Topiramate modulation of β1- and β3-homomeric GABAA receptors

Pharmacological Research 64 (2011) 44–52

Contents lists available at ScienceDirect

Pharmacological Research journal homepage: www.elsevier.com/locate/yphrs

Topiramate modulation of ␤1 - and ␤3 -homomeric GABAA receptors Timothy A. Simeone a,1 , Karen S. Wilcox a,b , H. Steve White a,b,∗ a b

Interdepartmental Program in Neuroscience, University of Utah, Salt Lake City, UT 84112, USA Department of Pharmacology and Toxicology, University of Utah, Salt Lake City, UT 84112, USA

a r t i c l e

i n f o

Article history: Received 15 July 2010 Received in revised form 10 March 2011 Accepted 14 March 2011 Keywords: GABAA receptor Beta homomer Topiramate

a b s t r a c t The broad spectrum anticonvulsant topiramate modulates multiple voltage-gated and ligand-gated channels, including ␥-aminobutyric acid type A (GABAA ) receptors. Previously, we found a strong ␤-subunit influence on the effects of topiramate on heteromeric GABAA receptors. Here, we tested the hypothesis that homomeric GABAA receptors comprised of either ␤1 - or ␤3 -subunits will contain a functional binding site for topiramate. For comparison, we also examined the effects of pentobarbital and loreclezole which exhibit ␤-subunit dependence as well. We expressed ␤1 - and ␤3 -homomeric receptors in Xenopus laevis oocytes and acquired electrophysiological responses using two-electrode voltage clamp techniques. Oocytes expressing ␤-homomers were insensitive to GABA and had hyperpolarized resting membrane potentials, decreased input resistances, increased holding currents and picrotoxin-induced outward currents consistent with the expression of non-ligand-mediated, spontaneous channel openings of ␤-homomers. Similar to picrotoxin, application of topiramate, pentobarbital and loreclezole inhibited ␤1 -homomers. In contrast, these compounds activated ␤3 -homomers. As with heteromeric receptors, topiramate and pentobarbital modulation of ␤1 - and ␤3 -homomers exhibited rebound currents indicating an open channel block or stabilization of desensitization. Interaction studies suggested competition between topiramate, loreclezole and pentobarbital for activation of ␤3 -homomers, whereas topiramate inhibitory actions were non-competitive with pentobarbital but competitive with loreclezole. In summary, ␤1 - and ␤3 -subunits have binding site(s) for topiramate that elicit functional effects with similarities to heteromeric receptor responses. From this foundation, contributions of residues and other subunits in binary and ternary heteromeric receptors can be explored to gain a complete understanding of topiramate actions on complex heteromeric GABAA receptors. © 2011 Elsevier Ltd. All rights reserved.

1. Introduction Topiramate (2,3:4,5-Di-O-isopropylidene-b-d-fructopyranose sulfamate) is a broad spectrum anticonvulsant that modulates the function of multiple voltage-gated and ligand-gated channels, including ␥-aminobutyric acid type A (GABAA ) receptors [11,14]. GABAA receptors mediate the majority of fast, inhibitory neurotransmission in the central nervous system and are composed of multiple protein subunits (i.e., ␣, ␤, ␥, ␦, ␧, ␪, ␲, and ␳) in a proposed pentameric arrangement [13]. The majority of native GABAA receptors are thought to contain either ␣ and ␤ or ␣, ␤, and ␥ subunits with uncertain stoichiometry [13]. Receptor subunit composition has tremendous effects on the

∗ Corresponding author at: University of Utah, Anticonvulsant Drug Development Project, Department of Pharmacology and Toxicology, 417 Wakara Way, Suite 3211, Salt Lake City, UT 84108, USA. Tel.: +1 801 581 6447; fax: +1 801 581 4049. E-mail address: [email protected] (H.S. White). 1 Present address: Department of Pharmacology, Creighton University School of Medicine, Omaha, NE 68178, USA. 1043-6618/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.phrs.2011.03.004

physiological and pharmacological responses of GABAA receptors [13]. We have previously reported that the effects of topiramate greatly depend on the subunit composition of GABAA receptors [14]. Specifically, topiramate directly activates and potentiates heteromeric GABAA receptors containing either ␤2 - or ␤3 -subunits, whereas receptor combinations containing ␤1 -subunits are either slightly potentiated or inhibited [14]. These studies suggest a ␤subunit dependence of topiramate and the possibility of a binding site on ␤-subunits. Recombinant expression models have demonstrated that functional homomeric receptors may be formed by the ␤1 and ␤3 subunit [2–4,7,18,19,29,30]. ␤-Homomeric receptors provide the opportunity to investigate both the specificity of pharmacological compounds and the structure/function of the ␤ subunits without the complications associated with receptor heterogeneity [19]. The current experiments took advantage of this reduced GABAA receptor expression system to test the hypothesis that topiramate has binding sites on ␤1 - and ␤3 -subunits that exhibit functional effects that may be related to effects on heteromeric GABAA receptors. For comparison, we also examined the effects of other

