Suramin

Suramin is a novel competitive antagonist selective to α1β2γ2 GABAA over ρ1 GABAC receptors

Hui Luo, Kristofer Wood, Fu-Dong Shi, Fenfei Gao, Yongchang Chang

PII: S0028-3908(18)30586-0
DOI: 10.1016/j.neuropharm.2018.08.036
Reference: NP 7321

To appear in: Neuropharmacology

Received Date: 1 May 2018
Revised Date: 12 August 2018
Accepted Date: 26 August 2018

Please cite this article as: Hui Luo, Kristofer Wood, Fu-Dong Shi, Fenfei Gao, Yongchang Chang, Suramin is a novel competitive antagonist selective to α1β2γ2 GABAA over ρ1 GABAC receptors, Neuropharmacology (2018), doi: 10.1016/j.neuropharm.2018.08.036

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Key words: suramin; GABAA receptor; GABAC receptor; competitive antagonist; two- electrode voltage-clamp

Abbreviations: GABA: γ-aminobutyric acid; OR2: oocyte Ringer’s solution; DEPC: diethyl pyrocarbonate; MS-222: Ethy 3-aminobenzoate methanesulfonate salt; SR 95531 hydrobromide: gabazine ,6-Imino-3-(4-methoxyphenyl)-1(6H)-pyridazinebutanoic acid hydrobromide; WT: wild type; GABAR: GABA receptor; RU5135: 3a-hydroxy-16- imino-17-aza-5b-androstan-11-one; HPA axis: hypothalamus-pituitary-adrenal axis

Abstract

GABAA and GABAC receptors are both GABA-gated chloride channels with distinct pharmacological properties, mainly in their sensitivity to bicuculline and gabazine. In this study, we found that suramin, a purinergic receptor antagonist, is a novel competitive antagonist selective to GABAA over GABAC receptors. Specifically, suramin antagonized the GABA-induced current and the spontaneous opening current of the wild type α1β2γ2 GABAA receptor with high-level expression in Xenopus oocytes. The antagonism was concentration dependent with an IC50 that varied depending on the concentration of GABA, and with the lowest IC50 of 0.43 µM when antagonizing the spontaneous current. Thus, its potency is slightly higher than bicuculline on the same GABAA receptor. Suramin also antagonized the mouse native brain GABA receptors micro-transplanted into the Xenopus oocytes with its potency depending on the GABA concentration. In addition, in the presence of two fixed concentrations of suramin, the GABA concentration response of the receptor was shifted to the right without reduction of the maximum current. Thus, our results are consistent with that suramin is a competitive antagonist for the α1β2γ2 GABAA receptor. Interestingly, the rank order of maximum allosteric inhibition (efficacy) of spontaneous current of the GABAA receptor by three competitive antagonists was suramin>bicuculline>gabazine, similar to the rank order of their molecular weight. In contrast, similar to bicuculline, suramin has much lower potency in antagonizing the GABA-induced current of the ρ1 GABAC receptor. In conclusion, we have identified a novel GABAA receptor competitive antagonist, which is selective to the α1β2γ2 over ρ1 GABA receptors.

1. Introduction

γ-aminobutyric acid (GABA) is the major inhibitory neurotransmitter in the brain, mediating synaptic and non-synaptic inhibition. GABA receptors can be classified into three types (A, B, and C) according to their pharmacological properties (Drew et al., 1984; Hill and Bowery, 1981). That is, GABAA receptors are sensitive to bicuculline inhibition, GABAB receptors are selectively activated by baclofen, and GABAC receptors are insensitive to both bicuculline and baclofen (Chebib and Johnston, 1999). Molecular cloning, exogenous expression, and functional testing have revealed that GABAA and GABAC receptors are ionotropic receptors or GABA-gated chloride channels (Chebib and Johnston, 1999; Macdonald and Olsen, 1994), whereas GABAB receptors are metabotropic receptors or G-protein coupled receptors (Bettler et al., 2004). The GABA- gated ion channels belong to the pentameric ligand-gated ion channel superfamily, which also include nicotinic receptors (Albuquerque et al., 2009), serotonin receptor type 3 (Lummis, 2012), zinc-activated ion channel (Davies et al., 2003), and glycine receptors (Lynch, 2009) in vertebrate. The chloride channels gated by glutamate (Wolstenholme, 2012), acetylcholine (Putrenko et al., 2005), serotonin (Ranganathan et al., 2000), and histamine (Zheng et al., 2002) in invertebrate and prokaryotic ligand- gated ion channels (Corringer et al., 2010) also belong to this receptor superfamily.

Most GABAA receptors are heteromeric GABA-gated chloride channels, consisting of α, β, and γ/δ subunits and their isoforms (Whiting et al., 2000), whereas the ρ subunit- containing GABA receptors show GABAC receptor pharmacological properties (Johnston, 1996), including insensitivity to bicuculline. The α1β2γ2 GABAA receptor is the most abundant subtype of the GABAA receptors in the brain (Whiting et al., 2000), whereas the ρ1 homomeric GABA receptor represents a typical GABAC receptor (Johnston, 1996). Despite the pharmacological (Bormann, 2000; Chebib, 2004) and functional (Amin and Weiss, 1994) differences between them, GABAA and GABAC receptors belong to the same ligand-gated ion channel family with similar overall ion channel structure and function. Thus, the International Union of Pharmacology recommended using GABAA-ρ receptors instead of the ρ GABAC receptors for ρ subunit containing GABA receptors (Olsen and Sieghart, 2008). Nevertheless, when we discuss the bicuculline-like pharmacological properties of these two types of GABA-gated ion channels, it is more convenient to use the GABAC receptor for the ρ1 GABA receptor due to the familiar pharmacological properties in response to the competitive antagonists.
Suramin is an antiparasitic drug. It has been used clinically to treat human African trypanosomiasis (sleeping sickness) for nearly a century (Bouteille et al., 2003). It is also a relatively nonselective purinergic P2X and P2Y receptor antagonist (Jacobson and Müller, 2016; Ralevic and Burnstock, 1998). Interestingly, recently suramin has been used in a clinical trial to treat autism with a promising initial outcome (Naviaux et al., 2017), suggesting potentially through its central targets. While the mechanism of its beneficial effect on autism has been proposed to be related to its anti-purinergic effect (Naviaux, 2018), suramin has been reported to have other central targets, such as GABA receptors and glutamate receptors in hippocampal neurons (Nakazawa et al., 1995). This prompted the need for detailed characterization of its effect on a typical GABAA receptor with a known subunit composition. In this study, we set to fully characterize the effects of suramin on the representative GABAA and GABAC receptors: a heteromeric α1β2γ2 GABAA receptor and a homomeric ρ1 GABAC receptor (also called ρ1 GABA receptor or GABAA-ρ1 receptor) expressed in Xenopus oocytes using two-electrode voltage-clamp.

