The antihelminthic moxidectin enhances tonic GABA currents in rodent hippocampal pyramidal neurons

Abstract

Macrocyclic lactones (MLs) are commonly used treatments for parasitic worm and insect infections in humans, livestock, and companion animals. MLs target the invertebrate glutamate-activated chloride channel that is not present in vertebrates. MLs are not entirely inert in vertebrates, though; they have been reported to have activity in heterologous expression systems consisting of ligand-gated ion channels that are present in the mammalian central nervous system (CNS). However, these compounds are typically not able to reach significant concentrations in the CNS because of the activity of the blood-brain barrier P-glycoprotein extrusion system. Despite this, these compounds are able to reach low levels in the CNS that may be useful in the design of novel “designer” ligand-receptor systems that can be used to directly investigate neuronal control of behavior in mammals and have potential for use in treating human neurological diseases. To determine whether MLs might affect neurons in intact brains, we investigated the activity of the ML moxidectin (MOX) at native GABA receptors. Specifically, we recorded tonic and phasic miniature inhibitory postsynaptic currents (mIPSCs) in ex vivo brain slices. Our data show that MOX potentiated tonic GABA currents in a dose-dependent manner but had no concomitant effects on phasic GABA currents (i.e., MOX had no effect on the amplitude, frequency, or decay kinetics of mIPSCs). These studies indicate that behavioral experiments that implement a ML-based novel ligand-receptor system should take care to control for potential effects of the ML on native tonic GABA receptors.

NEW & NOTEWORTHY We have identified a novel mechanism of action in the mammalian central nervous system for the antihelminthic moxidectin, commonly prescribed to animals worldwide and currently being evaluated for use in humans. Specifically, moxidectin applied to rodent brain slices selectively enhanced the tonic GABA conductance of hippocampal pyramidal neurons.

INTRODUCTION

Macrocyclic lactones (MLs) are a class of antihelminthic compounds that are commonly used in the treatment of parasitic worms in humans, livestock, and companion animals. These compounds, such as ivermectin (IVM), moxidectin (MOX), and others, are derived from avermectin, which is produced by the soil-dwelling bacterium Streptomyces avermectinius. Avermectin was originally discovered through a collaboration between Merck & Co. Inc. and the Kitasato Institute (Campbell 2012; Omura 2008). The profound success of the use of IVM in human and veterinarian medicine resulted in the awarding of the 2015 Nobel Prize in Physiology or Medicine to Satoshi Omura and William C. Campbell (https://www.nobelprize.org).

MLs act to control parasitic infections through the activation of glutamate-gated chloride channels (GluCls) that are critical for proper muscle control in invertebrates. These channels are completely absent in vertebrate nervous systems, thus making them an attractive target for the design of interventions that are not likely to produce significant side effects. However, MLs are known to bind to and allosterically modulate a number of other ionotropic, ligand-gated transmembrane receptors that are present in mammals including γ-aminobutyric acid type A (GABA_A) receptors (Krůsek and Zemková 1994), glycine receptors (Shan et al. 2001), neuronal α₇-nicotinic receptors (Krause et al. 1998) and purinergic P_2X4 receptors (Khakh et al. 1999). The successful use of MLs in treating infection with few to no adverse side effects is thought to be due to their relatively low affinity for these receptors compared with the GluCls, but at least some of the lack of side effects can be attributed to the active extrusion of avermectins from the central nervous system (CNS) by the blood-brain barrier protein P-glycoprotein (Kiki-Mvouaka et al. 2010; Schinkel et al. 1994). Although considered extremely safe, and having been administered to hundreds of millions of people and animals worldwide, MLs are lethal in individuals carrying mutations in the P-glycoprotein, which has been clearly demonstrated in some collies and inbred strains of mice (Lankas et al. 1997; Mealey et al. 2002; Ménez et al. 2012; Paul et al. 2000; Prichard et al. 2012).