T.A. Simeone et al. / Pharmacological Research 64 (2011) 44–52

compounds that exhibit ␤ subunit-dependence, pentobarbital and loreclezole. 2. Materials and methods 2.1. Oocyte isolation and injection As previously described [14,36], oocytes were removed from Xenopus laevis frogs anesthetized by tricaine (2 mg ml−1 ). Harvested ovarian lobes were defolliculated by incubation in 2 mg ml−1 of collagenase (Type IA) for 1–2 h at room temperature on an orbital shaker in calcium-free ND-96 solution containing in mM: 96 NaCl, 2 KCl, 1 MgCl2 , and 5 HEPES (pH = 7.4). The oocytes were rinsed 5–6 times with a Barth’s solution [in mM: 88 NaCl, 1 KCl, 0.41 CaCl2 , 0.33 Ca (NO3 )2 , 1 MgSO4 , 2.4 NaHCO3 , and 10 HEPES (pH = 7.4)], and selected stage V–VI oocytes were stored at 18.5 ◦ C in a Barth’s solution supplemented with 1 mM Na-Pyruvate (Sigma, St. Louis, MO), 0.01 mg ml−1 gentamycin (Sigma, MO), and an antibiotic–antimycotic solution containing 100 units ml−1 of penicillin, 100 ␮g ml−1 streptomycin and 0.25 ␮g ml−1 of Amphotericin B (Invitrogen, Carlsbad, CA). All procedures involving animals were in accordance with National Institutes of Health guidelines and were approved by the University of Utah Institutional Care and Use Committee. Circular pCIS2 plasmids containing cDNA encoding the rat ␤1 and ␤3 subunits were generously provided by Dr. Roy E. Twyman. Glass capillary tubes (World Precision Instruments, Sarasota, FL) were pulled to a fine tip on a micropipette puller (Sutter Instrument CO., Novato, CA) and broken back to an outside diameter of 21 ␮m. ␤1 and ␤3 subunit cDNA stocks were diluted to 0.1–0.4 ng nl−1 and injected into the nucleus of the oocyte (∼37 nl) with a nano-injector (World Precision Instruments) 24-h after isolation. 2.2. Electrophysiology Electrophysiological recordings were performed 1–4 days following injection and were conducted at room temperature (20–22 ◦ C) in a 100 ␮l chamber continuously perfused (6 ml/min) with a Ringer’s solution containing in mM: 115 NaCl, 2.5 KCl, 1.0 BaCl2 , and 10 HEPES (pH = 7.4). Two-electrode voltage-clamp recordings were obtained with a GeneClamp 500 amplifier (Axon Instruments, Union City, CA) using 3 M KCl-filled microelectrodes (1–5 M). Recordings were performed at a holding potential of −60 mV and captured with pClamp6 Fetchex and Clampex software (Axon Instruments). Pentobarbital is known to modulate and directly activate heteromeric and ␤-homomeric GABAA receptors [2,3,15]; therefore, 1 mM pentobarbital was initially applied to assess expression of ␤-homomers. If currents were not elicited with pentobarbital, a saturating concentration of picrotoxin (100 ␮M) was applied to confirm the lack of functional expression of spontaneously opening receptors. Xenopus oocytes expressing ␤1 - or ␤3 -homomeric receptors had significantly hyperpolarized resting membrane potentials (Vrest ), decreased input resistances (Rinput ), and required larger holding currents (Ihold ) to clamp the oocyte at −60 mV compared to oocytes that were sham-injected with water [Vrest (mV) = −17 ± 0.5 (sham), −33.0 ± 0.9 (␤1 ), −28.5 ± 1.2 (␤3 ); Rinput (M) = 0.64 ± 0.07 (sham), 0.12 ± 0.02 (␤1 ), 0.21 ± 0.04 (␤3 ); Ihold (nA) = −74 ± 9 (sham), −397 ± 38 (␤1 ), −312 ± 37 (␤3 ); n = 9, 44, 44; p < 0.001 vs. sham-injected oocytes with an one way ANOVA and an all pair-wise multiple comparison using the Holm-Sidak method]. These observations were similar to the lower input resistances of ␤-homomer expressing oocytes reported previously [7,12,18,19] and suggest an increased Cl− leak conductance. The ␤1 -homomers did not respond to 1 mM GABA (0/44 oocytes)