2. Materials and Methods

2.1 cDNA and cRNA Preparation. The cDNA encoding the wild type, rat α1, β2, γ2 and human ρ1 GABA receptor subunits were cloned into the pGEMHE vector in the T7 orientation. For cRNA synthesis, the cDNAs were individually amplified by PCR using the M13 forward and M13 reverse primer pair and high fidelity Phusion DNA polymerase (Thermo Fisher Scientific, Ipswich, MA, USA). The PCR products were then purified and served as the DNA templates for cRNA synthesis. The cRNAs encoding these GABA receptor subunits were transcribed by T7 RNA polymerase (Promega, Madison, WI, USA) using standard in vitro transcription protocol. After digestion of the DNA template by RNase-free DNase I, cRNAs were purified and re-suspended in diethyl pyrocarbonate (DEPC)-treated water. cRNA integrity was examined on a 1% agarose gel (Zhang, Xue et al. 2008), and cRNA concentration was determined by optical density measurements using a TECAN Infinite M200 PRO (Tecan Group Ltd., Switzerland).

2.2 Membrane protein preparation. Adult C57BL/6 mice (4 animals, weighing 20–30 g) were housed at constant temperature (22 ± 1 ◦C) and relative humidity (50%) under a regular light–dark schedule (light 07.00–19.00 h). Food and water were freely available. The animals were euthanized by decapitation under isoflurane anesthesia. The cerebral cortex and hippocampus were rapidly removed in ice cold PBS. The protocol for preparation of brain membrane from mice was approved by the Institutional Animal Care and Use Committee of the Barrow Neurological Institute and St. Joseph’s Hospital and Medical Center. The membrane proteins of the cerebral cortex and hippocampus were prepared according to the method described by Miledi, et al (Miledi et al., 2002). The cortex (~0.3 g) or hippocampus (~0.2g) were homogenized in a Teflon glass homogenizer with 2 ml of glycine buffer (200 mM glycine, 150 mM NaCl, 50 mM EGTA, 50 mM EDTA, and 300 mM sucrose; plus 25 µl of protease inhibitor cocktails (Sigma P8340), pH 9) separately on ice. The homogenate was centrifuged for 15 min at 9,500 g in a Beckmann centrifuge (TA-10.250 rotor). The supernatant was then centrifuged for 2 h at 100,000 g in a Beckmann centrifuge (SW32Ti rotor) at 4°C. The pellet was washed, and then resuspended in 5 mM glycine solution. The protein concentration of the membrane prep was measured by BCA method (Thermo SCIENTIFIC). The purified membrane proteins were used directly for microinjection or aliquoted and stored at – 80°C for later use.

2.3 Oocyte preparation and injection. Oocytes were harvested from female Xenopus laevis (Xenopus I, Ann Arbor, MI, USA), using the protocol approved by the Institutional Animal Care and Use Committee for this study and complied with the National Institutes of Health guide for the Care and Use of Laboratory Animals (https://grants.nih.gov/grants/olaw/guide-for-the-care-and-use-of-laboratory- animals.pdf). Briefly, frogs were anesthetized by 0.2% MS-222, and the ovarian lobes were surgically removed and placed in the incubation solution consisting of (in mM) 82.5 NaCl, 2.5 KCl, 1 MgCl2, 1 CaCl2, 1 Na2HPO4, 0.6 theophylline, 2.5 sodium pyruvate, and 5 HEPES; 50 U/ml penicillin, and 50 µg/ml streptomycin, pH 7.5. The frogs were given analgesic xylazine hydrochloride (10mg/kg, ip), antibiotic gentamicin (3mg/kg, ip), and anesthetic 0.25% bupivacaine (topical application) after surgery, and then allowed to recover from the surgery before being returned to the incubation tank. The ovarian lobes were cut into small pieces and digested with 1 Wunsch unit/ml liberase blendzyme 3 (Roche Applied Science, Indianapolis, IN, USA) with constant stirring at room temperature for 1.25 hours. The dispersed oocytes were thoroughly rinsed with the above solution. The stage VI oocytes were isolated with homemade sorter and hand selection, and incubated at 16 °C before injection. Micropipettes for injection were pulled from borosilicate glass on a Sutter P97 horizontal puller, and the tips were cut with forceps to ≈40 µm in diameter. The cRNAs (α:β:γ=1:1:1 mix, or ρ1 alone) were drawn up into the micropipette and injected into oocytes with a Nanoject II micro- injection system (Drummond, Broomall, PA, USA) at a total volume of ~60 nl, and total RNA of 12 ng. For membrane protein injection, oocytes were injected with the membrane proteins in 5 mM glycine (~ 90 nl; with the protein concentration of 1–3 mg/ml) in 10% sucrose solution (with incubation solution). The glycine solution (5mM) was used for sham injection control to the oocytes from the same batch.

2.4 Two-electrode voltage-clamp. Two to 5 days after injection, oocytes expressing GABA receptors were placed in a homemade, small volume chamber and a 8-channel manifold, continuously perfused with oocyte Ringer’s solution (OR2), which consisted of (in mM) 92.5 NaCl,
2.5 KCl, 1 CaCl2, 1 MgCl2 and 5 HEPES, pH 7.5. A ValveLink 8 computer-controlled perfusion system was used to switch solutions between the OR2 and a test solution. The perfusion chamber was grounded through an agar salt bridge to avoid possible alteration of the junction potential between the Ag/AgCl grounding electrode and solution by drugs. Oocytes were voltage-clamped at -70 mV to measure agonist-induced currents using an AxoClamp 900A amplifier (Molecular Devices, LLC, San Jose, CA, USA). The current signal was low-pass filtered at 20 Hz with the built-in 4-pole low-pass Bessel filter and digitized at 50 Hz with a Digidata1440A and pClamp 10 software (Molecular Devices).