MOX, a chemically derived analog of nemadectin, is generally better tolerated than IVM (Cobb and Boeckh 2009; Ménez et al. 2012; Prichard et al. 2012; Yang 2012) and has a slightly higher affinity for an engineered “designer” receptor-ligand system that makes it an attractive alternative to IVM for use in experiments where the addition of a heterologous or artificial control system to inhibit neuronal firing is desirable (Lynagh and Lynch 2010). The specific effects of MOX on mammalian ligand-gated transmembrane receptors in the CNS is relatively unknown. However, a large and sometimes inconsistent body of work has been directed toward determining the effect of MLs on mammalian GABA_A receptors, often utilizing heterologous systems (Geary and Moreno 2012; Zemkova et al. 2014). For instance, avermectin B_1a has been shown to enhance receptor sensitivity to GABA in Xenopus oocytes injected with mRNA isolated from chick brains (Sigel and Baur 1987), and in cultured rat cerebellar granule neurons avermectin B_1a enhanced ³⁶Cl⁻ uptake when coapplied with GABA, and this effect was blocked by GABA_A-receptor antagonists (Huang and Casida 1997), while in human embryonic kidney (HEK) 293 cells expressing rat recombinant α₁β₂γ_2S subunit-containing GABA_A receptors IVM was shown to directly activate a chloride conductance at micromolar concentrations with no coapplication enhancement of GABA currents (Adelsberger et al. 2000). In contrast, whole cell, exogenously applied, GABA-activated currents in cultured cortical and hippocampal neurons from embryonic rodents were potentiated by coapplication of IVM (Dawson et al. 2000; Krůsek and Zemková 1994). Likewise, additional studies have found that coapplication of a variety of avermectins enhanced GABA-activated currents in α₁β₂γ₂ subunit-expressing Xenopus oocytes (Dawson et al. 2000; Ménez et al. 2012).

Despite these inconsistencies in the exact mechanism of the MLs’ actions on mammalian GABA_A receptors and their effective extrusion from the CNS by the blood-brain barrier P-glycoprotein, several MLs have been investigated for potential in vivo activity as well as developed as a designer ligand-receptor tool for investigation of behavior in awake animals. Dawson et al. have shown that several ML analogs provide protection from pentylenetetrazole-induced seizures in mice (Dawson et al. 2000), thereby demonstrating a behavioral effect on native receptors in the CNS of wild-type animals. Taking advantage of the fact that these compounds cross the blood-brain barrier at a sufficiently high level to activate native receptors despite the activity of P-glycoprotein, the invertebrate GluCl channel was optimized for expression in mammalian neurons and mutagenized to reduce sensitivity to glutamate (Li et al. 2002; Slimko et al. 2002; Slimko and Lester 2003). This system has been used in combination with systemic administration of IVM to control behavior in mice expressing GluCl (Lerchner et al. 2007). A similar approach has been designed using the human α subunit of the glycine receptor that is suggested to offer certain technical advantages over the GluCl approach (Lynagh and Lynch 2010).

MATERIALS AND METHODS

All procedures were performed with protocols approved by the University of Utah Animal Care and Use Committee and in accordance with National Institutes of Health guidelines for the care and use of laboratory animals. CD-1 mice and Sprague-Dawley rats were maintained on a 12:12-h light-dark cycle with ad libitum access to food and water. Both male and female mice aged 1–2 mo and rats aged 14–27 days were anesthetized with isoflurane before decapitation. Brains were removed rapidly and transferred to ice-cold choline chloride artificial cerebrospinal fluid (aCSF) containing (in mM) 118 C₅H₁₄ClNO, 2.5 KCl, 2.5 CaCl₂, 10 MgCl₂, 1.2 NaH₂PO₄, 10 glucose, 3 kynurenic acid, and 25 NaHCO₃, bubbled with carbogen (95% O₂-5% CO₂) to yield pH 7.3–7.4. Coronal slices (350 μm) containing the dorsal CA1 region of the hippocampus were cut with a Leica VT-1000S Vibratome (Leica Microsystems, Wetzlar, Germany) and transferred to a holding chamber containing normal aCSF (in mM): 125 NaCl, 3 KCl, 1.3 CaCl₂, 1.3 MgSO₄, 1.2 NaH₂PO₄, 25 glucose, 3 kynurenic acid, and 25 NaHCO₃, and bubbled at 32–34°C, where they were kept for at least 1 h before experimentation.

