The GABA_B receptor as a target for antidepressant drug action

Abstract

Preclinical and clinical data suggest that a modification in GABA_B receptor expression and function may contribute to the symptoms of major depression and the response to antidepressants. This includes laboratory animal experiments demonstrating that antidepressants modify brain GABA_B receptor expression and function and that GABA_B receptor antagonists display antidepressant potential in animal models of this condition. Clinical and post-mortem studies reveal changes in GABAergic transmission associated with depression as well as depression-related changes in GABA_B subunit expression that are localized to the cortical depression network. Detailed in this review are the preclinical and clinical data implicating a role for the GABA_B receptor system in mediating symptoms of this disorder and its possible involvement in the response to antidepressants. Particular emphasis is placed on clinical and post-mortem studies, including previously unpublished work demonstrating regionally-selective modifications in GABA_B receptor subunit expression in brain samples obtained from depressed subjects. Together with the earlier preclinical studies, these new data point to a role for the GABA_B system in major depression and support the antidepressant potential of GABA_B receptor antagonists.

Introduction

It has been over five decades since the discovery of antidepressants (Kuhn, 1958). During that time, neuronal pathways and neurochemical systems that appear critical for mediating the symptoms of this disorder have been identified (Rajkowska et al., 1999; Nestler et al., 2001; Krystal et al., 2002; Mayberg, 2003; Seminowicz et al., 2004). These discoveries were due, in part, to the insights gained on the mechanisms of action of these drugs. Even with this progress, however, little has changed with regard to the types of agents employed to treat this condition, with this class still dominated by drugs that directly interact with monoamine systems (Kelsey and Nemeroff, 1998). This is not due to a lack of effort, as there remains a need for novel treatments (Enna and Williams, 2009). While newer antidepressants are generally safer than older agents, little progress has been made in decreasing the percentage of non-responders and in speeding the onset of action. Inasmuch as the first antidepressants were discovered empirically in the clinic, and there is no consistently identifiable neuropathology associated with this condition, the difficulties associated with discovering more efficacious antidepressants suggests that fundamental information is still lacking concerning the underlying neurobiological abnormalities responsible for this disorder. Indeed, it is possible that major depression is not a single entity unto itself, but rather a constellation of symptoms that only manifest in association with other psychiatric conditions. This could explain why no single mechanistic approach, such as enhancement of monoaminergic transmission, would be effective in all, or even the majority, of depressed patients because of the variable nature of the underlying cause.

Numerous attempts have been made to identify and develop antidepressants that target neurotransmitter systems other than those directly associated with the monoamines. Included have been antagonists for neurokinin-1 receptors (Herpfer and Lieb, 2005; Alvaro and Di Fabio, 2007), melanin-concentrating hormone-1 receptors (Shimazaki et al., 2006), corticotrophin-releasing factor-1 receptors (Valdez, 2009) and metabotropic glutamate receptors (Lesage and Steckler, 2010). While there are substantial preclinical data supporting the antidepressant potential of these agents, clinical studies have not as yet demonstrated their superiority over conventional therapies. This lack of success is likely due to several factors. One is that the animal models typically employed for screening antidepressant candidates were developed on the basis of their response to the clinically effective monoaminergic drugs. Although many of these models appear to have face validity, construct validity remains questionable given the lack of understanding of the disease process, as does the predictive validity for candidates that do not directly activate monoaminergic transmission. Moreover, as the response to many clinically effective psychotherapeutics involves interactions with multiple sites, it is possible that the most effective antidepressant may be one that targets several transmitter systems. As designing this type of drug is difficult given the number of possible target combinations, such an agent is more likely to be discovered with a pharmacometric approach (Enna and Williams, 2009).