2.5 Drug Preparation. GABA, picrotoxin, and zinc chloride were purchased from Sigma- Aldrich (St. Louis, MO, USA). Suramin hexasodium salt, 1(S),9(R)-(−)-Bicuculline methbromide, and SR 95531 hydrobromide (also named gabazine) were purchased from Tocris (Bio-Techne Corporation, Minneapolis, MN). The stock solutions of all drugs, except picrotoxin, were prepared in 18 MΩ water. Picrotoxin stock solution (3 mM) was prepared in OR2. All stock solutions were stored in small aliquots at -20°C before use.

2.6 Data Analysis. For each oocyte, concentration response or inhibition curves were fit to the Hill equation using least squares methods (Prism 6.0, GraphPad Software, Inc., San Diego, CA, USA) to derive EC50 or IC50 values (the concentration required for inducing a half maximal current, or half maximal inhibition), Hill coefficients (the slope factor), and maximum current. Current responses to different drug concentrations in each oocyte were normalized to the fitted maximum current, allowing the normalized data to be averaged across multiple oocytes and to be fit again to the Hill equation for plots. Statistical significance of the differences between the log(EC50) or log(IC50) values for comparison between multiple concentration responses or inhibitions was determined by two-sided grouped student t test in the Microsoft Excel 2013 or one-way ANOVA with post-hoc Tukey’s multiple comparisons test in GraphPad Prism. All data are presented as mean ± SEM (standard error).

3. Results and Discussion

3.1. Suramin antagonized the GABA-induced current of the α1β2γ2 GABAA receptor concentration dependently.

We tested suramin’s inhibitory effect on the GABA-induced current of the α1β2γ2 GABAA receptor expressed in the Xenopus oocytes using two-electrode voltage-clamp. Figure 1 shows that 5 µM GABA-induced current was reduced by co-application of suramin with increasing concentrations. The concentration dependent inhibition could be adequately fitted by the single Hill equation, suggesting suramin likely only has a single affinity binding site(s) on the receptor. This result demonstrates that suramin is an antagonist of this typical GABAA receptor. The log(IC50) of suramin inhibition of the 5 µM GABA-induced current was -5.621±0.069 (M) (IC50=2.4 µM). Thus, suramin is a relatively potent antagonist of the α1β2γ2 GABAA receptor.

This result is in contrast to the report of suramin effect on the GABA-induced current in rat hippocampal neurons with an IC50>100 µM (Nakazawa et al., 1995). Since we tested suramin’s effect on rat α1β2γ2 GABAA receptor, the sensitivity difference clearly does not arise from difference in species. The GABA concentration used in our case is 5 µM, whereas in their case is 10 µM. A two fold difference in GABA concentration cannot explain >40 fold difference in IC50. Thus, additional differences in suramin’s potency on GABA receptors expressed in oocytes vs. in neurons could arise from the multiple subtypes of GABAA receptors in neurons vs. the single subtype in the oocyte expression system.

3.2. Suramin’s potency of antagonizing GABA effect was GABA concentration dependent.

Antagonists can be classified into two categories according to their mechanism of action: competitive antagonists and non-competitive antagonists. Competitive antagonists share the binding site in the receptor with their competing agonists. Thus, they will compete with the agonists for the same orthosteric binding site. In contrast, non-competitive antagonists bind to other (allosteric) site(s) or directly block the channel (Chen et al., 2006). One way to differentiate these two mechanisms of action by an antagonist is to test whether the antagonist’s sensitivity is different in antagonizing the currents induced by different concentrations of the agonist. Figure 2 shows that a 10- fold increase in GABA concentration shifted suramin concentration dependent inhibition to the right. The log(IC50) of suramin inhibition of the 50 µM GABA-induced current was -4.458±0.135 (M) (IC50=34.8 µM). The statistical test for the difference between the log(IC50) values of suramin in inhibiting the currents induced by two different GABA concentrations resulted in a p value of 4.40125E-07. Thus, the current induced by the higher GABA concentration requires higher suramin concentrations to block it. The result is consistent with that suramin is a competitive antagonist of the α1β2γ2 GABAA receptor.

Our result is in contrast to the reported inhibitory effect of suramin on the GABA current in rat hippocampal neurons, in which the fraction of suramin inhibition of the GABA- induced current is not dependent on GABA concentration (Nakazawa et al., 1995). In their case, 100 µM suramin blocked about 40% of the currents induced by both 10 µM GABA and 100 µM GABA. In contrast, our result shows that 100 µM suramin nearly completely inhibited 5 µM GABA-induced current, but only inhibited 50 µM GABA- induced current by 66%. The reason for the difference between our data in expressed GABAA receptor and the reported one in hippocampal neurons is not clear. Since we used complete concentration inhibition relationship to show that the IC50 for suramin inhibition is dependent on GABA concentration, we have higher confidence to conclude that suramin is a competitive antagonist of the GABAA receptor.

3.3. Suramin shifted GABA concentration-response to the right without maximum current reduction.

To provide another line of evidence to support the competitive nature of suramin antagonism on the GABAA receptor, we further performed GABA concentration responses in the absence or presence different concentrations of suramin. In the presence of a competitive antagonist, we expected that higher concentrations of the agonist are needed to activate the receptor. That is, the GABA concentration-response would be shifted to the right in the presence of suramin. Figure 3. shows that indeed in the presence of suramin in two different concentrations, the GABA concentration response shifted to the right to different extent with a greater shift in the presence of higher suramin concentration. At the same time, the maximum currents induced by GABA were not reduced in the presence of suramin. This result further supports our conclusion that suramin is a competitive antagonist of the α1β2γ2 GABAA receptor. Again, the difference in supporting two different antagonizing mechanisms of suramin on the GABA receptor between our results and the reported results in neurons is not clear.