45

and ␤3 -homomers were relatively unresponsive when exposed to 1 mM GABA [i.e., 5/44 oocytes exhibited comparatively small inward currents (<25 nA)] (data not shown). Note: Responses of pharmacological compounds may vary between non-mammalian and mammalian expression systems; however, murine and rat (used in this study) ␤1 - and ␤3 -homomers have been shown to have functional expression in Xenopus oocytes and mammalian cells [2,4,7]. 2.3. Analysis of membrane currents The current magnitudes were measured from digitized traces in pClamp6 Fetchan or pClamp8 Clampfit (Axon Instruments). Concentration–response curves were constructed with Prism 4.0 software (Graphpad, San Diego, CA) and the equation I = Imin + (Imax − Imin ){[1/(1 + (EC50 /[agonist])nH ], where [agonist] is the drug concentration, EC50 is the concentration of drug eliciting a half-maximal response, and nH is the Hill coefficient. IC50 (i.e., concentration of drug eliciting a half-maximal inhibition) values were determined from fits to a logistic equation of similar form. Data are reported, unless stated otherwise, as the mean ± SEM; n is the number of cells tested. Statistical significance was assessed with Student’s paired or unpaired t-test or one way ANOVA as appropriate. Current–voltage (I–V) relationships were determined as previously described [14]. The theoretical chloride equilibrium potential was calculated using the intracellular ion concentrations in X. laevis oocytes measured by Barish [1] (Cl− : 33.4 ± 1.2 mM) and the Nernst equation (ECl − = −32.4 ± 0.9 mV). Data were analyzed using pClamp8 Clampfit software (Axon Instruments). Compound interactions during inhibition of ␤1 -homomers were quantified using a method described by Wollmuth [20]. Block by 100 ␮M picrotoxin was considered as maximal—thus block by pentobarbital, topiramate, and loreclezole were normalized to the block by 100 ␮M picrotoxin. The IC50 values for pentobarbital and topiramate calculated by the above equation were used, whereas the IC50 value for loreclezole (0.223 mM) was estimated with the Langmuir isotherm (assuming the law of mass action where mH = 1) of form y = 1/(1 + [loreclezole]/IC50 ), where y is the fraction of unblocked channels and IC50 is the concentration for half-maximal block. It was of interest to determine whether the blocking actions of the compounds were competitive or non-competitive. If they block competitively, then the total number of channels occupied by the blocking compounds would be the sum of those channels occupied by compound x and z in the presence of a competitive antagonist (z and x, respectively): occupiedx,z (competitive) = {1/[1 + (1/ x )(1 +  z )]} + {1/[1 + (1/ z ) (1 +  x )]} with  x and  z defined as [x]/KD(x) and [z]/KD(z) where KD = (IC50 )mH . In cocktails of x and z, the fraction of unblocked channels (1 − occupiedx,z ) would be: yx,z (competitive) = 1/(1 +  x +  z ). Alternatively, if they block at independent sites, the total number of channels occupied would be the sum of those channels occupied by x and z minus those channels occupied by both: occupiedx,z (non-competitive) = {1/[1 + (1/ x )]} + {1/[1 + (1/ z )]} − {1/[1 + (1/ x )]}{1/[1 + (1/ z )]}; and the fraction of unblocked channels would be: yx,z (noncompetitive) = 1/[(1 +  x )(1 +  z )]. 2.4. Drug solutions All drugs were prepared fresh before each recording session. A stock solution of 100 mM picrotoxin (Sigma) was prepared in DMSO (Sigma). GABAA receptor ␤-homomers were unaffected by 0.1% DMSO. On the day of experiments, the highest experimental doses of the following drugs were prepared in frog Ringer’s solution: 10 mM topiramate, 3 mM pentobarbital, and 100 ␮M lore-

46

T.A. Simeone et al. / Pharmacological Research 64 (2011) 44–52

Fig. 1. Inhibition of spontaneous channel openings of ␤1 - and ␤3 -homomeric receptors by picrotoxin. (A) Representative traces illustrating the concentration-dependent outward currents evoked by picrotoxin (PTX). The effects of picrotoxin were slow to wash out, therefore, increasing concentrations were consecutively applied. (B) Picrotoxin concentration–response curves for ␤1 -homomers (n = 7; filled circles; Hill coefficient: 0.96 ± 0.01) and ␤3 -homomers (n = 9; open circles; Hill coefficient: 0.68 ± 0.02). Data were normalized to inhibition observed upon application of 300 ␮M picrotoxin and are presented as the mean ± SEM. (C) Representative graphs depicting the picrotoxin I–V relationships of ␤1 - and ␤3 -homomers. Data points were obtained using a ramp protocol from −70 mV to +60 mV.

clezole. All drugs were serially diluted in Ringer’s solution to the lowest experimental concentration. 3. Results 3.1. Spontaneous activity of ˇ1 - and ˇ3 -homomeric receptors To determine the presence of spontaneously open channels, we applied the chloride-channel blocker picrotoxin to oocytes expressing either ␤1 - or ␤3 -homomers. Picrotoxin application resulted in concentration-dependent apparent outward currents indicating inhibition of the spontaneous activity of ␤1 - and ␤3 -homomeric receptors (Fig. 1A and B). The IC50 s of picrotoxin inhibition of ␤1 - and ␤3 -homomers were similar to those of other studies on ␤-homomeric and heteromeric GABAA receptors [12,18,19]. The non-liganded current blocked by 100 ␮M picrotoxin was signif-

icantly different between ␤1 - and ␤3 -homomers (225 ± 25 and 52 ± 8 nA, respectively; p ≤ 0.001) even though the holding currents required to clamp the oocytes at −60 mV were similar (−397 nA vs. −312 nA; sham Ih = −79 nA). This contradictory finding may reflect intrinsic differences of single channel properties [7,31] and/or efficacy of picrotoxin block of ␤1 - and ␤3 -homomers. Supporting the latter possibility, felbamate inhibits ␤3 -homomers up to 120% of picrotoxin inhibition, whereas ␤1 -homomers are inhibited to only 80% of picrotoxin inhibition (Simeone et al., unpublished). Furthermore, picrotoxin (up to 1 mM) exhibits only partial inhibition of the spontaneous openings of ␣1 ␤3 ␧ receptors [35]. The reversal potentials (Erev ) obtained from current–voltage (I–V) relationships were not significantly different from the value predicted by the Nernst equation and confirmed that the portion of leak current blocked by 100 ␮M picrotoxin was due to chloride ion flux in oocytes expressing either ␤1 - or ␤3 -homomers

T.A. Simeone et al. / Pharmacological Research 64 (2011) 44–52

47

Fig. 2. Inhibition of ␤1 -homomeric receptor spontaneous channel openings by pentobarbital, topiramate and loreclezole. (A) Traces of representative responses to each compound from individual oocytes are presented. (B) Concentration–response curves for pentobarbital (PTB; filled circles; Hill coefficient: 1.06 ± 0.11), topiramate (TPM; open circles; Hill coefficient: 1.00 ± 0.12) and loreclezole (LOR; filled squares) normalized to 100 ␮M picrotoxin (PTX) block (n = 6, 6 and 6, respectively).