The chances of successfully developing novel antidepressants are likely to be enhanced if there is direct evidence from preclinical, clinical or post-mortem studies directly linking the intended target with major depression in humans. The GABA_B receptor is a candidate that fulfills this criterion (Table 1). Thus, since its discovery in the early 1980s (Bowery, 2010), animal studies have indicated that the GABA_B receptor system is modified by chronic administration of antidepressants (Enna and Bowery, 2004). These findings, in turn, led to clinical studies aimed at identifying the consequences of stimulating this receptor system in depressed patients. Moreover, post-mortem brain studies demonstrate alterations in GABAergic neurons and possible changes in brain GABA synthesis in depressed subjects (Cryan and Slattery, 2010). Most recently, post-mortem analyses has indicated changes in GABA_B receptor subunit gene expression in discrete areas of the brain within the proposed cortical depression network (Klempan et al., 2009; Sequeira et al., 2009).

Table 1.

Chronological listing of selected reports on the relationship between GABA_B receptors and major depression

Authors	Date	Type of report	Journal	Findings
Pilc and Lloyd	1984	Preclinical	Life Sci	Chronic antidepressant administration increases GABAB receptors binding in the rat brain
Lloyd et al.	1985	Preclinical	J Pharmacol Exp Ther	GABAB receptor number in brain is increased by electroshock or antidepressant administration in the rat
Suzdak and Gianutsos	1986	Preclinical	Eur J Pharmacol	Repeated antidepressant or GABAB agonist administration modifies GABAB receptor binding and function in rat brain
Gray and Green	1987	Preclinical	Br J Pharmacol	Antidepressants or electroconvulsive shock increases GABAB receptor function in mouse brain
Cross et al.	1988	Clinical	Psychiatry Res	GABAB receptor binding profiles similar between suicide and control
Martin et al.	1989	Preclinical	Neuropsychobiology	Antidepressant drugs reverse decrements in GABAB receptor expression in a rat model of major depression
Marchesi et al.	1991	Clinical	Psychoneuroendocrinology	GABAB regulation of growth hormone in major depression
Post et al.	1991	Clinical	Int Clin Psychopharmacol	Baclofen exacerbates symptoms of major depression
Arranz et al.	1992	Clinical	Neuropsychobiology	GABAB receptor binding profiles in suicide and control subjects
O'Flynn and Dinan	1993	Clinical	Am J Psychiatry	GABAB regulation of growth hormone release in major depression
Pratt and Bowery	1993	Preclinical	Br J Pharmacol	Repeated administration of GABAB antagonist or desipramine up-regulates GABAB receptor binding in rat brain
Petty	1995	Review	J Affect Disord	A GABAergic hypothesis of depression
Nakagawa et al.	1996	Preclinical	Brain Res	Baclofen displays antidepressant activity in a rat model of depression
Davis et al.	1997	Clinical	Psychoneuroendocrinology	Lack of growth hormone response to baclofen in patients with major depression
Nakagawa et al.	1999	Preclinical	Eur J Pharmacol	GABAB antagonist reduces helplessness in a rat model of major depression
Sanacora et al.	2000	Clinical	Crit Rev Neurobiol	Neuroimaging and GABAergic function in major depression
Krystal et al.	2002	Clinical	Mol Psychiatry	Magnetic resonance spectroscopy indicates reduced cortical GABA levels in depression
Sands et al.	2003b	Preclinical	Life Sci	GABAB receptor subunit gene expression in rat brain is differentially modified in models of major depression and schizophrenia
Froestl et al.	2004	Preclinical	Biochem Pharmacol	GABAB antagonist SGS742 displays antidepressant activity in animal models of depression
Mombereau et al.	2004	Preclinical	Neuropsychopharmacology	GABAB1 subunit deletion mutant mice displays antidepressant phenotype in the forced-swim test
Sands et al.	2004	Preclinical	Biochem Pharmacol	Repeated administration of antidepressants modifies GABAB receptor function in rat hippocampus
Enna and Bowery	2004	Review	Biochem Pharmacol	Review of physiological and pharmacological manipulations that alter GABAB receptor expression and function
Fatemi et al.	2005	Clinical	Schizophr Res	GAD expression is altered in mood disorders and schizophrenia
Mombereau et al.	2005	Preclinical	Neuroreport	GABAB2 receptor subunit deletion mutant mice displays antidepressant phenotype in forced-swim test
Slattery et al.	2005	Preclinical	J Pharmacol Exp Ther	GABAB receptor antagonists display antidepressant activity in rodent models of major depression
McCarson et al.	2005	Preclinical	Biochem Pharmacol	Effect of antidepressants on GABAB receptor expression in rat spinal cord is state-dependent
Frieling and Bleich	2006	Review	Eur Arch Psychiatry Clin Neurosci	Tranylcypromine and GABAB receptor function
McCarson et al.	2006	Preclinical	Brain Res	Antidepressant administration or repeated stress alters GABAB receptor expression and function in rat spinal cord
Nowak et al.	2006	Preclinical	Br J Pharmacol	GABAB receptor antagonists display antidepressant activity in rodent models of major depression
Vigot et al.	2006	Preclinical	Neuron	GABAB receptor subunits regulate receptor location and function in mouse hippocampal neurons
Bielau et al.	2007	Clinical	Ann NY Acad Sci	GAD staining reveals altered GABAergic neuron terminal density in post-mortem brain samples from depressed patients
Rajkowska et al.	2007	Clinical	Neuropsychopharmacology	Reduced GABAergic neuron number in prefrontal cortex in major depression
Cornelisse et al.	2007	Preclinical	J Neurophysiol	Selective serotonin reuptake inhibitor treatment reduces GABAB receptor function in rat brain
Frankowska et al.	2007	Preclinical	Pharmacol Rep	Baclofen displays antidepressant activity in animal models of major depression
Klempan et al.	2009	Clinical	Mol Psychiatry	GABAA and GABAB receptor expressions are altered in the prefrontal cortex of suicides
Maciag et al.	2009	Clinical	Biol Psychiatry	Reductions in cortical GABAergic neurons in major depression
Sequeira et al.	2009	Clinical	PLoS One	Gene expression linkage analysis implicates GABAB mechanisms in major depression
Levinson et al.	2010	Clinical	Biol Psychiatry	GABAB-mediated cortical silence is prolonged in major depression
Cryan and Slattery	2010	Review	Advances in Pharmacology	Overview of research implicating a role for GABAB receptors in major depression