3.4. Suramin had higher potency in blocking spontaneous current of the wild type α1β2γ2 GABAA receptor.

From above experimental results, we can conclude that suramin is a GABAA receptor competitive antagonist, similar to bicuculline and gabazine. Since suramin’s IC50 in antagonizing GABA response varies depending on the GABA concentration, it is necessary to normalize it. To estimate suramin’s inhibitory constant (Ki), we can use Cheng-Prusoff/Chou equation (Cheng and Prusoff, 1973; Chou, 1974) in the form of Ki = IC50/(1 + [A]/EC50) (Wagner and Czajkowski, 2001). Since the inhibition by the highest concentration of suramin was nearly saturated for the 5-µM GABA-induced current but not for 50-µM GABA-induced current, we calculated the Ki based on its inhibition on the 5-µM GABA-induced current. The resulting Ki was 1.3 µM. However, it may not represent its true affinity. To measure the true affinity, we examined the concentration response of suramin on the spontaneously opening current of the wild type GABAA receptor with high expression in the absence of GABA (without competing agonist).

Figure 4 shows that suramin induced an apparently outward current in the oocytes highly expressing the α1β2γ2 GABAA receptor. This is not due to its effect on native oocyte channels, since no suramin effect was observed in the sham-injected control oocytes (data not shown). Instead, it is the inhibition of spontaneously opening current of the wild type GABAA receptor by suramin. The inhibition of the spontaneous current of the GABA receptor by suramin was concentration dependent, with a sub-micromolar IC50 (0.43 µM, log(IC50)=-6.36±0.05 M). This is slightly more potent than the bicuculline allosteric inhibition of the spontaneously opening mutant α1β2γ2 GABAA receptor (IC50=1.1 µM) (Chang and Weiss, 1999) or the current induced by the allosteric activators alphaxalone (IC50= 0.9 µM) or pentobarbital (IC50= 1.0 µM) of the wild type GABAA receptor (Ueno et al., 1997).

It is known that agonist affinity to its receptor depends on the state of the receptor (Colquhoun, 1998; Edelstein and Changeux, 1996). In the resting state, the agonist affinity to its receptor is low. However, in the activated state, the agonist affinity to the receptor is much higher. In case of the α1β2γ2 GABAA receptor, the estimated GABA affinities are 78.5 µM in the closed resting states and 0.12 µM in the open states (Chang and Weiss, 1999). Desensitization of a receptor can further increase the binding affinity to its agonist, and the desensitized state is the most stable state with the highest affinity to an agonist. The desensitized state of the α1β2γ2 GABAA receptor has the affinity of 0.04 µM as measured by radio-ligand binding in single oocytes (Chang et al., 2002). In contrast, binding of a competitive antagonist to the desensitized receptor can facilitate receptor recovery from desensitization to the resting state (Xu et al., 2016), suggesting that the resting state is the most stable state for a receptor to bind to its competitive antagonist. Thus, the IC50 of suramin inhibition of spontaneous current of the wild type α1β2γ2 GABAA receptor would represent the true affinity of the receptor to this competitive antagonist.

3.5. Size dependent block of the spontaneous current of the α1β2γ2 GABAA receptor by all three competitive antagonists.

Bicuculline and gabazine have been reported to be allosteric inhibitors of the GABAA receptor current activated by the allosteric activators of neurosteroids and barbiturates (Ueno et al., 1997), both of which have the binding sites in the transmembrane domain (Amin, 1999; Greenfield et al., 2002; Hosie et al., 2006). Suramin inhibition of the spontaneous current of the GABAA receptor (Figure 4) is also allosteric in nature. It would be interesting to compare these three allosteric inhibitors for their extent of allosteric inhibition of the spontaneous current of the GABAA receptor. Using a saturation concentration of the non-competitive antagonist picrotoxin as the control, we compared the maximum allosteric inhibition of the spontaneous opening of the GABAA receptor by a saturation concentration of these competitive antagonists.

Figure 5 shows that 100 µM suramin clearly had higher percentage of inhibition than bicuculline, whereas gabazine had the lowest percentage of inhibition. Higher fraction of inhibition by bicuculline than by gabazine is also observed in their antagonizing spontaneously opening mutant GABAA receptors (Chang and Weiss, 1999), as well as in antagonizing neurosteroid- or barbiturate-activated current of the wild type GABAA receptors (Ueno et al., 1997). The molecular weight of suramin is 1,297 Daltons, which is larger than that for bicuculline methobromide (462 Daltons), while the molecular weight of gabazine hydrobromide is the smallest (368 Daltons). The rank order of the molecular weight of these compounds is the same as the rank order of their maximum inhibition on the spontaneous current of the GABAA receptor, suggesting that the size of these competitive antagonists is related to the extent of their allosteric inhibition of the receptor gating.

The crystal structures of the acetylcholine binding protein in the presence of different agonists and competitive antagonists reveal that binding of a competitive antagonist to its receptor in general results in a more open position of the binding loop C (loop C uncapping) (Brams et al., 2011). Larger sized peptide antagonists result in a larger uncapping. On the contrary, agonists induce the loop C capping, causing contraction of the binding pocket (Celie et al., 2004; Hansen et al., 2005). Loop C capping has been proposed to initiate channel opening (Lee et al., 2008; Lee and Sine, 2005). However, further studies suggest that loop C capping may not be directly coupled to channel opening, but instead it stabilizes the receptor in the high affinity (for agonist) state to increase open probability (Jadey and Auerbach, 2012; Purohit and Auerbach, 2013).

The inference of this is that loop C uncapping would stabilize the receptor in the resting closed state. Thus, a reasonable explanation of our observation in Figure 5 is that the size of a competitive antagonist would be an important determinant for the extent of loop C uncapping by an antagonist, therefore the extent of stabilizing receptor in the resting state. This could be the potential mechanism for the efficacy difference in the inhibition of the spontaneously opening channels by these competitive antagonists with different sizes.

3.6. Suramin had low potency on the GABA-induced current of the ρ1 GABA receptor.

It is known that the GABAA receptor competitive antagonists, such as bicuculline and gabazine, have a selectivity to GABAA receptors over GABAC receptors. If suramin is a GABAA receptor competitive antagonist, then it would act like bicuculline and gabazine. That is, it would have low potency on the GABAC receptors. To test this hypothesis, we examined the effect of suramin on the ρ1 homomeric GABA receptor, which has the GABAC receptor pharmacology, i.e. relative insensitivity to bicuculline and gabazine.

