The role of GABA_A receptors in mediating the effects of alcohol in the central nervous system

Abstract

Ethanol is a chemically simple compound that produces many well-known effects in humans. The prevailing idea for many years was that ethanol and other alcohols exerted their effects on the central nervous system (CNS) by non-selectively disrupting the lipid bilayers of neurons. However, in recent years, there has been an accumulation of evidence pointing to the importance of ligand-gated ion channels (LGICs) in mediating the effects of ethanol. Of these LGICs, γ-aminobutyric acid type A (GABA_A) receptors appear to occupy a central role in mediating the effects of ethanol in the CNS. GABA is the primary inhibitory neurotransmitter in the mammalian CNS, and activation of GABA_A receptors by GABA tends to decrease neuronal excitability. This article reviews several aspects of GABA_A receptor and ethanol interactions, including the evidence for short- and long-term modulation of GABA_A receptors by ethanol and evidence for a GABA_A receptor-related genetic component of alcoholism.

Introduction

Alcohol is the most frequently abused drug in our society. The abuse of this substance has a societal impact not only in terms of the general health of the population, but also in economic output, where it is estimated that loss of productivity costs many billions of dollars each year.¹ Alcohol abuse has many long-term effects that result in premature death and an increased propensity for serious illness.² Additionally, fetal alcohol syndrome is a major health issue worldwide^3,4,5,6,7 and often leads to health problems throughout life.^8,9 Another alcohol-related health issue is ethanol–drug interactions which, in some cases, can be fatal.^10,11,12

Alcohol, or more specifically, ethanol, produces several effects in humans. It is a central nervous system (CNS) depressant and shares many of the effects of other CNS depressants, such as sedatives, hypnotics and anesthetic agents. Although it has dramatic effects on the CNS, ethanol is not a particularly potent drug. For example, threshold effects generally do not occur until the concentration of ethanol in the blood is relatively high (5–10 mmol/L).¹³ However, a single serving of a strong spirit can contain as much as 12 g of ethanol, and thus it is possible to consume large amounts of ethanol in a single sitting and quickly reach these blood concentrations.

The effects of ethanol in humans are well documented.¹³ At low blood concentrations, there is a feeling of euphoria or disinhibition. As the concentration of ethanol in the blood increases, motor function is impaired and speech becomes slurred. With blood alcohol concentrations between 200 mg/dL and 300 mg/dL (i.e., between 43 mmol/L and 65 mmol/L), vomiting can occur and the subject can fall into a stupor. Blood concentrations higher than this can result in coma, and at high blood concentrations (500 mg/dL or 109 mmol/L), there is the potential for respiratory failure and death.¹⁴

For many years, it was believed that these effects were mediated through the non-specific disruption of neuronal lipid bilayers by ethanol. However, it is now generally accepted that ethanol acts by binding with and altering the function of specific proteins, particularly membrane-bound ligand-gated ion channels and voltage-dependent ion channels.^15,16,17,18 Additionally, there is mounting evidence that ethanol can alter the function of second-messenger proteins.^19,20

Membrane-bound proteins and ethanol

The 2 major types of membrane-bound proteins that are directly affected by pharmacologically relevant concentrations of ethanol (i.e., concentrations up to 100 mmol/L or 460 mg/dL, at which point ethanol can be lethal in humans) are ligand-gated ion channels (LGICs) and voltage-dependent calcium channels.²¹ LGICs are a family of neurotransmitter receptors that are widely distributed in the mammalian CNS and play a major role in synaptic transmission and the regulation of neuronal excitability. In particular, the gamma-aminobutyric acid type A (GABA_A), N-methyl-D-aspartate (NMDA), glycine,²² neuronal nicotinic²³ and 5-hydroxytryptamine type 3 (5-HT₃) receptors²⁴ are LGICs that have been shown to be directly modulated by ethanol. Ethanol does not modulate these receptors in a non-specific fashion; interestingly, it potentiates ligand-gated currents at some receptors but inhibits them at others. For example, it is well documented that acute ethanol exposure potentiates these currents at GABA_A and glycine receptors²⁵ but inhibits them at NMDA receptors.²⁶ Voltage-gated calcium channels play key roles in neurotransmitter release, hormone secretion, gene regulation and differentiation. Ethanol, when administered acutely, has been shown to block voltage-gated calcium channels at pharmacologically relevant concentrations.^27,28