Slices were transferred individually to a submerged recording chamber perfused at ~3 ml/min with carbogen-gassed recording solution [normal aCSF + 1 μM tetrodotoxin (TTX) + 1:1,000 DMSO). Bath temperature was maintained at 34 ± 2°C with an inline heater. Microelectrodes (3–6 MΩ) were pulled from borosilicate glass and filled with a solution containing (in mM) 135 Cs-methanesulfonate, 8 NaCl, 10 HEPES, 2 Mg₂-ATP, 0.3 Na₃-GTP, 0.5 EGTA, and 7 phosphocreatine (pH = 7.3, ~300–310 mosM). Neurons were visually identified with infrared differential interference contrast optics on an upright microscope. Whole cell recordings were collected from pyramidal neurons in the dorsal CA1 area with Clampex software (Molecular Devices, Sunnyvale, CA) and a MultiClamp 700B (Molecular Devices). Signals were filtered at 3 kHz and digitized at 20 kHz by a Digidata 1322A (Molecular Devices).

Whole cell recordings of miniature inhibitory postsynaptic currents (mIPSCs) were obtained by voltage clamping the cell at +10 mV (with series resistance compensation) while blocking Na⁺ channel-dependent transmitter release with 1 μM TTX. Once the baseline holding current reached a steady level (~10–15 min after depolarization), recordings began. After a stable 4-min baseline, perfusion solution was switched to an identical solution containing MOX. Drug was applied for 10 min, followed by a direct bath application of 10 mM picrotoxin (PTX). Tonic current was measured as the peak of the all-points histogram of the recording during the last minute of the baseline, MOX, and PTX periods.

Events were detected with the Clampfit template matching function. Decay kinetics were determined by first compiling 40–60 individual (nonoverlapping) events from the last 2 min of the baseline and MOX periods. The composite average of each period was then fit with a standard double exponential equation:

$$f(x)=A_1{\hspace{0.17em}\times \mathrm{e}}^{\left(-t\text{/}{\tau }_1\right)}+A_2\times {\text{e}}^{\left(-t\text{/}{\tau }_2\right)}$$

where A is the amplitude of the current that decayed with the corresponding τ. The weighted τ (τ_w) was calculated from the following equation:

$${\tau }_w=\left[A_1/\left(A_1\text{+}{A}_2\right)\right]\times {\tau }_1+\left[A_2/\left(A_1\text{+}{A}_2\right)\right]\times {\tau }_2$$

Values given in the text are averages ± SD. When measurements were compared before and after drug application, a paired t-test was used to test for significance. When between-groups measurements were compared, an ANOVA was used to test for significance.

RESULTS

Moxidectin potentiates tonic GABA currents in a dose-dependent manner.

MLs have previously been shown to activate or potentiate mammalian GABA_A receptors in heterologous receptor expression systems or dissociated neurons. We therefore tested the effect of MOX, a derivative of nemadectin (Prichard et al. 2012), on naturally occurring GABA_A receptors in live neurons in brain slices prepared from adult and juvenile rodents. mIPSCs were recorded in the whole cell patch-clamp configuration from CA1 pyramidal neurons in the dorsal hippocampus. Bath perfusion of 1 μM MOX induced a steadily increasing outward shift in the baseline holding current, consistent with the activation of tonic conducting GABA_A receptors (Fig. 1, A1 and A2). After a 10-min application of MOX, the GABA_A receptor antagonist PTX, added directly to the bath, inactivated GABA_A receptors and revealed the portion of the holding current that was contributed by the activation/potentiation of these receptors (Fig. 1A3). The measured tonic GABA current can be seen as a shift in the all-points histograms calculated during the baseline, MOX, and PTX periods of the recording (Fig. 1A4). All-points histograms are a conversion of the raw whole cell recordings during a 1-min period of time into histogram format, and the peak of these histograms can be read as the average holding current that is necessary to maintain voltage clamp (Glykys and Mody 2007). The histograms are plotted on the same axis and therefore demonstrate the positive shift in holding current induced by application of MOX and the negative shift induced by PTX (relative to baseline). The magnitude of the shift from baseline to MOX is fully reversed by the application of PTX, and therefore the induced current is defined as PTX-sensitive tonic GABA. If no new current is induced by application of MOX, there should be no shift in the baseline holding current during drug application. Indeed, application of a lower concentration of MOX revealed the effect to be dose dependent, as 10 nM MOX had no effect on the tonic GABA_A current (Fig. 1B).