Figure 6A shows that indeed 5-µM GABA-induced current of the ρ1 homomeric GABA receptor is essentially insensitive to 100 µM suramin and 100 µM bicuculline, confirming that suramin’s selectivity for GABAA over GABAC receptor is similar to bicuculline.However, the ρ1 GABA receptor is not completely insensitive to the GABAA receptor competitive antagonists. When we reduced the GABA concentration to 0.9 µM, both 100 µM bicuculline and 100 µM suramin partially blocked the GABA-induced current of the ρ1 GABA receptor (Figure 6B). Higher percentage of the inhibition of the 0.9-µM GABA- induced current by 100-µM suramin than that by 100-µM bicuculline is in parallel to their potency difference in inhibiting the α1β2γ2 GABAA receptor.

3.7. Suramin competitively antagonized mouse native GABA receptors micro- transplanted into the Xenopus oocytes

In order to confirm that suramin can also antagonize the native GABAA receptors in a similar way, we purified the membrane proteins from mouse cerebral cortex and hippocampus, and micro-transplanted them into the Xenopus oocytes. The oocytes injected with the membrane proteins responded to GABA in a concentration dependent manner with the GABA EC50s of 63.3 ± 5.8 µM for cortex and 53.2 ± 3.5 µM for hippocampus. In contrast, the sham-injected oocytes with 5mM glycine did not respond to GABA (data not shown). We then selected two GABA concentrations (50 and 500 µM) to test suramin sensitivity. Figure 7. shows that indeed suramin antagonized the GABA-induced currents in the oocytes injected with the membrane proteins purified from the mouse brain (cerebral cortex and hippocampus). Similarly, a higher GABA- concentration requires higher suramin to antagonize the current to the same level. This suramin antagonism on the GABA-induced current was observed in the oocytes expressing either cortex membrane proteins or hippocampus membrane proteins. The result is consistent with that suramin is also a competitive antagonist of the mixed populations of the GABA-gated ion channels from the mouse brain and expressed in Xenopus oocytes.

In summary, we have provided several lines of evidence to support that suramin is a relatively potent novel competitive antagonist of the GABAA receptor, selective to the α1β2γ2 over ρ1 GABA receptors, similar to bicuculline. Discovery of suramin as a competitive GABAA receptor antagonist is unexpected given no structural similarity between suramin and bicuculline or gabazine. However, all known GABAA receptor competitive antagonists, including bicuculline, gabazine, salicylidene salicylhydrazide, RU5135, and 4-(3-biphenyl-5-(4-piperidyl)-3-isoxazoleare, are structurally unrelated. Thus, it is difficult to study the structure-activity relationship among these compounds for their binding affinity to the GABAA receptor.

3.8. Significance

Bicuculline and gabazine have been used as a tool to functionally dissect GABAA receptor components in neurons. Clearly, suramin cannot be added to the repertoire of the tools for GABAA receptor functional isolation, since it is also a non-selective antagonist of purinergic receptors (Lambertucci et al., 2015). However, our finding would suggest alternative interpretation when considering the mechanism of action of suramin in vivo application. That is, its mechanism of action in vivo could be through GABAA receptors. Unlike other GABAA receptor competitive antagonists, suramin is a clinically prescribed medication. It can be used for clinical trials. In fact, low dose suramin has been used in a clinical trial recently with promising beneficial effect on autism (Naviaux et al., 2017). Similar beneficial effect has been observed in an autism animal model (Naviaux et al., 2015). It has been suggested that this effect is through its anti-purinergic receptor action (Naviaux, 2018). Suramin has its IC50s in micro-molar range when antagonizing several subtypes of P2X receptor (Lambertucci et al., 2015). Here, we show that suramin is a competitive antagonist of a typical GABAA receptor with an IC50 in low micro-molar range. Since GABAA receptors are widely expressed in neurons, as well as in immune cells (Barragan et al., 2015), it is equally likely that suramin would influence brain function through GABAA receptors as opposed to through purinergic receptors. Interestingly, peripheral application (intraperitoneally) of the GABAA receptor competitive antagonist, bicuculline, can block the hypothalamus- pituitary-adrenal (HPA) axis activation induced by ghrelin (Gastón et al., 2017). Since the HPA axis is more reactive to stress in autism spectrum disorders (Jacobson, 2014; Spratt et al., 2012), our study would suggest that the beneficial effect of suramin on autism observed in the clinical trial could be through normalizing HPA axis response in autistic patients by antagonizing GABAA receptors.

The function of the HPA axis is under the GABAergic influence with the mixed effects of inhibition and enhancement (Mikkelsen et al., 2005). Potentiation of the GABAA receptor function by the GABAA receptor α1 subunit selective positive allosteric modulator, zolpidem, increases the HPA activity (Mikkelsen et al., 2008; Mikkelsen et al., 2005), whereas enhancing non-α1 subunit containing GABAA receptors with L-818,417 decreases the activity (Mikkelsen et al., 2008). However, overall effect of the subunit non-selective GABAA receptor positive allosteric modulator, such as diazepam, on the HPA axis activity is to increase the activity (although to a lesser extent) (Mikkelsen et al., 2005). This could explain the paradoxical effect of diazepam on autism: potentiation of inhibitory GABA receptor can worsen the behavioral responses (Marrosu et al., 1987). The paradoxical effect of the GABAA receptor positive modulators was explained by GABA-mediated excitation (Bruining et al., 2015; Lemonnier et al., 2012). Autism is a developmental disorder with the disturbed neuronal maturation (Casanova and Casanova, 2014). This is also the rationale to use bumetanide (reducing intracellular chloride concentration to reverse the GABAA receptor-mediated excitation) to treat the autistic children in a clinical trial (Bruining et al., 2015; Lemonnier et al., 2012). Thus, the antagonism of GABAA receptors by suramin could be one of potential mechanisms for its normalizing effect on the hyperactive HPA axis. Alternatively, suramin may have direct impact on adrenal cortex. In fact, suramin can inhibit cortisol secretion and adrenocortical cell growth in vitro (Dorfinger, et al, 1991).

Nevertheless, our finding of suramin is a GABAA receptor competitive antagonist would provide an interesting alternative potential mechanism of action of suramin on autism, and would be a new clue for future investigation. For example, it would help design experiments with controls using more selective GABAA receptor antagonists or P2X receptors antagonists in animal models or clinical trials to differentiate these two possibilities.

