (-) subunit interface of the receptor [14,17]. However, benzodiazepines have also been proposed to target a low-affinity binding site located in the transmembrane domains of both αβ and αβγ GABA A Rs [18]. Thus, the benzodiazepines have been proposed to possess a nM activity component arising exclusively from α 1,2,3,5 βγ GABA A Rs and a μM activity component that potentially could involve all GABA A Rs [18]. Furthermore, several recent studies have proposed the existence of a low-affinity binding site for some benzodiazepines and other benzodiazepine-site ligands such as CGS 9895, LAU 177 and Ro 15-4513 in the extracellular α (+) /β (-) subunit interface of the GABA A R, a site homologous to the extracellular high-affinity α (+) /γ (-) binding site for benzodiazepines in the αβγ GABA A R [19][20][21][22][23].
The selectivity profile of a particular benzodiazepine at the different α 1,2,3,5 βγ GABA A R subtypes is believed to correlate to its clinical efficacy and adverse effects. Insights gained from studies of knock-in mice expressing benzodiazepine-insensitive subtypes have provided the rationale for the development of positive allosteric modulators (PAMs) of α 1 -containing subtypes as hypnotics, of PAMs of α 2 -or α 3 -containing subtypes as anxiolytics and analgesics, and of negative allosteric modulators (NAMs) of α 5 -containing receptors as cognitive enhancers [12,13,15,16,24]. Moreover, other studies suggest that the anticonvulsive and anxiolytic effects of benzodiazepines may be mediated via α 2 -containing GABA A Rs, whereas modulation of these subtypes has little or no sedative effects [25][26][27][28][29]. However, experiments in rodents indicate that the anticonvulsant effects require modulation of more than one specific α-subunit-containing GABA A R and that modulation of different subtypes may act synergistically [29]. To date, most benzodiazepines reported capable of modulating GABA A R signaling are 1,4benzodiazepines. One of the few exceptions is the 1,5-benzodiazepine clobazam (Fig. 1). Clobazam (Onfi) has recently been approved in the United States and is indicated for the adjunctive treatment of seizures associated with Lennox-Gastaut syndrome in patients aged 2 years and older, and outside the United States the drug is routinely administered for anxiety disorders and epilepsy [30][31][32]. Interestingly, clobazam has been reported to exhibit different in vivo adverse effects and addiction liability profile than the 1,4-benzodiazepines, including clonazepam (Klonopin, Fig. 1), an antiepileptic drug approved for the treatment of Lennox-Gastaut syndrome (petit mal variant), and akinetic and myoclonic seizures. For example, studies suggest that clobazam induces fewer psychomotor disturbances than clonazepam when dosed at clinically effective concentrations in healthy volunteers [33,34]. Moreover, preclinical studies suggest that clobazam in contrast to clonazepam and other 1,4-benzodiazepines may exert more specific anticonvulsant/antiepileptic over sedative effects [35,36]. The major active metabolite of clobazam, N-desmethylclobazam ( Fig. 1), has been shown to have a longer plasma half-life than the parent compound (79 h vs. 36 h), resulting in greater metabolite plasma concentrations following long-term clobazam dosing in humans [37]. Hence, the metabolite is likely to contribute significantly to the clinical efficacy of clobazam, and interestingly N-desmethylclobazam has been reported to produce fewer adverse effects than clobazam [38,39].
Considering that clobazam has been administered for various indications in the clinic for decades, the current insight into the pharmacological characteristics of clobazam and N-desmethylclobazam at GABA A Rs is surprisingly limited [40]. In a recent study, we delineated the binding characteristics of clobazam, N-desmethylclobam and the 1,4-benzodiazepine clonazepam at native GABA A Rs in rat brain membranes and at human α 1,2,3,5 β 2 γ 2S GABA A Rs expressed in HEK293 cells in a [ 3 H]flumazenil competition binding assay [41]. In the present study, we have characterized the functional properties of clobazam and N-desmethylclobazam at human α 1 β 2 γ 2S , α 2 β 2 γ 2S , α 3 β 2 γ 2S , α 5 β 2 γ 2S and α 6 β 2 δ GABA A Rs expressed in Xenopus laevis oocytes using two-electrode voltage clamp (TEVC) electrophysiology and compared these functionalities to that exhibited by clonazepam.