Because of ethanol's capacity to change neuronal excitability by interacting with the proteins mentioned above, it is not surprising that after long-term exposure to ethanol, neurons adapt to its presence by altering the expression of their complement of these receptors and ion channels.

Current therapies to treat alcohol addiction

Current therapeutics for alcohol-related health problems include the drugs disulfiram, naltrexone and acamprosate.^29,30

Disulfiram inhibits aldehyde dehydrogenase, one of the enzymes involved in alcohol metabolism, which converts acetaldehyde to acetic acid.^31,32 Administration of disulfiram in the absence of alcohol has little or no effect. However, if alcohol is consumed, there is a build up of acetaldehyde, and this results in several unpleasant effects such as tachycardia, nausea, vomiting and hyperventilation, often accompanied by feelings of anxiety or panic.³³ Compliance taking this drug is often low.

Evidence for a link between opioid receptors and alcohol has been shown, and it has been postulated that opioid receptors are involved in the reinforcing effects of chronic ethanol consumption.³⁴ With this idea in mind, naltrexone, an opioid receptor antagonist, was tested for its effects on ethanol consumption and proved to be effective in decreasing it.^35,36 Naltrexone has been described as an anti-craving medication because clinical trials have shown that it helps relieve the urge that alcoholics have to consume ethanol.

Acamprosate is a relative newcomer as a therapy for alcohol abuse and is best used to maintain abstinence in patients who have stopped drinking.³⁷ Acamprosate (N-acetylhomotaurine) has a structure similar to GABA and has been shown to interact with presynaptic GABA_B receptors, increasing the release of GABA from presynaptic terminals.³⁸ Additionally, it appears to inhibit calcium ion influx through voltage-dependent calcium channels and NMDA receptors.³⁹ The 5-HT₃ receptor antagonist, ondansetron, has shown some efficacy in the treatment of early-onset alcoholism. This group of patients is considered to have a biological predisposition to alcohol abuse. It has been postulated that ondansetron acts by reducing the craving for alcohol.⁴⁰

Benzodiazepines are often used to treat some of the symptoms of alcohol abuse, especially during withdrawal, but have yet to be shown as effective drugs in treating addiction itself.^41,42

Ethanol and GABA_A receptors

As described above, ethanol interacts with and modifies the function of a number of membrane-bound proteins. However, it is likely that some play more important roles than others in mediating the effects of ethanol in the CNS.¹⁵ There is a large body of evidence showing that GABA_A receptors play central roles in both the short- and long-term effects of ethanol in the CNS.⁴³ GABA_A receptors belong to a family of transmembrane ligand-gated ion channels that includes the nicotinic acetylcholine, glycine and 5-HT₃ receptors.⁴⁴ These receptors are responsible for rapid neuronal transmission in the mammalian CNS. GABA_A receptors primarily occur in the postsynaptic membrane, although there is evidence that certain subtypes may occur extra-synaptically.⁴⁵ These receptors are the site of action of a number of drugs, including barbiturates, benzodiazepines and anesthetics. The GABA_A receptors are pentameric receptors with a high degree of homology with nicotinic receptors.^46,47 GABA_A receptors have a multitude of subunits, including 6!#!alpha;, 4β, 3γ, 2ρ, δ, ε, Θ and p.⁴⁸ They have a rich pharmacology, and this is dependent upon the particular subunits that are present within the receptor pentamer.^49,50 Each subunit has an extracellular N-terminal domain that typically contains ligand-recognition sites and 4 membrane-spanning domains.⁵¹ The second membrane-spanning domain forms the lining of the ion channel. The subunits are arranged in a radial fashion such that they surround a central ion pore that opens in the presence of ligand. Once the ion channel opens, ion transport follows the electrochemical gradient that is established across the neuronal membrane. In the case of GABA_A receptors, the ion pore conducts chloride ions. GABA is classified as an inhibitory amino acid neurotransmitter because the influx of chloride ions into the postsynaptic cell after the activation of these receptors moves the postsynaptic membrane potential further away from its firing threshold. The discrete distribution of receptor subtypes suggests that each has a specific function within the CNS.^52,53 In mammalian tissue, the most common receptor subtype contains α1, β2 and γ2 subunits.