4. Declaration of interest

Authors have no competing interests to declare.

Authorship Contributions

YC, FG, FDS designed experiments, contributed reagents/materials/analysis tools; HL and KW performed experiments; HL and YC analyzed data; YC wrote the draft; YC, FG, and HL revised the manuscript.

Funding

This work was supported by Barrow Neurological Foundation (to YC).

Figure legends

Figure 1. Suramin antagonized the GABA-induced current of the α1β2γ2 GABAA receptor concentration dependently. A. Representative current traces of the 5 µM GABA-induced currents from an α1β2γ2 GABAA receptor-expressing oocyte, in the absence or presence of increasing concentration of suramin. B. The normalized and averaged concentration inhibition of suramin on the 5 µM GABA-induced current (N=10). The continuous line is the least-squares fit of the Hill equation to the concentration inhibition data. The resulting log(IC50) is given in the text.

Figure 2. The sensitivity of the suramin antagonism was GABA concentration dependent. A. Representative current traces of the 50 µM GABA-induced currents from an α1β2γ2 GABAA receptor-expressing oocyte, in the absence or presence of increasing concentration of suramin. B. The normalized and averaged concentration inhibition of suramin on the 50 µM GABA-induced current (N=8) compared with the replotted data from Figure 1. The continuous lines are the least-squares fit of the Hill equation to the concentration inhibition data. The resulting log(IC50) is given in the text.

Figure 3. Suramin shifted GABA concentration-response curve to the right without reducing the maximum current. A. Representative traces of the GABA-induced currents in the absence or presence of 10 or 100 µM of suramin. The last traces (thick traces) for GABA responses in the presence of suramin are the GABA control (316 µM without suramin) for normalization. B. Normalized and averaged GABA concentration-response in the absence (N=10, group A) or presence of 10 µM (N=9, group B) or 100 µM (N=10, group C) of suramin. The continuous lines are the least-squares fit of the Hill equation to the concentration inhibition data. The resulting log(EC50)s were -5.22±0.03, -5.08±0.03,
-4.74±0.08 (M). One-way ANOVA test results were F=19.88 and P=0.0010. The post hoc Tukey’s multiple comparisons test shows that there were statistical significance between A-B, A-C, and B-C groups. Note the maxima of GABA concentration response normalized to the control traces in the presence of suramin were slightly higher than 1 (1.07 and 1.06 respectively). For convenience of comparison of the concentration response shift, we plotted with the current levels normalized to their respective maximum. Nevertheless, the maxima of the GABA-induced current were not reduced in the presence of suramin.

Figure 4. Suramin antagonism of spontaneous current of the α1β2γ2 GABAA receptor.A. Representative traces of the concentration dependent inhibition of the spontaneous inward currents by suramin. B. Normalized and averaged suramin concentration- inhibition of the spontaneous current, assuming maximum inhibition to zero current level (N=13). However, as showing in Figure 5, the maximum inhibition by suramin did not completely block the current. The continuous lines are the least-squares fit of the Hill equation to the concentration inhibition data. The resulting log(IC50) was -6.36±0.05 M (IC50 = 0.43 µM).