Electrophysiological recordings
Electrophysiological recordings were performed using the TEVC technique on Xenopus oocytes expressing various GABA A R combinations using a protocol adapted from previous studies [44,45]. Oocytes were placed in a recording chamber and gravity perfused with Ringer buffer [115 mM NaCl, 2.5 mM KCl, 10 mM HEPES (pH 7.5), 1.8 mM CaCl 2 , 0.1 mM MgCl 2 ]. Cells were impaled with agar-plugged 0.5-1 MO electrodes containing 3 M KCl and voltage clamped at −70 mV by a Gene Clamp 500B amplifier (Axon Instruments, Union City, CA, USA) and recorded with pClamp 10 (Windows version, Molecular Devices, LLC, Sunnyvale, CA, USA). The oocytes were continuously perfused with Ringer buffer, and the test compounds were applied in the perfusate. Experiments were performed at room temperature and each data point represents the mean ± S.E.M. value of recordings performed on at least two oocytes from at least two different batches of oocytes. The recorded baseline-to-peak current amplitudes were analyzed using Clampfit 10.1 (Axon Instruments, Union City, CA, USA). Analogously to the procedures used in a recent study [43], the incorporation of the γ 2S subunit into the GABA A Rs assembled at the cell surface of α 1,2,3,5 β 2 γ 2S -expressing oocytes was confirmed on a routinely basis with 100 μM ZnCl 2 [46]. The presence of δ in cell surface-expressed receptors in α 6 β 2 δ-injected oocytes was confirmed on a routinely basis using the δ-GABA A R selective PAM DS2 (1 μM) [47]. Furthermore, taking advantage of the differential sensitivity of αβ and αβδ receptors to Zn 2+ as antagonist [45,[48][49][50], 1 μM ZnCl 2 was used to verify that a homogenous population of ternary α 6 β 2 δ receptor complexes was expressed in these oocytes.
The GABA EC 20 (the GABA concentration eliciting 20% of the maximum effect) for each receptor subtype was determined on two oocytes on each day of the experiment. A maximum concentration of GABA was applied until the peak of the response was observed, usually within 30 seconds. When two consecutive applications of the maximum GABA concentration were observed to elicit responses of similar effect (± 5%), 3-4 different concentrations of GABA were applied to the perfusate until the peak of the response was observed. Two to five minutes of wash time between each application were allowed to prevent receptor desensitization. Data for the GABA concentrations was normalized to the maximal response elicited by GABA on each oocyte, and the concentration-response curves were fitted in Prism GraphPad 5.0a (GraphPad Software, Inc. La Jolla, CA, USA) by nonlinear regression using the equation for sigmoidal dosage-response with variable slope (Equation Bottom = response at the bottom plateau; EC 50 = concentration giving rise to 50% of the maximum response; Top = response at the top plateau; X = logarithm of the concentration; Y = response. The GABA EC 20 response was calculated using Equation 2 (F = 20), and this concentration was subsequently used on a given experimental day. (2) log EC 50 The functional characteristics of clobazam, N-desmethylclobazam, clonazepam, diazepam, and zolpidem on the GABA A Rs were determined by co-application of different concentrations of the compounds with GABA EC 20 . The test compounds were pre-applied 30 seconds prior to the co-application with GABA EC 20 . At the end of an experiment, a maximum concentration of GABA was applied in the perfusate to determine the maximum response elicited by GABA through the receptor, which served as the internal standard and as a control of any potential drift in the system during the recordings. Two to five minutes of wash time between each application were permitted to overcome receptor desensitization. Data for the benzodiazepines was normalized to the responses elicited by GABA EC 20 at the receptor (the EC 20 response was defined as 100%). Concentration-response curves for the benzodiazepines were fitted using Equation 1.

Data analysis
Using GraphPad Prism 4 (GraphPad Software, Inc., La Jolla, CA, USA), the pEC 50 and I max values were evaluated for statistical differences across the receptor subtypes per compound using one-way ANOVA with Tukey's Multiple-Comparison Post-hoc Test, where P<0.05 was considered significant.