Fig. 1.

MOX enhances tonic GABA currents in hippocampal pyramidal neurons. Tonic and phasic GABA currents were recorded in CA1 pyramidal neurons held at +10 mV. A: examples of 1 min of raw data from each of the measured periods, baseline (A1), 10 min after perfusion of 1 μM MOX (A2), and after direct bath application of a saturating concentration of picrotoxin (A3), demonstrate the shift in holding current due to activation of tonic GABA_A currents and block thereof. The all-points histograms of the whole current recording during these periods (A4) demonstrate that MOX induces an increase in the tonic GABA_A current to 296 pA that is blocked by picrotoxin (which also reveals the baseline tonic current of 76 pA). B: in a separate cell, the lower concentration of 10 nM MOX did not increase the tonic GABA current (B2), demonstrating that the effect of MOX is dose dependent. All-points histograms (B4) of the baseline whole current (B1) and 10 min after application of MOX (B2) compared with picrotoxin (B3) demonstrate that the tonic current was 31 pA at baseline and 32 pA in the presence of MOX.

This dose-dependent effect of MOX was quantified for 10 nM, 100 nM, 300 nM, and 1 μM concentrations and can be seen in the bar graphs comparing PTX-sensitive current recorded in the absence and presence of MOX in Fig. 2A. The 1 μM concentration induced an average 3.7-fold increase in the average tonic GABA current (n = 10 pyramidal neurons; baseline = 44.5 ± 30.7 pA, 1 μM MOX = 165.4 ± 109.5 pA; P < 0.05, paired t-test). The 300 nM concentration induced an average 4.6-fold increase in the average tonic GABA current (n = 11 pyramidal neurons; baseline = 42.2 ± 36.2 pA, 300 nM MOX = 198.5 ± 107.2 pA; P < 0.05, paired t-test). The 100 nM concentration induced an average 2.2-fold increase (n = 13 pyramidal neurons; baseline = 44.6 ± 26.5 pA, 100 nM MOX = 78.5 ± 49.1 pA; P < 0.05, paired t-test). The 10 nM concentration had no significant effect on tonic GABA currents (n = 10 pyramidal neurons; baseline = 41.0 ± 33.8 pA, 10 nM MOX = 38.4 ± 24.4 pA). A fit of the dose-response curve (Fig. 2B) determined the EC₅₀ to be 152 nM MOX. The values plotted in Fig. 2B are the MOX-induced tonic current that was calculated as the average PTX-sensitive current during baseline subtracted from the average PTX-sensitive current at 10 min after MOX application for each concentration. When the magnitudes of the MOX-induced currents were compared, we observed a significant increase from 10 nM to 300 nM, from 10 nM to 1 μM, from 100 nM to 300 nM, and from 100 nM to 1 μM (1-way ANOVA, P < 0.05, Tukey’s multiple comparison test).

Fig. 2.

Effects of MOX on tonic GABA are dose dependent. A: averages ± SD for the baseline tonic GABA current in hippocampal pyramidal neurons (solid bars) and the shift in tonic GABA current after 10-min application of different concentrations of MOX (checkered boxes). Asterisks indicate there was a significant increase after 100 nM, 300 nM, and 1 μM MOX (P < 0.05, individual paired t-tests on the before and after measurements within each group). B: MOX-induced current was calculated as the average picrotoxin-sensitive current in the presence of MOX minus the average picrotoxin-sensitive current in the absence of MOX and plotted against the log of the concentration to fit a dose-response curve (solid line). The fit of these data resulted in an EC₅₀ of 152 nM MOX.