(Fig. 1C). In fact, the Vrest of oocytes expressing ␤-homomers (see Section 2) was not significantly different from the theoretical ECl − (−32.4 mV) indicating a major contribution of Cl− to the leak current. These experiments demonstrate that rat ␤1 homomeric and ␤3 -homomeric receptors expressed in oocytes mediate non-ligand-mediated, spontaneous chloride currents that are insensitive to GABA (see Section 2), but are sensitive to picrotoxin. We compared topiramate actions to pentobarbital, known to gate ␤-homomers [2,3,19,29], and loreclezole, known for ␤-subunit dependence in heteromeric GABAA receptors [16,17,21]. Topiramate and pentobarbital have similar actions on heteromeric receptors, i.e., both compounds potentiate currents, directly activate receptors, exhibit rebound currents, and may compete for a shared transduction/gating mechanism [14,25]. Loreclezole also has barbiturate-like effects on heteromeric receptors except it lacks rebound currents [16,17]. We hypothesized that all of the compounds would modulate ␤-homomers, thus confirming functional binding sites on ␤-subunits. Moreover, we predicted that the responses would diverge from heteromeric receptor responses enabling deduction of subunit contributions and forming the basis for future structure/function experiments.

highest concentration of each compound [pentobarbital (3 mM), topiramate (10 mM), and loreclezole (100 ␮M)] blocked the chloride current by 87 ± 17% (n = 6), 85 ± 20% (n = 6) and 31 ± 3% (n = 6), respectively, relative to the block by 100 ␮M picrotoxin (Fig. 2B). A full concentration response was unattainable for loreclezole due to solubility limitations. Upon termination of applications of pentobarbital and topiramate, the current rapidly overshot the baseline and became a transient inward current that slowly returned to baseline (Fig. 2). These rebound currents were concentration-dependent, however, the maximum effect was not reached by the highest concentrations tested. Therefore, accurate EC50 s could not be calculated, but are estimated to be ∼1 mM and ∼4–5 mM for pentobarbital and topiramate rebound currents, respectively. The presence of these rebound currents suggests that the ␤1 -homomer passes through an open state upon relief from pentobarbital and topiramate inhibition. As with picrotoxin, I–V relationships for pentobarbital, topiramate, felbamate and loreclezole actions on oocytes expressing either ␤1 - or ␤3 -homomers confirmed that the portion of leak current modulated by each compound was carried by chloride ions and did not deviate significantly from the Nernst equation prediction (data not shown). Collectively, these results confirm the presence of functional binding sites on ␤1 -subunits.

3.2. Topiramate, pentobarbital and loreclezole inhibit ˇ1 -homomeric receptors Similar to picrotoxin, topiramate, pentobarbital and loreclezole inhibited spontaneous activity of the ␤1 -subunit homomers in a concentration-dependent manner (Fig. 2). At the end of these experiments, a saturating concentration of picrotoxin (100 ␮M) was applied in order to compare the efficacy of the compounds. The

3.3. Topiramate, pentobarbital and loreclezole activate ˇ3 -homomeric receptors It could not be determined whether the observed effects on ␤3 -homomer function were due to potentiation of spontaneous channel openings or activation of additional ␤3 -homomeric recep-

48

T.A. Simeone et al. / Pharmacological Research 64 (2011) 44–52

Fig. 3. ␤3 -Homomeric receptors are activated by pentobarbital, topiramate and loreclezole. (A) Traces of representative responses to each compound from individual oocytes are presented. (B) Concentration–response curves of the peak currents evoked by pentobarbital (PTB; filled circles; n = 4; Hill coefficient: 1.01 ± 0.04) and topiramate (TPM; open circles; n = 7; Hill coefficient: 0.97 ± 0.04). High concentrations of pentobarbital and topiramate resulted in reduced levels of current, revealing inhibitory effects. Therefore, currents were normalized to the largest peak currents evoked by 1 mM pentobarbital or 3 mM topiramate and fit with a modified Hill equation that considered both positive and negative modulations (see Section 2). The calculated IC50 s for pentobarbital and topiramate are 1.65 ± 0.61 mM and 8.59 ± 2.07 mM, respectively, with Hill coefficients of 1.08 ± 0.03 and 1.14 ± 0.07, respectively. Note in A the rebound currents following removal of pentobarbital and topiramate (arrows).

tors. Hence, the term ‘activation’ is used to describe positive modulations of ␤3 -homomer function. In contrast to the inhibitory effects on ␤1 -subunits, topiramate, pentobarbital and loreclezole activated ␤3 -homomers in a concentration-dependent manner (Fig. 3). Reduced activation was elicited by 1 mM and 3 mM pentobarbital and 10 mM topiramate and was followed by rebound currents. These rebound currents were similar to barbiturate and anesthetic actions on heteromeric GABAA receptors and may reflect relief from a low affinity channel block of ␤3 -homomers by pentobarbital and topiramate, or a transient passage from a desensitized state through an open state before channel closure [2–4,10,14,15,19]. If these rebound currents are viewed as revealing the full effect of pentobarbital and topiramate and are included in the concentration response [22,25], the EC50 s significantly decrease from 0.84 ± 0.06 to 0.55 ± 0.04 mM for pentobarbital and from 2.14 ± 0.22 to 0.94 ± 0.04 mM for topiramate (p < 0.01; paired t-test). Loreclezole activation did not saturate and concentrations higher than 100 ␮M were unattainable due to insolubility, thus the EC50 could not be calculated. Apparent receptor desensitization was observed with increasing concentrations similar to previous reports on heteromeric receptors and may have contributed to a concentration-dependent outward current upon washout reflecting slow recovery from a desensitized state [5,22]. This outward