Figure 5. Maximum inhibition of the spontaneous current by suramin was higher than the maximum inhibition by bicuculline and gabazine. A. Current traces of inhibition of spontaneous inward current by 100 µM of suramin, bicuculline, or gabazine, and compared to the inhibition by 100 µM picrotoxin. Note that 100 µM is the saturation concentration for all of these antagonists. Zinc chloride (for αβγ vs. αβ or β homomer) served as the control to make sure that spontaneous current is not from αβ or β homomeric GABAA receptors. Twenty micro-molar of zinc chloride is the saturation concentration for αβ and β GABAA receptors (Xu et al., 2016). B. Normalized (to picrotoxin inhibition) and averaged inhibition (N=12). Note that among three competitive antagonists, suramin exhibited the highest inhibition, whereas gabazine had the lowest inhibition. One-way ANOVA test of the data normalized to the 2 µM GABA response in each oocyte allowed us to compare picrotoxin inhibition to other inhibitions for statistical significance (F=190.1, P<0.0001). The post hoc Tukey’s multiple comparisons test revealed that there was statistical significance for the difference between each pair. Figure 6. The ρ1 GABA receptor was less sensitive to suramin and bicuculline inhibition. A. Upper panel: currents induced by 5 µM GABA in the oocytes expressing the ρ1 GABA receptor were nearly insensitive to 100 µM suramin or 100 µM bicuculline. Lower panel: when GABA concentration was reduced to 0.9 µM, the GABA-induced current could be partially inhibited by 100 µM suramin or 100 µM bicuculline. B. Normalized (to the average of before and after controls) and averaged (N=12 for each group) inhibition of the GABA-induced currents by 100 µM suramin or 100 µM bicuculline. Figure 7. Suramin competitively antagonized mouse native GABA receptors micro- transplanted into the Xenopus oocytes. A. Concentration dependent inhibition of suramin antagonism of the GABA-induced currents recorded from the oocytes expressing the membrane proteins from mouse cerebral cortex. Upper panel: the raw current traces were antagonized by increasing concentrations of suramin. Lower panel: Normalized and averaged (N=8 for each group) inhibition of the GABA-induced currents by suramin. Error bars are smaller than the symbols. The IC50 of suramin in antagonizing 50 µM GABA was 0.71±0.07 µM, whereas the IC50 of suramin in antagonizing 500 µM GABA was 24.47 ± 1.64 µM. B. Concentration dependent inhibition of suramin antagonism of the GABA-induced currents recorded from the oocytes expressing the membrane proteins from mouse hippocampi. Upper panel: the raw current traces were antagonized by increasing concentrations of suramin. Lower panel: Normalized and averaged (N=8 for each group) inhibition of the GABA-induced currents by suramin. Error bars are smaller than the symbols. The IC50 of suramin in antagonizing 50 µM GABA was 2.50±0.19 µM, whereas the IC50 of suramin in antagonizing 500 µM GABA was 28.96 ± 2.89 µM. References Albuquerque, E., Pereira, E., Alkondon, M., Rogers, S., 2009. Mammalian nicotinic acetylcholine receptors: from structure to function. Phyiological Review 89, 73-120. Amin, J., 1999. A single hydrophobic residue confers barbiturate sensitivity to -aminobutyric acid type C receptor. Molecular Pharmacology 55, 411-423. Amin, J., Weiss, D., 1994. Homomeric  GABA channels: activation properties and domains. Receptors Channels 2, 227-236. Barragan, A., Weidner, J., Jin, Z., Korpi, E., Birnir, B., 2015. GABAergic signalling in the immune system. Acta Physiol (Oxford) 213, 819-827. Bettler, B., Kaupmann, K., Mosbacher, J., Gassmann, M., 2004. Molecular structure and physiological functions of GABA(B) receptors. Physiological Reviews 84, 835-867. Bormann, J., 2000. The ‘ABC’ of GABA receptors. Trends in Pharmacological Sciences 21. Bouteille, B., Oukem, O., Bisser, S., Dumas, M., 2003. Treatment perspectives for human African trypanosomiasis. Fundamental & Clinical Pharmacology 17, 171-181. Brams, M., Pandya, A., Kuzmin, D., van Elk, R., Krijnen, L., Yakel, J., Tsetlin, V., Smit, A., Ulens, C., 2011. A structural and mutagenic blueprint for molecular recognition of strychnine and d-tubocurarine by different cys-loop receptors. PLoS Biology 9, e1001034. Bruining, H., Passtoors, L., Goriounova, N., Jansen, F., Hakvoort, B., de Jonge, M., Poil, S., 2015. Paradoxical Benzodiazepine Response: A Rationale for Bumetanide in Neurodevelopmental Disorders? Pediatrics 136, e539-543. Casanova, E., Casanova, M., 2014. Genetics studies indicate that neural induction and early neuronal maturation are disturbed in autism. Front Cell Neurosci 8, 397. Celie, P., Rossum-Fikkert, S., Dijk, W., Brejc, K., Smit, A., Sixma, T., 2004. Nicotine and carbamylcholine binding to nicotinic acetylcholine receptors as studied in AChBP crystal structures. Neuron 41, 907-914. Chang, Y., Ghansah, E., Chen, Y., Ye, J., Weiss, D., 2002. Desensitization mechanism of GABA receptor revealed by single oocyte binding and receptor function. Journal of Neuroscience 22, 7982-7990. Chang, Y., Weiss, D., 1999. Allosteric activation mechanism of the  -aminobutyric acid type A receptor revealed by mutation of the conserved M2 leucine. Biophysical Journal 77, 2542-2551. Chebib, M., 2004. GABAC receptor ion channels. Clinical and Experimental Pharmacology and Physiology 31, 800-804. Chebib, M., Johnston, G., 1999. The 'ABC' of GABA receptors: a brief review. Clinical and Experimental Pharmacology and Physiology 26, 937–940. Chen, L., Durkin, K., Casida, J., 2006. Structural model for -aminobutyric acid receptor noncompetitive antagonist binding: Widely diverse structures fit the same site. Proceedings of the National Academy of Sciences of the United States of America 103, 5185-5190. Cheng, Y., Prusoff, W., 1973. Relationship between the inhibition constant (KI) and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction. Biochem Pharmacol 22, 3099-3108. Chou, T., 1974. Relationships between inhibition constants and fractional inhibition in enzyme- catalyzed reactions with different numbers of reactants, different reaction mechanisms, and different types and mechanisms of inhibition. Molecular Pharmacology 10, 235-247. Colquhoun, D., 1998. Binding, gating, affinity, and efficacy: the interpretation of structure- activity relationships for agonists and of the effects of mutating receptors. British Journal of Pharmacology 125, 923-948. Corringer, P., Baaden, M., Bocquet, N., Delarue, M., Dufresne, V., Nury, H., Prevost, M., Van Renterghem, C., 2010. Atomic structure and dynamics of pentameric ligand-gated ion channels: new insight from bacterial homologues. Journal of Physiology 588, 565-572. Davies, P., Wang, W., Hales, T., Kirkness, E., 2003. A novel class of ligand-gated ion channel is activated by Zn2+. Journal of Biological Chemistry 278, 712–717. Drew, C., Johnston, G., Weatherby, R., 1984. Bicuculline-insensitive GABA receptors: studies on the binding of (-)-baclofen to rat cerebellar membranes. Neuroscience Letters 52, 317-321. Edelstein, S., Changeux, J., 1996. Allosteric proteins after thirty years: the binding and state functions of the neuronal 7 nicotinic acetylcholine receptors. Experientia 52, 1083-1090. Gastón, M., Cid, M., Salvatierra, N., 2017. Bicuculline, a GABAA-receptor antagonist, blocked HPA axis activation induced by ghrelin under an acute stress. Behavioral Brain Research 320, 464-472. Greenfield, L. J., Zaman, S., Sutherland, M., Lummis, S., Niemeyer, M., Barnard, E., Macdonald, R., 2002. Mutation of the GABAA receptor M1 transmembrane proline increases GABA affinity and reduces barbiturate enhancement. Neuropharmacology 42, 502-521. Hansen, S., Sulzenbacher, G., Huxford, T., Marchot, P., Taylor, P., Bourne, Y., 2005. Structures of Aplysia AChBP complexes with nicotinic agonists and antagonists reveal distinctive binding interfaces and conformations. EMBO Journal 24, 3635-3646. Hill, D., Bowery, N., 1981. 