Results
Functional characterization of GABA and determination of GABA EC 20 values at human α 1 β 2 γ 2S , α 2 β 2 γ 2S , α 3 β 2 γ 2S , α 5 β 2 γ 2S and α 6 β 2 δ GABA A Rs expressed in Xenopus oocytes. Prior to the functional characterization of the benzodiazepines at the five GABA A R subtypes, the pharmacological properties of GABA at the receptors were determined. The concentration-response relationships displayed by GABA at the receptors are given in Fig. 2, and the pharmacological data are summarized in Table 1. The EC 50 and Hill slope values determined for the agonist were in good agreement with those observed in previous studies of these five receptors expressed in Xenopus oocytes [46,51].
EC 50 values are given in nM with pEC 50 ± S.E.M. values in brackets, I max values are given as % of the response evoked by GABA EC 20 at the receptor, and the numbers of experiments (N) are also given.

Receptor
EC 50 [pEC 50 Table 3. Statistical analysis of the functional properties of clobazam, N-desmethylclobazam, and clonazepam at the human P-values from one-way ANOVA testing if mean values (pEC 50 and I max ) experimentally determined for the compounds with Tukey's Multiple Comparison (P<0.05). As for the functional properties of clonazepam at the four receptors, only a significant difference was identified for its I max values at α 3 β 2 γ 2S and α 1 β 2 γ 2S (P<0.05).
Investigations into the benzodiazepine-mediated modulation of the α 1,2,3,5 β 2 γ 2S GABA A Rs As will be outlined in the Discussion section, the modulatory efficacies exhibited by benzodiazepines at αβγ GABA A Rs in recombinant expression systems in previous studies have varied considerably. To address this aspect, we compared the maximum responses mediated by saturating concentrations (3 μM) of clobazam, N-desmethylclobazam, and clonazepam at the GABA A Rs with the maximal responses mediated by a saturating concentration of diazepam, a 1,4-benzodiazepine often referred to as a "full benzodiazepine agonist". The maximum responses evoked by clobazam and N-desmethylclobazam when co-applied with GABA EC 20 at α 1 β 2 γ 2S and α 2 β 2 γ 2S GABA A Rs, respectively, were slightly but significantly smaller than those mediated by diazepam at the respective receptors (p<0.1; ordinary one-way ANOVA) ( Table 2, Figs. 3B and 4A). In contrast, the maximum responses mediated by clobazam at α 2,3,5 β 2 γ 2S , by N-desmethylclobazam at α 1,3,5 β 2 γ 2S and by clonazepam at all the four receptors did not differ significantly from those exhibited by diazepam ( Table 2, Figs. 3B and 4A). We also compared the maximum degree of potentiation mediated by clobazam at the α 2 β 2 γ 2S GABA A R with the maximal responses induced by a saturating concentration of zolpidem. Albeit not a benzodiazepine, zolpidem acts as a PAM through the high-affinity benzodiazepine site in α 1,2,3,5 βγ GABA A Rs, and the compound has previously been reported to be a "full benzodiazepine-site agonist" at the α 2 β 2 γ 2 subtype [52]. In this experiment cRNA for α 2 β 2 γ 2S was injected in a ratio of 1:1:5 (α 2 :β 2 :γ 2S ) to facilitate the expression of a homogenous population of γ 2S -containing complexes. The modulatory efficacies displayed by clobazam and zolpidem at the α 2 β 2 γ 2S receptors in these experiments did not differ significantly (Fig. 4B). Moreover, the respective degrees of potentiation mediated by clobazam and zolpidem at the receptors expressed in these oocytes did not differ significantly from those at oocytes injected with an α 2 : β 2 :γ 2S cRNA ratio of 1:1:1 ( Fig. 4B and data not shown).
In another series of experiments, we took advantage of the well-documented ability of zinc to discriminate between αβand αβγ GABA A Rs [46,53,54] to investigate whether oocytes injected with α 1,2,3,5 β 2 γ 2S cRNAs in a 1:1:1 subunit ratio express homogeneous populations of ternary receptor complexes. As can be seen from Fig. 5, the GABA EC 80 -evoked response through the α 1 β 2 GABA A R was almost completely eliminated by 100 μM Zn 2+ , whereas the presence of this concentration of the metal ion had negligible effect on the currents elicited in α 1 β 2 γ 2S -expressing oocytes. This strongly suggests that the functional properties of clobazam, N-desmethylclobazam and clonazepam at the α 1,2,3,5 β 2 γ 2S GABA A Rs have been determined at homogenous populations of γ 2S -containing receptors.