Within each subunit is a large N-terminal domain, which includes clusters or “loops” of amino acids thought to form the neurotransmitter binding domain(s),^54,55 4 transmembrane domains, of which the second (TM2) is thought to line the ion channel, and a large intracellular loop that contains consensus sites for phosphorylation by various kinases.^47,56 Interestingly, these receptors not only bind their neurotransmitter ligands, but can also interact with a number of compounds that bind at distant sites on the protein and allosterically modulate the actions of the neurotransmitter.^57,58 These modulatory agents include benzodiazepines, barbiturates, neurosteroids and anesthetics. With some exceptions, these modulators are not able to activate the receptors unless GABA is present and bound to the receptor. Indeed, the capacity of some of these compounds to modulate LGICs, and thus neuronal function, has been exploited in the development of therapeutic agents. Typically, those modulators that enhance the actions of GABA at the receptor have sedative or hypnotic effects. Benzodiazepines, one of the most commonly prescribed groups of psychoactive drugs, are positive allosteric modulators of GABA_A receptors.⁵⁹ Clinically useful benzodiazepines and benzodiazepine-like compounds act by potentiating the effects of GABA at the receptor. This is manifested as an increased probability of channel opening, thus leading to a greater hyperpolarization of the postsynaptic membrane and a further decrease in neuronal excitability than would be seen with GABA activation alone. Benzodiazepines are used to treat clinical symptoms such as anxiety, convulsions, muscle tension and insomnia. Not all ligands for this site are positive allosteric modulators, however. In fact, ligands for this particular binding site show a spectrum of efficacies.⁶⁰

Apart from positive allosteric modulators, there are those that act as negative allosteric modulators by decreasing the probability of channel opening events in the presence of GABA. In animal models, these compounds lead to anxiety, seizures or both; other compounds have no intrinsic efficacy at this site and are therefore classified as antagonists.⁴⁴

When GABA_A receptors are acutely exposed to ethanol, there is a potentiation of GABA-gated current.⁶¹ Some of the first experiments showing that the ionic flux generating this current is increased by ethanol were done using native receptors in rat brain microsacs and synaptoneurosome preparations. Exposure of these preparations to ethanol increased GABA-gated chloride uptake by as much as 260%.⁶² Subsequently, the effects of ethanol on recombinant receptors expressed in Xenopus oocytes were examined. These effects generally occur in a pharmacologically relevant range of concentrations (below 100 mmol/L) and potentiate the actions of GABA at GABA_A receptors. However, this issue is somewhat contentious; some laboratories report seeing no effect of alcohols at GABA_A receptors, and the range of concentrations at which effects are seen also seems to vary.^63,64,65 A study using α1β2γ2 receptors showed that ethanol potentiation occurred, but only in those receptors in which a particular splice variant of the γ2 subunit was present.⁶⁶ The “long” version of γ2 (γ2L), so-called because of the presence of an extra 8 amino acids in the intracellular loop, was deemed necessary for potentiation to occur. However, other laboratories have been unable to duplicate this finding and have shown that the effects of ethanol appear to be independent of the type of γ2 subunit that is contained within the oligomer.⁶⁷ So far, there appears to be no subunit dependence with respect to the effects of ethanol on GABA_A receptors,⁶⁸ with the exception of the α4β1δ subtype, which seems to be more sensitive to ethanol than other receptor subtypes tested to date.⁶⁹