rebound current of loreclezole (100 ␮M) constituted a 27 ± 8% (n = 4) block compared to 100 ␮M picrotoxin. It is estimated that the peak block by loreclezole would be ∼40–50% with an IC50 of ∼50–100 ␮M. However, these are crude estimates to be interpreted cautiously. Pentobarbital was the most efficacious of the compounds tested on Xenopus oocytes expressing ␤3 -homomers. Topiramate and loreclezole evoked currents that were a fraction of the current evoked by 100 ␮M pentobarbital [10 mM topiramate = 27 ± 9%, n = 8; and 100 ␮M loreclezole = 16 ± 5%, n = 4]. These results confirm the presence of functional binding sites on ␤3 -subunits. 3.4. Compound interactions Topiramate, pentobarbital and loreclezole modulated the ␤homomers in a similar manner, inhibiting ␤1 -homomers and activating ␤3 -homomers, suggesting the possibility of a shared mechanism of action. We probed interactions using sub-saturating concentrations of each compound. Co-application of 1 mM pentobarbital and either 1 mM topiramate or 100 ␮M loreclezole to ␤1 -homomers increased the pentobarbital block by ∼16–18%, whereas co-application of 100 ␮M loreclezole and 1 mM topiramate decreased the loreclezole block by ∼16% (Fig. 4). The method of Wollmuth [20] was used to determine the theoretical

T.A. Simeone et al. / Pharmacological Research 64 (2011) 44–52

49

Fig. 4. Compound interactions. (A) Inhibition of ␤1 -homomers by co-application of pentobarbital (PTB) and topiramate (TPM) (n = 5) or PTB and loreclezole (LOR) (n = 4) closely resembled values predicted by a non-competitive model of channel inhibition, whereas inhibition by co-application of TPM and LOR (n = 3) more closely conformed to a competitive model of channel inhibition. The dashed lines represent the calculated values of competitive (c; lower dashed line) and non-competitive (nc; upper dashed line) interactions of the respective compounds. Data presented as the mean ± SEM. (B) Activation of ␤3 -homomers by PTB was reduced upon co-application of either TPM or LOR (n = 3). LOR-mediated currents were reduced by TPM (n = 3) indicating possible competition for either a similar binding site or gating mechanism. Significantly different from PTB or LOR: *p < 0.05, **p < 0.001 by paired t-test.

percentage of channels that remain unblocked during competitive and non-competitive interactions of the compounds at these concentrations. According to theoretical calculations, a competitive block of ␤1 -homomers would result in 36%, 40% and 47% of channels remaining unblocked for the co-application combinations of pentobarbital–topiramate, pentobarbital–loreclezole and loreclezole–topiramate, respectively. A non-competitive block would result in fewer channels remaining un-blocked (i.e., 29% for pentobarbital–topiramate, 34% for pentobarbital–loreclezole and 41% for loreclezole–topiramate). Our experimental data indicated that the following percent of ␤1 -homomeric receptors remained unblocked: 28 ± 7% during pentobarbital–topiramate, 29 ± 11% during pentobarbital–loreclezole, and 62 ± 10% during loreclezole–topiramate (Fig. 4A). These data suggest a noncompetitive and additive interaction of pentobarbital–topiramate and pentobarbital–loreclezole, and a competitive interaction between topiramate and loreclezole. The loreclezole interactions must be viewed cautiously because the loreclezole block did not saturate and the IC50 was coarsely estimated with the Langmuir isotherm (see Section 2). Adjusting the loreclezole IC50 up to 1 mM increased the goodness of fit for the loreclezole–topiramate interaction suggesting that the loreclezole IC50 or the Hill coefficient or both were underestimated since both contribute to the loreclezole KD . [Note: Additional experiments are needed to reduce this source of error and full Schild analysis or mathematical modeling similar to Kindler et al. [6] should be performed before definitive conclusions can be drawn.] Consistent with a non-competitive additive interaction on receptor inhibition, pentobarbital–topiramate and pentobarbital–loreclezole applications increased the rebound current by 137 ± 88% and 27 ± 15%, respectively, compared to the rebound current of pentobarbital alone. In contrast, rebound currents that normally accompanied topiramate were suppressed with co-application of topiramate and loreclezole (data not shown). All three co-application combinations resulted in significant decreases of pentobarbital or loreclezole activation of ␤3 -homomers suggesting that topiramate, pentobarbital and loreclezole compete for either a shared binding site or gating mechanism (Fig. 4B). Additionally, co-application of topiramate increased the pentobarbital rebound current by 38 ± 17%, whereas loreclezole did not alter the pentobarbital rebound current.