Functional properties of clobazam, N-desmethylclobazam, and clonazepam at the human α 6 β 2 δ GABA A R expressed in Xenopus oocytes To investigate whether the functional properties of clobazam and its metabolite at αβδ GABA A Rs potentially differ from those of clonazepam, the three compounds were tested at a representative of these receptors, the human α 6 β 2 δ subtype ( Table 2). Representative traces recorded for clobazam, N-desmethylclobazam and clonazepam when pre-incubated and co-applied with GABA EC 20 at the receptor are presented in Fig. 6A, and concentration-response relationships determined for the compounds at the receptor are given in Fig. 6B.
All three modulators potentiated GABA EC 20 -mediated signaling in α 6 β 2 δ-oocytes in a concentration-dependent and reversible manner, exhibiting EC 50 values in the mid-to-high micromolar range at the receptor (Table 2). Thus, the modulatory potencies of clobazam, Ndesmethylclobazam and clonazepam at this receptor were 200-400, 1100-3100 and 1400-3900 fold lower than those exhibited by the respective benzodiazepines at the α 1,2,3,5 β 2 γ 2S GABA A Rs, respectively (Table 2). Strikingly, however, the modulatory efficacies of the two 1,5-benzodiazepines at this receptor were very different from those at the four α 1,2,3,5 β 2 γ 2S receptors. Whereas the degree of potentiation of the GABA EC 20 -evoked response through α 6 β 2 δ mediated by clonazepam was comparable to those at the α 1,2,3,5 β 2 γ 2S GABA A Rs, clobazam and in particular N-desmethylclobazam were much more efficacious PAMs at the receptor (Table 2, Fig. 6). Albeit the concentration-response relationship for N-desmethylclobazam at the α 6 β 2 δ receptor was not saturated within the concentration range tested, 1 mM of the drug was found to potentiate the GABA EC 20 -evoked response through the receptor by *24-fold ( Fig. 6A and 6B). Interestingly, a distinct inhibition phase in the concentration-response relationship was observed in some of these recordings for clobazam and clonazepam, with 1 mM of the modulator resulting in a lower degree of potentiation than 300 μM (Fig. 6A). Moreover, rebound currents were observed at these high modulator concentrations (Fig. 6A).
To elucidate the molecular basis for the high-efficacious modulation exerted by N-desmethylclobazam at the α 6 β 2 δ GABA A R, we characterized the functional characteristics of the compound at the binary α 6 β 2 receptor. N-desmethylclobazam also potentiated the GABA EC 20 -mediated signaling through this receptor in a concentration-dependent manner, exhibiting modulatory potency and efficacy not significantly different from those displayed at α 6 β 2 δ (Table 2, Fig. 6B). However, it should be stressed that the EC 50 and I max values given for N-desmethylclobazam at α 6 β 2 δ and α 6 β 2 are estimates, since neither of the concentration-response curves for the modulator at these two receptors reached saturation.

Discussion
In view of the clinical use of clobazam for the treatment of various diseases over the last decades, surprising little is known about the in vitro pharmacology of the drug. In the present study, we have performed an elaborate functional characterization of clobazam, its major active Functional Properties of Clobazam at GABAA Receptors metabolite N-desmethylclobazam and the 1,4-benzodiazepine clonazepam at the human α 1 β 2 γ 2S , α 2 β 2 γ 2S , α 3 β 2 γ 2S , α 5 β 2 γ 2S and α 6 β 2 δ GABA A Rs expressed in Xenopus oocytes. A detailed functional characterization of clonazepam at recombinant GABA A Rs has to our Functional Properties of Clobazam at GABAA Receptors knowledge not been published previously, and thus this study provides substantial insights into the molecular pharmacology of all three modulators.