Originally, it was believed that alcohols and anesthetics acted by non-selectively disturbing the lipid bilayers of neurons, thus exerting a global disruption in neuronal membrane protein activity. This idea was based on evidence showing that the lipid solubility of anesthetic agents and their potency was highly correlated.^70,71,72 One aspect of anesthesia not easily explained by the lipid hypothesis was the cut-off effect seen with many series of anesthetic agents, in which the carbon length of the anesthetic agent incrementally increased.⁷³ Typically, the potencies of these agents increase until a certain backbone length is reached, at which point the anesthetic no longer shows any effect. It was originally proposed that this phenomenon was due to anesthetics over a particular size or carbon length having poor solubility in lipid bilayers.^71,73

A turning point in the idea of alcohol and anesthetic targets was provided by experiments showing that the activity of a soluble enzyme, luciferase, could be modified by alcohols.⁷⁴ Further, this modification showed distinct “cut-offs” where, after a certain chain length was reached, the inhibitory effects were lost or did not increase further.⁷⁴ (See below.) As this protein was soluble and not associated with lipid, the best explanation for this phenomenon was that the protein contained a binding pocket of fixed dimensions that could only accommodate alcohols of a certain size. Additionally, it has been shown that long-chain anesthetics are able to partition into membranes long after they have lost their anesthetic effect,⁷⁵ thus negating the idea that the lipid cut-off was due to a partitioning effect. Other evidence for the protein theory comes from experiments showing that optical isomers of anesthetic agents, while having the same bilayer disrupting properties, nonetheless show very different effects on LGICs.^76,77 Additionally, agents that can disrupt the lipid bilayer do not necessarily induce anesthesia.⁷⁸

With respect to GABA_A receptors, a number of alcohols with different carbon numbers in their backbone (indicated in parenthesis), including methanol (3), butanol (4), hexanol (6), octanol (8), decanol (10) and dodecanol (12), potentiate GABA-gated current. The potencies of these alcohols tend to increase with chain length. However, once the length of the backbone exceeds 12 carbons, there is no longer an effect on GABA-gated current. Interestingly, dodecanol is the longest n-alcohol that is able to bring about a loss of the righting reflex in tadpoles.⁷⁹ This lack of effect with n-alcohols longer than dodecanol has been taken to show that there is a binding pocket with a defined size that is able to accommodate alcohols with carbon lengths up to 12, with those exceeding 12 being excluded on the basis of size.⁸⁰ Similar molecular cut-offs have been found for neuronal nicotinic, NMDA⁸¹ and AMPA receptors,⁸² although other studies suggest that this cut-off effect may be due to poor aqueous solubility.⁸³

A study of the single-channel properties of GABA_A receptors revealed that ethanol-induced potentiation of GABA-gated currents is due to an increase in the frequency and duration of channel opening and an increase in channel bursts and burst duration.⁸⁴ Additionally, the amount of time that the channel spends in the closed state is decreased. The summation of these effects leads to increased ion flux through the open channel in the presence of GABA and ethanol.⁸⁴

Evidence for a putative ethanol binding site on GABA_A receptors was provided by a study from Mihic et al.⁸⁵ Ethanol potentiates glycine-gated chloride current through glycine receptors composed of glycine α subunits, but it inhibits GABA-gated current through receptors composed entirely of the ρ subunit. By creating a series of chimeras between these 2 subunits, Ye and colleagues⁸⁶ identified regions of the subunits involved in mediating this differential ethanol response. They identified a region encompassing part of the TM2 and TM3 regions and the extracellular loop between as being important for the effects of ethanol. However, there was some question about whether this site was part of a transduction mechanism brought about by ethanol rather than the ethanol-binding site itself. Subsequent mutagenesis studies of this region identified an amino acid in TM2 as being an important component of ethanol sensitivity. It was shown that replacement of the serine residue in position 267 of the glycine receptor α1 subunit by those occupying a smaller volume can increase the alcohol cut-off effect described above, supporting the idea that this region does constitute a binding site for alcohol.⁸⁶ Conversely, replacing the residue at position 267 with amino acids of a larger size decreased the cut-off size of alcohols. Further manipulation of this region resulted in receptors at which glycine-gated currents were inhibited, rather than potentiated, by ethanol.