To determine whether these results were sufficiently powered, we conducted a post hoc power analysis using G*Power 3 (Faul et al. 2007). Based on the comparison of the means and standard deviations, the 1 μM data set produced an effect size of 1.5, and this yielded a statistical power of 0.99 (α = 0.05, n = 10 pairs, 1-tailed comparison of dependent means). Similarly, the 300 nM data set produced an effect size of 1.9 and a power of 0.99 (α = 0.05, n = 11 pairs, 1-tailed comparison of dependent means). The 100 nM data set produced an effect size of 0.8 and a power of 0.88 (α = 0.05, n = 13 pairs, 1-tailed comparison of dependent means). Finally, the 10 nM data set produced an effect size of 0.1 and a power of 0.1 (α = 0.05, n = 10 pairs, 1-tailed comparison of dependent means). At 10 nM there was no significant difference, and we therefore accept the null hypothesis that this concentration does not affect tonic GABA currents in CA1 pyramidal neurons. To further strengthen this conclusion, we conducted an a priori sample size estimate based on the effect size observed in our data set, and this calculation revealed that we would need ~597 pairs to confidently accept a difference of this small magnitude.

Moxidectin does not affect phasic GABA currents.

In contrast to the effect of MOX on tonic GABA_A currents, MOX had no effect on phasic GABA_A currents at any of the concentrations tested (Fig. 3). We found no change in the interevent interval, amplitude, or decay kinetics of mIPSCs after bath application of any concentration of MOX compared with baseline (P > 0.05, 1-way ANOVA for each parameter). To compile data from rats and mice at different ages, we normalized the postdrug values to the baseline values for each measurement to determine the relative change, if any, induced by the addition of MOX. At the highest concentration tested (1 μM), mIPSC amplitude was 0.93 ± 0.09 of the baseline after MOX treatment. The interevent interval in the presence of 1 μM MOX was 1.08 ± 0.29 of the baseline value, and the mIPSC decay kinetics in the presence of 1 μM MOX was 1.19 ± 0.24 of the baseline decay rates. At 300 nM MOX the fold change in interevent interval was 1.12 ± 0.31, amplitude was 0.85 ± 0.12, and decay rates were 1.05 ± 0.29. At 100 nM MOX the fold change in interevent interval was 1.01 ± 0.19, amplitude was 0.97 ± 0.0.9, and decay rates were 1.1 ± 0.17. Finally, at 10 nM MOX the fold change in interevent interval was 0.92 ± 0.15, amplitude was 0.97 ± 0.07, and decay rates were 1.04 ± 0.18.

Fig. 3.

MOX did not change phasic GABA currents in hippocampal pyramidal neurons. A: raw traces as in Fig. 1 expanded to demonstrate the mIPSCs (upward deflections) recorded during baseline (left) and 10 min after 1 μM MOX application (right). Amplitude, interevent interval, and decay kinetics of mIPSCs recorded during each period can be seen in the further expanded data at bottom. B: phasic properties of mIPSC amplitude, interevent interval, and decay kinetics are quantified for each concentration of MOX normalized to the baseline recording and plotted as means ± SD. In each case there were no significant changes in the presence of MOX at any concentration.

DISCUSSION

The primary finding of this study is that MOX, in a dose-dependent manner, selectively enhances tonic GABA_A currents and has no effect on phasic GABA_A currents. This mechanism of action for MOX was determined utilizing the naturally occurring varieties of GABA_A receptors that are found in hippocampal principal neurons in ex vivo rodent brain slices.

Does MOX potentiate tonic GABA through a presynaptic mechanism?

To our knowledge, there are no specific data regarding the activity of MOX at purinergic receptors; however, IVM is well known to be a specific allosteric modulator of the P_2X4 receptor (Coddou et al. 2011; Khakh et al. 1999; Priel and Silberberg 2004; Zemkova et al. 2014). Potentiation of ATP-induced P_2X4 currents by IVM has been shown to increase synaptic release (Lalo et al. 2007; Vavra et al. 2011). It is therefore possible that MOX, acting through P_2X4 receptors, induced an increase in synaptic release of GABA resulting in an enhancement of tonic GABA currents. However, in each of the studies demonstrating an effect of IVM on P_2X4-mediated GABA release, the effect of application of IVM was an increase in the frequency of phasic postsynaptic currents. We did not observe a similar change in the frequency of phasic currents in this study; therefore, it is more likely that the effect of MOX on tonic inhibition was through the direct potentiation of GABA-activated currents at the extra- and perisynaptic GABA_A receptors.