Outlined in this review are the preclinical and clinical data implicating a role for the GABA_B receptor system in mediating symptoms of major depression, and its possible involvement in the response to antidepressants. Particular emphasis is placed on clinical and post-mortem studies, with a detailed description of previously unpublished work demonstrating regionally selective modifications in GABA_B receptor subunit expression in brain samples obtained from depressed subjects. Together with the earlier clinical and preclinical work (Table 1), these findings point to a role for the GABA_B system in major depression and support the antidepressant potential for GABA_B receptor antagonists.

GABA_B receptors

Two pharmacologically and molecularly distinct GABA receptors have been identified, GABA_Aand GABA_B (Enna, 2007). Whereas the GABA_Asite is a pentameric, ligand-gated ion channel allosterically modulated by benzodiazepines and other anxiolytic and hypnotic agents, the GABA_B receptor, a G protein-coupled heterodimer, is the site of action for baclofen, a muscle relaxant (Bowery, 2010). A class III metabotropic site, the GABA_B receptor is composed of two 7-transmembrane spanning proteins (Kubo and Tateyama, 2005; Binet et al., 2006). While there are multiple subunit isoforms in various animal species, GABA_B1a, GABA_B1b and GABA_B2 predominate, with coupling of either of the GABA_B1 subtypes with GABA_B2 essential for insertion into the plasma membrane and receptor function (Kaupmann et al., 1998; Chronwall et al., 2001; Enna and Bowery, 2010). Although GABA_B receptors are widely distributed throughout the neuroaxis, subunit expression differs among brain regions (Bischoff et al., 1999; Towers et al., 2001; Vigot et al., 2006; Farb et al., 2007), suggesting the possible existence of regionally and molecularly distinct GABA_B receptor subtypes.