GABA_A receptors in self-administration

There is strong evidence that GABAergic systems are involved in mediating self-administration of ethanol in animal models, likely by stimulating reward circuitry in the mesolimbic system.⁸⁷ For the most part, there is a correlation between compounds that activate or potentiate GABAergic systems and increased ethanol intake,^88,89 while the reverse is true for compounds that block or inhibit GABAergic systems.^90,91,92 Benzodiazepines, which display a spectrum of efficacies at GABA_A receptors, likewise can increase or decrease self-administration depending on the efficacy of the particular drug being studied.⁴³ However, some studies have not seen this effect with benzodiazepine agents.^93,94 It is believed that the mesolimbic dopamine system is intimately involved in this behaviour and that GABA_A receptors play a role in mediating dopamine levels in this region.⁴³ For example, infusion of muscimol directly into the ventral tegmental area (VTA), a component of the mesolimbic system, results in a dose-dependent increase in dopamine levels.⁹⁵ Also, long-term ethanol exposure decreases GABA_A receptor α1 subunits in the VTA,⁹⁶ which would in turn result in an increase in dopamine release.³⁴ An inverse agonist selective for α5-containing receptors when infused into rat hippocampus was effective in reducing ethanol self-administration.⁹⁷ α5-Containing receptors are enriched in the CA1 and CA3 fields, both of which innervate the nucleus accumbens and amygdala, components of the mesolimbic system.

Long-term effects of ethanol on GABA_A receptors

Long-term exposure to ethanol affects the baseline of neuronal excitability.^98,99 Animals chronically treated with ethanol and then withdrawn are typically more prone to seizure activity than naive controls. Several studies have shown that GABA_A receptor subunit mRNA and protein levels change during repeated ethanol exposure. At the functional level, examination of receptors in brain preparations derived from chronically treated rats shows that ethanol no longer potentiates GABA-gated chloride flux to as great an extent as is seen in non-treated rats. Additionally, the response to allosteric modulators such as benzodiazepine-site ligands is altered.¹⁰⁰ Positive allosteric modulators such as flunitrazepam appear to be less effective in potentiating the GABA response, whereas negative allosteric modulators such as Ro 15-4513 appear to have an enhanced effect.¹⁰¹ Additionally, the response to neurosteroids, a group of endogenous allosteric modulators, is altered.¹⁰²

Some of the more consistent findings in studies of repeated alcohol exposure and changes in GABA_A receptor subunit mRNA and protein involve changes in relative levels of α1 and α4 protein and mRNA. In animals exposed to ethanol, the mRNA and protein levels of α1 decrease, whereas those for α4 increase in certain brain regions such as cerebral cortex.^103,104 The reason for this change is not understood, but it appears to be a common adaptive change. We have observed this pattern of change in the brains of rats repeatedly exposed to other allosteric modulators such as benzodiazepines,¹⁰⁵ and others have observed increases in α4 expression in response to progesterone exposure.¹⁰⁶