Each of the three benzodiazepines exhibited similar EC 50 values as a PAM at the four α 1,2,3,5 β 2 γ 2S GABA A Rs, and the maximum responses evoked by saturating concentrations of each of the modulators upon co-application with GABA EC 20 at these four subtypes were also very similar (Table 2). Although statistical analysis identified differences between some of the potencies and efficacies displayed by the respective compounds at the four receptors (Table 3), these differences are not considered pertinent from a biological perspective. It is important to stress that the α 1,2,3,5 β 2 γ 2S receptors included in this study only constitute a selection of all αβγ receptors targeted by benzodiazepines in clinically relevant concentrations. However, the fact that benzodiazepines have been reported to be less potent and less efficacious modulators of αβγ 1 and αβγ 3 receptors than of the corresponding αβγ 2 receptors [55][56][57] combined with the very restricted expression of γ 1 and γ 3 in the CNS [58,59] strongly suggests that the contributions of these receptor assemblies to the overall clinical effects of benzodiazepines are negligible. More importantly, the identity of the β subunit in the αβγ 2 GABA A R complex is not believed to influence benzodiazepine pharmacology substantially [56,57], just as there to our knowledge are no reports of benzodiazepines exhibiting significantly different pharmacological properties at γ 2S -and γ 2L -containing GABA A Rs. Thus, although we can not exclude the possibility that the functional properties of clobazam, N-desmethylclobazam and/or clonazepam at α 1,2,3,5 βγ 2 complexes comprising β 1 β 3 and/or γ 2L subunits could differ from those observed at the receptors in this study, we propose that the three benzodiazepines are likely to act as nonselective PAMs at all α 1,2,3,5 βγ 2 receptors.
In our recent study of the binding properties of the three benzodiazepines, clobazam and Ndesmethylclobazam displayed slightly but significantly higher binding affinities at the α 2 β 2 γ 2S GABA A R compared to the α 1 β 2 γ 2S subtype (2.5-and 4.3-fold, respectively), whereas clonazepam exhibited significantly higher binding affinities to α 1,2,5 β 2 γ 2S subtypes than to the α 3 β 2 γ 2S receptor (2.8-to 3.4-fold) [41]. The fact that these binding subtype-preferences are not mirrored in the functional profiles of the modulators is not particular surprising, given the different methodologies used to assess binding affinities and functional potencies of modulators. In support of this, another interesting observation that can be extracted from the two studies is that the 200-1000 fold higher binding affinities displayed by clonazepam compared to clobazam and N-desmethylclobazam at the four α 1,2,3,5 β 2 γ 2S receptors in the [ 3 H]flumazenil binding assay translate into considerably less pronounced differences (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19) fold) between the EC 50 values of the 1,4-benzodiazepine and the two 1,5-benzodiazepines at the receptors in the oocyte recordings (Table 2) [41]. These observations may reflect differences in the degrees of receptor desensitization in the two assays.
A comprehensive literature search has only identified two previous studies of the functional properties of clobazam and N-desmethylclobazam at recombinant GABA A Rs [55,60]. Clobazam has been reported to exert negligible potentiation of GABA EC 5 -EC 10 -evoked responses through α 1 β 2 γ 1 receptors in Xenopus oocytes, suggesting that the presence of this γ subunit in the GABA A R complex alters the functionality of the 1,5-benzodiazepine substantially [55]. Of particular interest for this study, Fisher has performed a direct comparison of the modulation mediated by clonazepam, clobazam, and N-desmethylclobazam on the current elicited by 3 μM GABA (EC 10 -EC 20 ) through the rat α 3 β 3 γ 2L GABA A R expressed in HEK-293T cells by patch clamp electrophysiology [60]. The EC 50 values determined for clonazepam, clobazam, and Ndesmethylclobazam at the rat α 3 β 3 γ 2L receptor in the Fisher study were 90 nM, 490 nM and 550 nM, respectively, which are in good agreement with the potencies exhibited by three benzodiazepines at the human α 3 β 2 γ 2S receptor in this study (Table 2) [60]. However, Fisher found clobazam and diazepam to be substantially more efficacious potentiators of rat α 3 β 3 γ 2L