3H-baclofen and 3H-GABA bind to bicuculline-insensitive GABA B sites in rat brain. Nature 290, 149-152. Hosie, A., Wilkins, M., da Silva, H., Smart, T., 2006. Endogenous neurosteroids regulate GABAA receptors through two discrete transmembrane sites. Nature 444, 486-489. Jacobson, K., Müller, C., 2016. Medicinal chemistry of adenosine, P2Y and P2X receptors. Neuropharmacology 104, 31-49. Jacobson, L., 2014. Hypothalamic-pituitary-adrenocortical axis: neuropsychiatric aspects. Comprehensive Physiology 4, 715-738. Jadey, S., Auerbach, A., 2012. An integrated catch-and-hold mechanism activates nicotinic acetylcholine receptors. Journal of General Physiology 140, 17–28. Johnston, G., 1996. GABAC receptors: relatively simple transmitter-gated ion channels? Trends in Pharmacological Sciences 17, 319-323. Lambertucci, C., Dal Ben, D., Buccioni, M., Marucci, G., Thomas, A., Volpini, R., 2015. Medicinal chemistry of P2X receptors: agonists and orthosteric antagonists. Current Medicinal Chemistry 22, 915-928. Lee, W., Free, C., Sine, S., 2008. Nicotinic receptor interloop proline anchors  and Cys loops in coupling agonist binding to channel gating. Journal of General Physiology 132, 265– 278. Lee, W., Sine, S., 2005. Principal pathway coupling agonist binding to channel gating in nicotinic receptors. Nature 438, 243-247. Lemonnier, E., Degrez, C., Phelep, M., Tyzio, R., Josse, F., Grandgeorge, M., Hadjikhani, N., Ben-Ari, Y., 2012. A randomised controlled trial of bumetanide in the treatment of autism in children. Transl Psychiatry 2, e202. Lummis, S., 2012. 5-HT3 receptors. Journal of Biological Chemistry 287, 40239-40245. Lynch, J., 2009. Native glycine receptor subtypes and their physiological roles. Neuropharmacology 56, 303-309. Macdonald, R., Olsen, R., 1994. GABAA receptor channels. Annual Review of Neuroscience 17, 569-602. Marrosu, F., Marrosu, G., Rachel, M., Biggio, G., 1987. Paradoxical reactions elicited by diazepam in children with classic autism. Funct Neurol 2, 355-361. Mikkelsen, J., Bundzikova, J., Larsen, M., Hansen, H., Kiss, A., 2008. GABA regulates the rat hypothalamic-pituitary-adrenocortical axis via different GABAA receptor -subtypes. Ann NY Acad Sci 1148, 384-392. Mikkelsen, J., Søderman, A., Kiss, A., Mirza, N., 2005. Effects of benzodiazepines receptor agonists on the hypothalamic-pituitary-adrenocortical axis. Eur J Pharmacol 519, 223-230. Miledi, R., Eusebi, F., Martínez-Torres, A., Palma, E., Trettel, F., 2002. Expression of functional neurotransmitter receptors in Xenopus oocytes after injection of human brain membranes. Proceedings of the National Academy of Sciences of the United States of America 99, 13238- 13242. Nakazawa, K., Inoue, K., Ito, K., Koizumi, S., Inoue, K., 1995. Inhibition by suramin and reactive blue 2 of GABA and glutamate receptor channels in rat hippocampal neurons. Naunyn Schmiedebergs Arch Pharmacol 351, 202-208. Naviaux, J., Wang, L., Li, K., Bright, A., Alaynick, W., Williams, K., Powell, S., Naviaux, R., 2015. Antipurinergic therapy corrects the autism-like features in the Fragile X (Fmr1 knockout) mouse model. Molecular Autism 6, 1. Naviaux, R., 2018. Antipurinergic therapy for autism—An in-depth review. Mitochondrion in press. Naviaux, R. K., Curtis, B., Li, K., Naviaux, J., Bright, A., Reiner, G., Westerfield, M., Goh, S., Alaynick, W., Wang, L., Capparelli, E., Adams, C., Sun, J., Jain, S., He, F., Arellano, D., Mash, L., Chukoskie, L., Lincoln, A., Townsend, J., 2017. Low-dose suramin in autism spectrum disorder: a small, phase I/II, randomized clinical trial. Ann Clin Transl Neurol 4, 491–505.
Olsen, R., Sieghart, W., 2008. International Union of Pharmacology. LXX. Subtypes of - aminobutyric acid (A) receptors: classification on the basis of subunit composition, pharmacology, and function. Update. Pharmacological Reviews 60, 243-260.
Purohit, P., Auerbach, A., 2013. Loop C and the mechanism of acetylcholine receptor–channel gating. Journal of General Physiology 141, 467-478.
Putrenko, I., Zakikhani, M., Dent, J., 2005. A family of acetylcholine-gated chloride channel subunits in Caenorhabditis elegans. Journal of Biological Chemistry 280, 6392-6398.
Ralevic, V., Burnstock, G., 1998. Receptors for purines and pyrimidines. Pharmacological Reviews 50, 413–492.
Ranganathan, R., Cannon, S., Horvitz, H., 2000. MOD-1 is a serotonin-gated chloride channel that modulates locomotory behaviour in C. elegans. Nature 408, 470-475.
Spratt, E., Nicholas, J., Brady, K., Carpenter, L., Hatcher, C., Meekins, K., Furlanetto, R., Charles, J., 2012. Enhanced cortisol response to stress in children in autism. Journal of Autism and Developmental Disorders 42, 75-81.
Ueno, S., Bracamontes, J., Zorumski, C., Weiss, D., Steinbach, J., 1997. Bicuculline and gabazine are allosteric inhibitors of channel opening of the GABAA receptor. Journal of Neuroscience 17, 625-634.
Wagner, D., Czajkowski, C., 2001. Structure and dynamics of the GABA binding pocket: a narrowing cleft that constricts during activation. Journal of Neuroscience 21, 67-74.
Whiting, P., Wafford, K., McKernan, R., 2000. Pharmacological subtypes of GABAA receptors based on subunit composition. In: Martin, D., Olsen, R., (Eds), GABA in the Nervous System: The View at Fifty Years. Lippincott Williams & Wilkins, Philadelphia, pp. 113-126.
Wolstenholme, A., 2012. Glutamate-gated chloride channels. Journal of Biological Chemistry 287, 40232-40238.
Xu, X., Roberts, D., Zhu, G., Chang, Y., 2016. Competitive antagonists facilitate the recovery from desensitization of α1β2γ2 GABAA receptors expressed in Xenopus oocytes. Acta Pharmacol Sin 37, 1020-1030.
Zheng, Y., Hirschberg, B., Yuan, J., Wang, A., Hunt, D., Ludmerer, S., Schmatz, D., Cully, D., 2002. Identification of two novel Drosophila melanogaster histamine-gated chloride channel subunits expressed in the eye. Journal of Biological Chemistry 277, 2000-2005.