The activation of GABA_B receptors causes neuronal hyperpolarization by decreasing Ca²⁺ and increasing K⁺ membrane efflux, the latter through direct coupling to Kir3 K⁺ channels (Enna, 2001; Ladera et al., 2008; Fernandez-Alacid et al., 2009; Pinard et al., 2010). Located both pre- and post-synaptically, GABA_B receptors influence cAMP production through coupling to G_i and G_o. The stimulation of this site can either inhibit or enhance formation of this second messenger depending upon whether there is a simultaneous activation of a G_s-coupled site in the same neuronal compartment (Karbon and Enna, 1985). Thus, the response to GABA_B receptor stimulation or inhibition may vary as a function of the pre-existing state of the affected cell. Given their widespread distribution, and myriad of effects on second messenger production and ion channel activity, it is not surprising that laboratory animal and human studies indicate that alterations in the GABA_B receptor system contribute to the symptoms of a host of clinical conditions, including seizures, cognitive deficits, depression, anxiety, spasticity, drug abuse, schizophrenia, pain and gastro-oesophageal reflux disease (Enna, 2001; Enna and Bowery, 2004; Froestl, 2010).

Because of these findings, efforts have been expended to develop orthosteric GABA_B receptor agonists and antagonists, and allosteric modulators (Froestl, 2010). Included are agonists such as arbaclofen placarbil (Gerson et al., 2010; Lal et al., 2009), a baclofen prodrug, and lesogabaran (Bredenoord, 2009), and orthosteric antagonists such as CGP36742, CGP54626 and SCH50911 (Froestl, 2010). Positive allosteric modulators, including CGP7930 and GS39783, have been designed to more selectively activate subsets of the GABA_B receptor system and thereby minimize the side effects encountered with orthosteric agonists (Cryan et al., 2004; Mannoury la Cour et al., 2008).

While evidence suggests the possibility of pharmacologically distinct GABA_B receptors (Bonanno and Raiteri, 1993a,b; Cunningham and Enna, 1996), their existence has been a matter of dispute (Waldmeier et al., 1994). Indeed, the discovery that heterodimerization is required for receptor function, and the fact that the number of GABA_B receptor subunits is limited, argue against variability among orthosteric binding sites, which are exclusively located on the GABA_B1subunit. The identification of pharmacologically distinguishable sites is important therapeutically as generalized activation or inhibition of GABA_B receptors would be anticipated to be accompanied by numerous side effects, as is the case with baclofen, an orthosteric GABA_B receptor agonist. It is encouraging therefore that in recent years the evidence supporting pharmacologically distinct GABA_Breceptors has grown (Pinard et al., 2010). Data are accumulating to suggest that the differential relative affinities and responses reported for GABA_B receptor agonists and antagonists could be due to differences in the expression or function of regulators of G-protein signaling (RGS) proteins, which can influence GABA_B receptor and K⁺ channel responsiveness (Mutneja et al., 2005). Moreover, four sequence-related cytosolic proteins have been discovered that bind as tetramers to the C-terminal domain of the GABA_B2 subunit, influencing the pharmacology and kinetics of the receptor response (Pinard et al., 2010; Schwenk et al., 2010).

Thus, the functional responsiveness of GABA_B receptors is dependent upon the production of GABA_B1 and GABA_B2 subunits, the coupling of the latter to G proteins, the phosphorylation state of the receptor, and the scaffolding provided by RGS and cytosolic proteins. Taken together, these observations, along with earlier studies, provide evidence for possible pharmacological heterogeneity among GABA_B receptors. Such findings are crucial for customizing compounds to selectively interact with subsets of GABA_B receptors in developing new therapies for the treatment of CNS disorders, including major depression.

GABA_B receptors and depression

It has been speculated for some time that GABA_B receptors are modified in depression and in response to antidepressant therapies (Enna and Bowery, 2004) (Table 1). While there have been conflicting findings (Cryan and Slattery, 2010), the preclinical studies generally indicate that chronic, but not acute, administration of antidepressants or electroconvulsive shock increases the number and function of GABA_B receptors in rodent brain (Pilc and Lloyd, 1984; Lloyd et al., 1985; Suzdak and Gianutsos, 1986; Gray and Green, 1987; Martin et al., 1989; Pratt and Bowery, 1993; Frieling and Bleich, 2006; Cornelisse et al., 2007), and that brain GABA_B receptor number is decreased in rat models of depression (Martin et al., 1989). Notably, antidepressant-induced increases in GABA_B receptor binding occur in only certain rat brain regions (Pratt and Bowery, 1993).