Ethanol can pass freely through the lipid bilayer of cells and affects several intracellular proteins, including many involved in second-messenger pathways. In particular, many studies have indicated that protein kinase C (PKC) is affected by exposure to ethanol.¹⁹ Ethanol-induced alteration of PKC function leads to changes in the expression or function of voltage dependent calcium channels¹⁰⁷ and ligand-gated ion channels.¹⁰⁸ In brain tissue from mice lacking PKCε, GABA receptors are highly sensitive to allosteric modulation by ethanol, with 20 mmol/L ethanol bringing about twice the potentiation of GABA agonist-induced chloride ion flux compared with controls.¹⁰⁹ Additionally, these mice displayed less ethanol self-administration. Similarly, in a line of mutant mice lacking PKC-γ, ethanol lost its potentiating effect on GABA_A receptors.¹¹⁰ These mice show a distinct increase in the consumption of ethanol and appear to have a higher level of impulsive behaviour than wild-type mice. Interestingly, decreased sensitivity to the effects of ethanol seems to be predictive of a high probability of alcoholism in humans (see below).

Protein kinase A (PKA) activity is affected by ethanol and, in turn, appears to alter the function of other proteins.²⁰ In some rat brain regions, modulation of GABA_A receptors by ethanol will occur only if the region has been pre-exposed to β-adrenergic stimulating agents.¹¹¹ This shows that increased cyclic adenosine monophosphate (cAMP) levels, and therefore an increase in PKA activity, are needed for ethanol to potentiate GABA-gated currents. In support of this are findings that repeated ethanol administration results in increased PKA activity and cAMP levels and alters the subcellular location of PKA.^112,113,114 Knockout mice in which one of the regulatory subunits of PKA was disrupted were much less sensitive to the sedative effects of ethanol and consumed more ethanol than controls.¹¹⁵

GABA_A receptor subunit knockouts and ethanol

Several GABA_A receptor subunit knockout mice have been developed over the past several years. Because it was hypothesized that the long splice variant of the γ2 subunit was required for ethanol responses, this gene was disrupted to determine its role in ethanol sensitivity in mice. This line of mice showed no difference between wild-type mice in terms of the behavioural effects of ethanol.¹¹⁶ A line of mice in which the α6 subunit had been knocked out also showed no difference in response to ethanol when compared with wild-type animals.¹¹⁷ However, a different outcome was seen with δ subunit knockout mice; these animals consumed less ethanol, were less sensitive to withdrawal and displayed less ethanol-induced seizure protection than wild-type mice.¹¹⁸ Interestingly, δ subunits have been found to co-localize with α4 subunits in various brain regions,¹¹⁹ and, as mentioned previously, expression of this subunit increases in models of repeated ethanol exposure and appears to be a neuroadaptive response to the presence of ethanol.

GABA_A receptor as a target for therapeutics

A large body of evidence indicates that pharmacological manipulation of GABA_A receptors may be one way to treat alcohol-related diseases. Numerous behavioural studies have shown that allosteric modulators of GABA_A receptors, specifically those that recognize the benzodiazepine binding site, can reverse the effects of ethanol. One compound in particular, the imidazobenzodiazepine Ro 15-4513, has been reported to be particularly effective in reversing the effects of ethanol intoxication in rats.¹²⁰ This compound is a negative allosteric modulator. At the molecular level, Ro 15-4513 has been shown to reverse the potentiating effects of ethanol at recombinant receptors expressed in Xenopus oocytes. However, because of its anxiogenic and proconvulsant activity, Ro 15-4513 has not been tested as an alcohol antagonist in humans. Interestingly, this drug has also been shown to protect against ethanol-induced damage of the gastric mucosa by interacting with GABA_A receptors in the stomach.¹²¹ Ro 15-4513 binds to subtypes that contain any α, β and γ subunits. It binds with particularly high affinity to α5β2γ2 subtypes, with an affinity approximately 10-fold higher than for α1-containing receptors.