Application of 1 mM topiramate alone did not induce rebound currents (Fig. 3). If the mechanism of rebound currents in ␤3 homomers is similar to block of ␤1 -homomers, this may reflect a non-competitive, additive inhibitory effect of pentobarbital and topiramate. 4. Discussion The present study confirmed functional binding sites for topiramate, pentobarbital and loreclezole on ␤-subunits and demonstrated that these compounds inhibit ␤1 -homomers and activate ␤3 -homomers. To our knowledge, this is the first report of specific modulation of ␤-homomers by topiramate and loreclezole, and supports previous studies of pentobarbital action [2,3,19,29]. 4.1. Predictions based on ˇ-homomeric receptor modulation Based on the current study, we predict that (1) compounds that activate ␤3 -homomers will directly activate and potentiate binary and ternary heteromeric receptors containing ␤3 -subunits and (2) compounds that inhibit ␤1 -homomers will inhibit or weakly potentiate binary and ternary heteromeric receptors containing ␤1 -subunits. Comparing published studies on recombinant heteromeric and endogenous neuronal GABAA receptors to our results (Table 1), the actions of topiramate and loreclezole are correctly predicted. In contrast, pentobarbital potentiates and directly activates all heteromeric receptors regardless of ␤-subunit. We speculate that these differences may be related to the partial agonistic actions of topiramate and loreclezole and the full agonistic actions of pentobarbital and as detailed below may reflect a greater contribution of additional subunit types to mediate the effects of pentobarbital. 4.2. Topiramate Topiramate modulation of heteromeric receptors is dependent on the ␤ subunit [14]. Topiramate directly activates and enhances ␣x ␤3 or ␣x ␤3 ␥2 receptor function, whereas ␣1,2,5 ␤1 ␥2S receptors are slightly enhanced and ␣1 ␤1 and ␣4,6 ␤1 ␥2S receptors are inhibited [14]. Topiramate has increased efficacy of direct activation of ␣4 ␤3 ␥2S and inhibition of ␣4 ␤1 ␥2S receptors, and minimal

50

T.A. Simeone et al. / Pharmacological Research 64 (2011) 44–52

Table 1 Comparison of compound modulatory potencies and efficacies on homomeric, heteromeric and native neuronal GABAA receptors.

Inhibition (i) and enhancement (e) of ␤-homomers were compared to 100 ␮M picrotoxin and 100 ␮M pentobarbital, respectively. Inhibition (i) and enhancement (e) of heteromeric and native receptors are relative to a control GABA response. N.D., not determined. Dashed lines separate reported data from different studies. Data for the native receptors was obtained from cultured and acutely dissociated cortical or hippocampal neurons. Topiramate and loreclezole inhibited or failed to potentiate GABAcurrents of a subset of neurons in each study. a This study. b Simeone et al. [14]. c Wafford et al. [16]. d Thompson et al. [15]. e Wingrove et al. [17]. f Fisher et al. [21]. g Simeone, unpublished observations. h Serafini et al. [22]. i Feng et al. [23]. j White et al. [24]. k Rho et al. [10]. l Tietz et al. [26]. m Kapur et al. [27]. n Saxena and Macdonald [28]. o Mangan et al. [9].

direct activation of ␣6 ␤3 ␥2S [14]. The effects of ␦-subunits on topiramate modulation are currently unknown, but based on pentobarbital studies [23,32], we suspect increased efficacy especially for ␣4 ␤3 ␦ receptors. The present findings suggest the ␤-subunit

possesses the primary topiramate binding site(s), and the ␣1,2,5 and ␥2 subunits may either contribute to the topiramate binding site or allosterically modulate the receptor response to topiramate.

T.A. Simeone et al. / Pharmacological Research 64 (2011) 44–52

4.3. Pentobarbital As mentioned, pentobarbital directly activates and/or enhances GABA-currents of every ␣␤ or ␣␤␥ receptor subunit combination including those expressing ␤1 -subunits [15] suggesting that the ␣-subunit promotes pentobarbital-mediated positive modulation, both potentiation and direct activation. Pentobarbital allosterically interacts with a binding pocket for etomidate, another intravenous general anesthetic which is also ␤ subunit-dependent, between the ␣1,2,3,5 and ␤1,2,3 subunits [8] indicating that ␣-subunits either contribute to a pentobarbital binding site or allosterically modulates the pentobarbital binding site or gating mechanism. Of the ␣-subunits, ␣6 confers the highest affinity and efficacy, whereas ␣4 lacks direct activation but retains potentiation [15,34]. Additionally, the ␦-subunit increases the apparent efficacy of both potentiation and direct activation by pentobarbital [23,32]. 4.4. Loreclezole In the current study, loreclezole inhibited the spontaneous activity of ␤1 -homomers and activated channel openings of ␤3 homomers. Similar to topiramate, loreclezole enhancement of GABA-currents and direct activation of heteromeric receptors are dependent on the ␤ subunit. The loreclezole potentiation of heteromeric receptors containing either ␤3 or ␤2 subunits has a ∼300-fold greater affinity than the potentiation of ␤1 -containing receptors (Table 1) [16,17]. Binary ␣␤1 receptors are potentiated by loreclezole indicating that the ␣-subunit may allosterically modulate existing ␤-homomer gating mechanisms. Interestingly, loreclezole inhibited GABAA -mediated tonic inhibitory currents of cultured hippocampal pyramidal neurons expressing ␣4 , ␤1 , and ␦ subunits [9] suggesting a lack of ␣4 and ␦ modulation of the ␤1 dependent loreclezole effects in this receptor combination, similar to our previous findings with topiramate effects on ␣4 ␤1 ␥2 receptors [14]. 4.5. Activation and inhibition of ˇ-homomeric receptors Topiramate and pentobarbital elicited similar responses from homomeric receptors, activating ␤3 -homomers and inhibiting ␤1 homomers with the accompaniment of rebound currents to both effects. Pentobarbital effects on ␤-homomers are dependent on residue 265 in the second transmembrane domain, which is a serine for ␤1 subunits and an asparagine for ␤3 subunits [2]. Reciprocal replacement of this residue results in pentobarbital activation of mutant ␤1 S265N-homomers and inhibition of mutant ␤3 N265S-homomers. This residue participates in ion channel transduction/gating rather than a pentobarbital binding site and is important for several other ␤-subunit dependent compounds including loreclezole and etomidate [2,17,30]. Previously, we determined topiramate is a partial agonist for heteromeric receptors containing ␤3 -subunits that competes with pentobarbital for a transduction/gating mechanism [14]. Our current observations indicate that this competitive interaction exists for ␤3 -homomers. The ␤3 N265 residue is a likely candidate for competition between topiramate, pentobarbital and loreclezole. We speculate that inhibition and activation occur via two distinct mechanisms, and that ␤1 -subunits contain an inhibitory mechanism and ␤3 -subunits contain inhibitory and activation mechanisms. In this scenario, ␤1 S265N and ␤3 N265S mutant subunits gain and lose the activation mechanisms, respectively, while retaining inhibitory functions. According to our above predictions (Section 4.1), heteromeric receptors containing ␤3 N265S should lack both potentiation and direct activation. Indeed, the ␤-subunit dependent anesthetic etomidate does not potentiate or activate ␣␤3 N265S receptors; however, pentobarbital potentia-