Does MOX target specific GABA_A receptors?

GABA_A currents are mediated through pentameric ligand-gated chloride channels that are assembled from 19 different genes/subunits. Functionally, these receptors provide either a short-duration, fast-decaying current, termed phasic, that is the inhibitory postsynaptic current resulting from the synaptic release of GABA, or a noninactivating long-lasting tonic current associated with extra- and perisynaptic localization of channels (Farrant and Nusser 2005). Tonic GABA-activated currents in CA1 pyramidal neurons are conducted by α₅ subunit-containing GABA_A receptors (Caraiscos et al. 2004; Glykys and Mody 2006; Martin et al. 2010; Sperk et al. 1997;). Therefore, our data offer one possible explanation for the discrepancies reported for the activity of MOX or other MLs at GABA_A receptors in heterologous expression systems (see introduction); that is, these studies have typically used synaptic receptor subunit compositions containing the α₁ subunit. Dawson et al. have reported a differential affinity of MLs for GABA receptors containing different β subunits, with a fivefold increase in selectivity conferred by β₃-containing receptors (Dawson et al. 2000). Indeed, the α₅ subunit is thought to naturally pair with the β₃ subunit in the hippocampus (Sur et al. 1998). It is therefore likely that MOX specifically potentiates tonic GABA currents through α₅β₃-containing receptors.

Is MOX likely to potentiate tonic GABA currents in vivo at supratherapeutic doses?

After subcutaneous injection of MOX in wild-type mice with the typical treatment dose of 0.2 mg/kg, brain concentration of MOX reaches 6 ng/g at 24 h (Kiki-Mvouaka et al. 2010). Assuming that the brain weight of an adult mouse is ~0.5 g and the volume is measured at ~500 mm³ or 0.5 ml (Maheswaran et al. 2009; Vincent et al. 2010), we estimate that the brain concentration of MOX at 24 h after a normal treatment dose is ~9 nM. Our data demonstrate that this concentration would be insufficient to activate tonic GABA_A currents in hippocampal neurons. However, when the blood-brain barrier P-glycoprotein is not functional, as in mdr1ab(−/−) mice, MOX brain concentrations can increase 10-fold at 24 h after treatment to 65 ng/g or ~102 nM (Kiki-Mvouaka et al. 2010). This is the range at which we begin to see a small enhancement of tonic GABA_A currents. Impaired motor coordination has been reported in wild-type rats administered 10- to 100-fold higher than normal treatment doses (Rodrigues-Alves et al. 2009), which is within the concentration range where we confidently observed a significant enhancement of tonic GABA current. These data suggest that higher systemic doses of MOX may increase brain concentrations to levels that are sufficient to affect behavior. Additional CNS effects have been reported for systemic doses well above the therapeutic range for numerous MLs screened in mice for efficacy in the pentylenetetrazole-induced seizure model and rotarod test (Dawson et al. 2000).

In conclusion, our data could be considered as a guide for future experiments involving the use of designer ligand-receptor-based systems for behavioral studies specifically involving the administration of MLs or MOX. In this regard, MOX should be given at concentrations low enough to avoid complications induced by activation of native tonic GABA_A receptors.

GRANTS

This work was supported by National Institute of Neurological Disorders and Stroke Grant NS-079274.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.S. and F.E.D. conceived and designed research; J.S. and A.G. performed experiments; J.S. and A.G. analyzed data; J.S. and F.E.D. interpreted results of experiments; J.S. prepared figures; J.S. drafted manuscript; J.S., A.G., and F.E.D. edited and revised manuscript; J.S., A.G., and F.E.D. approved final version of manuscript.