currents (I max values of 487% and 508% of the responses evoked by GABA EC 10 -EC 20 , respectively) than N-desmethylclobazam and clonazepam (I max values of 270% and 263%, respectively) [60]. Whereas the modulatory efficacies of N-desmethylclobazam and clonazepam at α 3 β 3 γ 2L are in good agreement with those at the α 3 β 2 γ 2S receptor in the present study, the higher efficacies exhibited by clobazam and diazepam in the Fisher study clearly contrast the comparable I max values determined for the four PAMs at the α 3 β 2 γ 2S receptor in this study ( Table 2, Figs. 3 and 4). Several factors might explain this apparent discrepancy, including the different receptors studied (rat α 3 β 3 γ 2L vs. human α 3 β 2 γ 2S ), the different expression systems (HEK-293T cells vs. Xenopus oocytes), and the different recording techniques (patch clamp vs. TEVC electrophysiology). We propose that the precise determination of the GABA EC 20 used for the characterization of the benzodiazepines in the present study can have facilitated a more precise determination of the degree of maximum potentiation of the GABA-evoked responses than the GABA EC 10 -EC 20 concentrations employed in the Fisher study. On the other hand, the faster application rate of the compounds in patch clamp recordings using HEK-293T cells may have resulted in less concomitant desensitization of the receptors during application than in the Xenopus oocyte recordings system, which is characterized by slower exchange rates. Thus, we cannot exclude the possibility that concurrent desensitization of the receptors upon co-application of GABA EC 20 with high benzodiazepine concentrations could constitute a ceiling effect with respect to the degree of maximum response elicited by the benzodiazepinebound receptor, and that this effect potentially can have masked putative differential efficacies of the benzodiazepines at the receptors. However, as will be outlined below, the divergent efficacies reported for benzodiazepines at GABA A Rs expressed in oocytes in the literature strongly suggest that establishing a "true efficacy" for any given benzodiazepine is not trivial.
In the absence of other previous studies of the functional characteristics of clobazam, N-desmethylclobazam and/or clonazepam at GABA A Rs expressed in Xenopus oocytes, we investigated whether the determined modulatory efficacies for these compounds could be considered reliable by comparing them to the efficacies mediated by two reference benzodiazepine-site modulators. A saturating concentration of the prototypic benzodiazepine diazepam was observed to induce comparable or slightly higher responses than those mediated by saturating concentrations of clobazam, N-desmethylclobazam and clonazepam at the four α 1,2,3,5 β 2 γ 2S subtypes (Figs. 3B and 4A). Furthermore, the maximum responses induced by zolpidem at α 1 β 2 γ 2S and α 2 β 2 γ 2S receptors were comparable to those mediated by clobazam ( Fig. 4B and data not shown). Unfortunately, the efficacies determined for diazepam and zolpidem at GABA A Rs expressed in oocytes in previous studies have varied considerably. Although often referred to as a "full benzodiazepine agonist," diazepam has exhibited very different efficacies as a PAM of GABA EC 10 -to EC 20 -evoked currents through α 1 β 2 γ 2 receptors in previous studies, including degrees of potentiation in the same 2.3-to 3.4-fold range as exhibited by the modulator at the α 1,2,3,5 β 2 γ 2S receptors in this study [61][62][63]. Likewise, whereas zolpidem has been reported to potentiate the GABA EC 15 -to EC 20 -evoked responses through α 1 β 2 γ 2 and α 2 β 2 γ 2 receptors 4-to 6-fold in some studies [52,64,65], its maximum modulation of GABAevoked currents through the receptors in other studies have been very similar to the 2.7-fold potentiation observed at α 2 β 2 γ 2S in this study [61,62,66]. In conclusion, the results in this study indicate that clobazam and N-desmethylclobazam are equally efficacious or almost as efficacious as the 1,4-benzodiazepines clonazepam and diazepam at the human α 1,2,3,5 β 2 γ 2S receptors. However, considering the variation in the efficacies reported for standard benzodiazepines in the literature, the absolute degrees of potentiation exerted by the two 1,5-benzodiazepines at the receptors in other recording set-ups could potentially differ from those observed in this study.