51

tion and direct activation are unaffected [2]. Mutation of threonine 262 to glutamine in ␣1 ␤1 T262Q receptors abolishes pentobarbital potentiation, but spares direct activation and inhibition [33]. These findings underscore the complexity of pentobarbital modulation and illustrate that the addition of other subunit types may establish new binding sites, modify existing binding sites or allosterically modulate transduction pathways that introduce a higher affinity potentiation aspect to channel activation. As seen with heteromeric receptors, topiramate and pentobarbital elicited rebound currents indicating inhibition of ␤1 - and ␤3 -homomers by either an open channel block or stabilization of a desensitized state [10,14,19]. Loreclezole lacked significant inward rebound currents for either ␤-homomer, but this may be due to strong desensitization with slow reversal kinetics [5,21,23]. Interaction experiments suggested a non-competitive inhibition between pentobarbital and either topiramate or loreclezole, whereas a competitive block between topiramate and loreclezole may exist. Similarly, co-application of pentobarbital with either topiramate or loreclezole to heteromeric ␣2 ␤3 ␥2 receptors increases rebound currents [14]. Overall, these observations suggest that ␤-homomeric receptors are an appropriate reduced system to examine the activation, inhibition and rebound currents elicited by these compounds. 5. Conclusion The existence of ␤-homomeric GABAA receptors in native systems is unlikely, but experimentation with homomeric receptors provides a foundation to deduce ␤-subunit contributions to receptor pharmacological properties and to elucidate basic functional binding sites via structure/function studies. The current study demonstrates that topiramate, pentobarbital, and loreclezole have functional binding sites on ␤1 - and ␤3 -subunits. From this foundation, contributions of other subunits in binary and ternary heteromeric receptors can be explored to gain a complete understanding of actions on complex heteromeric GABAA receptors. Future structure/function experiments based on the current findings will attempt to identify residues involved in binding and gating for topiramate. Acknowledgements The authors would like to sincerely thank Dr. Michael McIntosh (University of Utah) and associates for supplying Xenopus oocytes. We also thank Dr. Kristina A. Simeone for critical comments on this manuscript. This work was supported by an unrestricted grant by R. W. Johnson Pharmaceutical Research and Development, Springhouse, PA (HSW) and a Health Future Foundation award (TAS). References [1] Barish ME. A transient calcium-dependent chloride current in the immature Xenopus oocyte. J Physiol 1983;342:309–25. [2] Cestari IN, Min KT, Kulli JC, Yang J. Identification of an amino acid defining the distinct properties of murine ␤1 and ␤3 subunit-containing GABAA receptors. J Neurochem 2000;74:827–38. [3] Cestari IN, Uchida I, Li L, Burt D, Yang J. The agonistic action of pentobarbital on GABAA ␤-subunit homomeric receptors. Neuroreport 1996;7:943–7. [4] Davies PA, Kirkness EF, Hales TG. Modulation by general anaesthetics of rat GABAA receptors comprised of ␣1␤3 and ␤3 subunits expressed in human embryonic kidney 293 cells. Brit J Pharmacol 1997;120:899–909. [5] Donnelly JL, MacDonald RL. Loreclezole enhances apparent desensitization of recombinant GABAA receptor currents. Neuropharmacology 1996;35:1233–41. [6] Kindler CH, Verotta D, Gray AT, Gropper MA, Yost CS. Additive inhibition of nicotinic acetylcholine receptors by corticosteroids and the neuromuscular blocking drug vecuronium. Anesthesiology 2000;92(3):821–32. [7] Krishek BJ, Moss SJ, Smart TG. Homomeric ␤1 ␥-aminobutyric acidA receptorion channels: evaluation of pharmacological and physiological properties. Mol Pharmacol 1996;49:494–504.