Other compounds that act at the benzodiazepine binding site, for example, members of the β-carboline structural family, also show potential for being alcohol antagonists. FG 7142, a β-carboline inverse agonist, has been shown to block the effects of ethanol in various GABA_A receptor-containing models. However, like Ro 15-4513, its proconvulsant activity precludes its use as a therapeutic agent. Allosteric modulators with no intrinsic activity at the benzodiazepine site, which therefore have the potential to be used therapeutically, show some promise in ameliorating the effects of ethanol. Although these compounds have not yet been shown to antagonize the potentiating effects of ethanol at GABA_A receptors, they do seem able to antagonize some of the complex behaviours associated with ethanol intake.^122,123 This has prompted some speculation that an endogenous compound produced during chronic ethanol ingestion is prevented from binding to GABA_A receptors by these compounds. Flumazenil (also known as Ro 15-1788), an imidazobenzodiazepine antagonist, has been the subject of several trials. There is conflicting evidence with respect to its ability to reduce ethanol tolerance and dependence,^{124,125,126,127} but it seems to be somewhat effective in lessening withdrawal-associated anxiety and convulsions.¹²⁸ Both flumazenil and ZK 93426 (a β-carboline antagonist) have been shown to reduce alcohol-induced aggression in rats and monkey.¹²⁹ Flumazenil is also reported to block ethanol-induced sleep when injected into the medial preoptic area of the hypothalamus of rats.¹³⁰ Other antagonists such as CGS 8216 have also been shown to modulate some of the behaviours associated with ethanol ingestion, and drugs that block the chloride channel (picrotoxin) or are GABA antagonists (bicuculline) have shown some potential in antagonizing the effects of ethanol.^131,132,133

Genetic component to alcoholism

It has been long-debated whether alcohol dependence is a due to environment or genotype. A genetic component for alcohol-related disorders is supported by many many studies.¹³⁴ Early genetic studies revealed polymorphisms in the gene coding for aldehyde dehydrogenase, an enzyme that plays a major role in the metabolism of ethanol. There are 2 common polymorphisms in these genes. One causes an increase in acetaldehyde production, the other a decrease in the rate of its removal. A build-up of acetaldehyde results, producing symptoms such as headache, vasodilation and nausea. At high enough concentrations, acetaldehyde can produce symptoms similar to those of disulfiram.^134,135 Poor alcohol metabolism due to these polymorphisms is prevalent in Chinese and Japanese populations. The side-effects of poor metabolism seem to, in turn, reduce the probability of alcoholism.^136,137

Several strains of rodents have been bred to be sensitive to ethanol, and there appears to be a GABAergic component underlying this trait. For example, long sleep (LS) and short sleep (SS) mice are strains that were selectively bred on the basis of the duration of sleep induced by acute exposure to ethanol. When the mRNA obtained from the brains of these mice was injected into Xenopus oocytes, the GABA_A receptors that were expressed were differentially modulated by ethanol.¹³⁸ mRNA from SS mice produced receptors at which ethanol antagonized GABA-gated currents, and mRNA from LS mice produced receptors at which ethanol potentiated GABA-gated currents. In a line of alcohol non-tolerant (ANT) rats, GABA_A receptors are highly sensitive to allosteric modulators of GABA_A receptors such as benzodiazepines. When GABA_A receptor subunits from this line were cloned and sequenced, it was discovered that the α6 subunits contained a mutation at position 100, where a glutamine residue replaced an arginine residue.¹³⁹ Interestingly, the amino acid occupying this position has been shown to be a major determinant of benzodiazepine affinity and efficacy,¹⁴⁰ but not of ethanol sensitivity, and so its significance in the ANT phenotype is uncertain.¹⁴¹ GABA_A receptor polymorphism appears to be the underlying cause in differences among mice strains with different reactions to alcohol withdrawal.