The α 6 β 2 δ receptor was included in this study as a representative of the δ-containing GABA A Rs, and clobazam, N-desmethylclobazam and clonazepam all displayed mid-to-high micromolar potencies as PAMs of this receptor (Fig. 6, Table 2). Apart from the obvious absence of an α (+) /γ (-) subunit interface in the α 6 β 2 δ complex, the distinct functional characteristics exhibited by the three benzodiazepines at this receptor support the notion of them acting through another allosteric site than the classical high-affinity benzodiazepine binding site, and the comparable potencies displayed by N-desmethylclobazam as a PAM at the α 6 β 2 and α 6 β 2 δ receptors strongly suggest that this site is comprised within the αβ regions of the αβδ complex (Fig. 6B, Table 2). Some of the characteristics displayed by the benzodiazepines at the α 6 β 2 δ GABA A R could be argued to be indicative of a binding site in the transmembrane domain of the receptor. The tendencies towards bell-shaped concentration-response curves observed for clobazam and clonazepam as well as the rebound currents observed at high concentrations of the modulators are certainly reminiscent of the characteristics previously reported for PAMs/ ago-PAMs acting through transmembrane domains of GABA A Rs [55,[67][68][69][70][71][72]. Moreover, whereas PAMs targeting extracellular non-cannonical subunit interfaces in GABA A Rs and other Cys-loop receptors historically predominantly have been found to increase agonist potency without affecting the maximal agonist response through the receptors significantly [20,73,74], several PAMs acting through the transmembrane domains of the receptors have been shown capable of increasing agonist-evoked maximal peak currents, analogously to the highefficacious potentiation of α 6 β 2 δ signaling mediated by N-desmethylclobazam [67,[74][75][76]. However, this phenotypic difference between PAMs acting through extracellular and transmembrane regions of Cys-loop receptors does not appear to be black-and-white, as illustrated by the pronounced enhancement of GABA efficacy at the α 1 β 3 δ GABA A R mediated by LAU 177 through a site in the extracellular α 1 (+) /β 3 (-) interface [22] and by the augmentation of agonist efficacy exerted by the anthelmintic drug morantel through the extracellular β 2 (+) /α 3 (-) interface of the α 3 β 2 nicotinic acetylcholine receptor [77]. Thus, not having explored the modes of action of clobazam, N-desmethylclobazam and clonazepam at the α 6 β 2 δ GABA A R in detail, we will refrain from speculations about whether the modulators target a low-affinity binding site located in the transmembrane domain, in the extracellular α (+) /β (-) subunit interface, or in another region of the receptor complex. The nature and efficacies of the modulation exerted by clobazam and N-desmethylclobazam at other αβ and αβδ receptors could potentially differ from those observed for the compounds at α 6 β 2 and α 6 β 2 δ. However, the two 1,5-benzodiazepines are likely to be weak modulators at all αβand αβδ GABA A Rs, thus mediating their effects at these receptors at substantially higher concentrations than those required to modulate αβγ receptors. Although the μM activity component of benzodiazepines previously has been proposed to contribute to the CNS depression observed at high in vivo concentrations [18], we find it highly improbable that this component contributes significantly to the clinical efficacy of clobazam/N-desmethylclobazam. On the other hand, the highly efficacious potentiation of α 6 β 2 δ GABA A R signaling exerted by N-desmethylclobazam is quite interesting from a molecular perspective, and we propose that the 1,5-benzodiazepine could constitute an interesting lead scaffold for the development of novel allosteric modulators targeting this elusive low-affinity benzodiazepine binding site in the GABA A Rs.

Conclusion
The present study represents the first elaborate in vitro functional characterization of clobazam, N-desmethylclobazam and clonazepam at recombinant human GABA A Rs. While both 1,5-benzodiazepines are potent PAMs of human α 1,2,3,5 β 2 γ 2S receptors, they appear to be non-selective both in terms of potency and efficacy, a characteristic they share with many classical 1,4-benzodiazepines including clonazepam (Table 2) [15,16]. Obviously we can not completely exclude the possibility that the functionalities of clobazam and/or N-desmethylclobazam at one or a few of the GABA A R subtypes not included in this study could differ substantially from those of clonazepam and other 1,4-benzodiazepines, and that this difference could contribute to the distinct properties observed for clobazam compared with the classical 1,4benzodiazepines in the clinic. However, judging from the results in this study we propose that these clinical differences are more likely to be rooted in other factors than the in vitro pharmacological properties of the modulators, such as their respective pharmacokinetic characteristics.