52

T.A. Simeone et al. / Pharmacological Research 64 (2011) 44–52

[8] Li GD, Chiara DC, Sawyer GW, Husain SS, Olsen RW, Cohen JB. Identification of a GABAA receptor anesthetic binding site at subunit interfaces by photolabeling with an etomidate analog. J Neurosci 2006;26:11599–605. [9] Mangan PS, Sun C, Carpenter M, Goodkin HP, Sieghart W, Kapur J. Cultured hippocampal pyramidal neurons express two kinds of GABAA receptors. Mol Pharmacol 2005;67:775–88. [10] Rho JM, Donevan SD, Rogawski MA. Direct activation of GABAA receptors by barbiturates in cultured rat hippocampal neurons. J Physiol 1996;497:509–22. [11] Shank RP, Gardocki JF, Streeter AJ, Maryanoff BE. An overview of the preclinical aspects of topiramate: pharmacology, pharmacokinetics, and mechanism of action. Epilepsia 2000;41(Suppl. 1):S3–9. [12] Sigel E, Baur R, Malherbe P, Mohler H. The rat ␤1 -subunit of the GABAA receptor forms a picrotoxin-sensitive anion channel open in the absence of GABA. FEBS Lett 1989;257:377–9. [13] Simeone TA, Donevan SD, Rho JM. Molecular biology and ontogeny of gammaaminobutyric acid (GABA) receptors in the mammalian central nervous system. J Child Neurol 2003;18:39–48. [14] Simeone TA, Wilcox KS, White HS. Subunit selectivity of topiramate modulation of heteromeric GABA(A) receptors. Neuropharmacology 2006;50:845–57. [15] Thompson SA, Whiting PJ, Wafford KA. Barbiturate interactions at the human GABAA receptor: dependence on receptor subunit combination. Br J Pharmacol 1996;117:521–7. [16] Wafford KA, Bain CJ, McKernan RM, Wingrove PB, Whiting PJ, Kemp JA. A novel allosteric modulatory site on the GABAA receptor ␤ subunit. Neuron 1994;12:775–82. [17] Wingrove PB, Wafford KA, Bain C, Whiting PJ. The modulatory action of loreclezole at the ␥-aminobutyric acid type A receptor is determined by a single amino acid in the ␤2 and ␤3 subunit. Proc Natl Acad Sci USA 1994;91:4569–73. [18] Wooltorton JR, McDonald BJ, Moss SJ, Smart TG. Identification of a Zn2+ binding site on the murine GABAA receptor complex: dependence on the second transmembrane domain of ␤ subunits. J Physiol 1997;505:633–40. [19] Wooltorton JR, Moss SJ, Smart TG. Pharmacological and physiological characterization of murine homomeric ␤3 GABAA receptors. Eur J Neurosci 1997;9:2225–35. [20] Wollmuth LP. Multiple binding sites in Ih channels of rod photoreceptors from tiger salamanders. Pflugers Arch – Eur J Physiol 1995;430:34–43. [21] Fisher JL, Hinkle DJ, Macdonald RL. Loreclezole inhibition of recombinant ␣1␤1␥2L GABAA receptor single channel currents. Neuropharmacology 2000;39:235–45. [22] Serafini R, Bracamontes J, Steinbach JH. Structural domains of the human GABAA receptor ␤3 subunit involved in the actions of pentobarbital. J Physiol 2000;524:649–76.

[23] Feng HJ, Bianchi MT, Macdonald RL. Pentobarbital differentially modulates ␣1␤1␦and ␣1␤1␥2L GABAA receptor currents. Mol Pharmacol 2004;66:988–1003. [24] White HS, Brown SD, Woodhead JH, Skeen GA, Wolf HH. Topiramate modulates GABA-evoked currents in murine cortical neurons by a nonbenzodiazepine mechanism. Epilepsia 2000;41(Suppl. 1):S17–20. [25] Rho JM, Donevan SD, Rogawski MA. Barbiturate-like actions of the propanediol dicarbamates felbamate and meprobamate. J Pharmacol Exp Ther 1997;280:1383–91. [26] Tietz EI, Kapur J, Macdonald RL. Functional GABAA receptor heterogeneity of acutely dissociated hippocampal CA1 pyramidal cells. J Neurophysiol 1999;81:1575–86. [27] Kapur J, Macdonald RL. Pharmacological properties of ␥-aminobutyric acidA receptors from acutely dissociated rat dentate granule cells. Mol Pharmacol 1996;50:458–66. [28] Saxena NC, Macdonald RL. Assembly of GABAA receptor subunits: role of the delta subunit. J Neurosci 1994;14:7077–86. [29] Sanna E, Garau F, Harris A. Novel properties of homomeric ␤1 ␥-aminobutyric acid type A receptors: actions of the anesthetics propofol and pentobarbital. Mol Pharmacol 1995;47:213–7. [30] Halliwell RF, Thomas P, Patten D, James CH, Martinez-Torres A, Miledi R, et al. Subunit-selective modulation of GABAA receptors by the nonsteroidal anti-inflammatory agent, mefenamic acid. Eur J Neurosci 1999;11: 2897–905. [31] Fisher JL, Macdonald RL. Single channel properties of recombinant GABAA receptors containing ␥2 or ␦ subtypes expressed with ␣1 and ␤3 subtypes in mouse L929 cells. J Physiol 1997;505:283–97. [32] Brown N, Kerby J, Bonnert TP, Whiting PJ, Wafford KA. Pharmacological characterization of a novel cell line expressing human alpha(4)beta(3)delta GABA(A) receptors. Br J Pharmacol 2002;136:965–74. [33] Dalziel JE, Cox GB, Gage PW, Birnir B. Mutant human alpha(1)beta(1)(T262Q) GABA(A) receptors are directly activated but not modulated by pentobarbital. Eur J Pharmacol 1999;385:283–6. [34] Wafford KA, Thompson SA, Thomas D, Sikela J, Wilcox AS, Whiting PJ. Functional characterization of human gamma aminobutyric acid A receptors containing the alpha 4 subunit. Mol Pharmacol 1996;50:670–8. [35] Maksay G, Thompson SA, Wafford KA. The pharmacology of spontaneously open ␣1 ␤3 ␧ GABAA receptor-ionophores. Neuropharmacology 2003;44:994–1002. [36] Simeone TA, Otto JF, Wilcox KS, White HS. Felbamate is a subunit selective modulator of recombinant gamma-aminobutyric acid type A receptors expressed in Xenopus oocytes. Eur J Pharmacol 2006;552:31–5.