Quantitative trait loci (QTL) studies are based on statistical analysis of the association of a measurable trait and molecular markers that segregate with them. As the position of the markers is known, the relation between the trait(s) and regions of specific chromosomes can be determined. These studies are useful to determine the relation between complex traits that likely involve more than 1 gene and genotype. Application of this technique to understanding the genetic basis of alcohol-related behaviours has met with some success and has implicated the GABAergic system in these processes.¹⁴² The C57BL/6J and DBA/2J mouse lines have high and low acute ethanol withdrawal sensitivity, respectively. QTL mapping shows that a cluster of genes on chromosome 11 affects the severity of acute alcohol withdrawal.¹⁴³ Included in this cluster are the genes for the GABA_A receptor α1, α6, β2 and γ2 subunits. Comparison of the γ2 gene of the 2 mice strains shows that the γ2 subunit gene differs at position 11, with alanine present in the high withdrawal sensitivity mice and a threonine present in the low withdrawal sensitivity mice.¹⁴³ This allelic variation is highly correlated with susceptibility to behaviours such as withdrawal sensitivity and ethanol-induced motor incoordination.¹⁴⁴ How this polymorphism affects receptor function is unknown, but it has been proposed that it might disrupt an α-helical region of the N-terminus.

Several studies have also implicated a region of mouse chromosome 1 in mediating alcohol-related behaviours.¹⁴⁵ This region also contains genes involved in diazepam withdrawal convulsions and pentobarbital withdrawal. The identity of these genes has yet to be resolved. QTL mapping has also revealed a cluster of genes on mouse chromosome 2 that include those coding for glutamic acid decarboxylase, the enzyme that synthesizes GABA from glutamate, as being a determinant of ethanol withdrawal behaviours.¹⁴² Interestingly, these same regions overlap with those linked to pentobarbital withdrawal.¹⁴⁵ Other neurotransmitter systems, namely those for dopamine and serotonin, have also been shown in QTL studies to be linked to alcohol-related behaviours.^146,147

Polymorphisms in GABA_A receptor subunits have also been linked to alcohol response in humans. One study showed that the offspring of alcoholics tend to be less sensitive to the effects of ethanol than control groups, and this was a strong predictor for the development of alcoholism.¹⁴⁸ Genetic analysis of this group revealed a polymorphism in the GABA_A receptor α6 subunit, specifically a switch from proline at position 385 to serine.¹⁴⁹ Again, as was noted in the ANT mouse line, this mutation seems to be more important in benzodiazepine modulation and has not been shown to affect modulation of the receptor by ethanol. Results of another study in 2 psychiatric populations suggest that the serotonin1B gene might be linked to alcoholism in which aggression is displayed.¹⁵⁰

The role of GABA_A receptors in fetal alcohol syndrome

Fetal alcohol syndrome is a collection of abnormalities, including mental retardation, behavioural and developmental problems, as well as physical abnormalities (e.g., facial malformations).¹⁵¹ The developing CNS appears to be extremely sensitive to the effects of ethanol, especially during a period of rapid brain growth in the third trimester.¹⁵² There is evidence that GABA_A and NMDA receptors in the developing brain are involved in mediating the damage brought about by ethanol.^153,154 With respect to GABA receptors, it appears that the potentiating effects of ethanol initiate a pro-apoptotic cascade that leads to neuronal cell death.¹⁵⁵ The pattern of cell death brought about by ethanol matched that caused by exposure to benzodiazepines and barbiturates, both positive allosteric modulators of GABA_A receptors.¹⁵⁴ Relatively brief exposure of the developing brain to ethanol during periods of neuronal growth can lead to the deaths of millions of neurons in animal models. This neuronal death is likely the cause of the reduced brain mass and thinner cerebral cortex typically noted in those with fetal alcohol syndrome. The mechanism by which the apoptotic cascade begins through GABA_A receptor activation is unclear, but a link between these receptors and calcium influx via L-type voltage-gated calcium channels has been shown in a model system composed of proliferating neural precursor cells.¹⁵⁶

Conclusions

The effects of ethanol on proteins in the CNS are complex and involve many different systems. It appears that ligand-gated ion channels and voltage-gated calcium channels are important targets for this drug because their function, type and numbers are altered by short- and long-term exposure to ethanol. A large body of work ranging from biochemical to genetic studies points to the importance of GABA_A receptors in mediating the effects of ethanol in the CNS. It is possible that drugs targeting these receptors could be a component of therapies designed to battle alcohol abuse and dependence.