Although these manipulations generally increase GABA_Breceptor sensitivity, changes in the magnitude and direction of GABA_B subunit expression vary with the central nervous system area examined. Thus, it appears that regionally selective changes in the production, assembly and processing of brain GABA_B subunits may be important homeostatic mechanisms for controlling CNS activity. For example, antidepressant-induced increases in GABA_B receptor binding occur in only certain areas of the rat brain frontal cortex, including laminas I and VI. In contrast, chronic administration of these agents causes no changes in GABA_B binding in laminas II, III and V (Pratt and Bowery, 1993). This regional specificity may explain why some have been unable to detect such changes and suggests this receptor modification is not due to a direct interaction of the drug with GABA_B receptors, but rather is secondary to an effect on some other, most likely monoaminergic, system (Slattery et al., 2005).

With the cloning of the GABA_B receptor subunit genes (Kaupmann et al., 1997; 1998; Jones et al., 1998; White et al., 1998; Chronwall et al., 2001), it became possible to examine whether the changes noted in GABA_B receptor binding and function are due to modifications in the transcription or translation of these proteins (McCarson and Enna, 1999; Sands et al., 2003a,b; 2004; McCarson et al., 2005; 2006;). Studies indicate that chronic administration of antidepressants, stress or pain differentially modifies GABA_B receptor subunit gene expression and receptor responsiveness in rat spinal cord and hippocampus, and alters GABA_B receptor responsiveness in these subjects. The chronic administration of classical antidepressants generally increases GABA_B receptor function and GABA_B1a gene expression in rat hippocampus and dorsal horn of the spinal cord, while having a variable effect on the expression of the GABA_B2 subunit gene. These data suggest that antidepressants cause an up-regulation in GABA_B receptor expression and function by decreasing GABAergic tone, supporting the notion that depression is characterized by an overabundance of brain GABAergic activity (Sands et al., 2004).

The antidepressant-induced increase in GABA_B receptor function is evidenced by an enhancement in baclofen-stimulated cAMP production in brain and spinal cord tissue obtained from animals that are chronically administered any one of a number of such agents (Sands et al., 2003a; 2004; McCarson et al., 2006) Likewise, repeated administration of amitryptyline or electroconvulsive shock enhances baclofen-induced inhibition of K⁺-stimulated serotonin release from mouse frontal cortex (Gray and Green, 1987). Both of these findings are consistent with other data indicating these drugs increase GABA_B receptor number in this brain area.

An enhancement of receptor number and responsiveness could suggest that antidepressants either correct a depression-related underactive system or decrease GABAergic tone, leading to a supersentive receptor state. While there has been a report suggesting that baclofen displays antidepressant activity in some animal models of depression (Frankowska et al., 2007), the overwhelming weight of evidence suggests that a decrease in GABA_B receptor activity is more typically associated with an antidepressant response. Thus, GABA_B receptor antagonists display antidepressant properties in most, but not all (Mombereau et al., 2004), animal models of this condition (Nakagawa et al., 1999; Froestl et al., 2004; Mombereau et al., 2005; Nowak et al., 2006), and GABA_Breceptor stimulation exacerbates learned helplessness in rats (Nakagawa et al., 1996), a behaviour interpreted as a model for human depression. Also, mice lacking functional GABA_B receptors behave as though they were receiving an antidepressant, suggesting regionally selective enhancements in brain GABAergic function in depression (Mombereau et al., 2004). Interpretation of these findings is limited, however, by the possibility that this behavioural response is due to secondary adaptive changes in these genetically modified animals rather than being a faithful representation of a phenotype that results solely from a selective decline in GABA_B receptor activity.

It has been reported that GABA_B receptor positive allosteric modulators display weak anxiolytic activity in some (Cryan et al., 2004; Jacobson and Cryan, 2008), but not all (Jacobson and Cryan, 2008; Paterson and Hanania, 2010), animal models of anxiety. Moreover, anxiety-like behaviour has been noted in GABA_B receptor-deficient mice (Mombereau et al., 2005).These findings suggest that GABA_B receptor blockade might exacerbate or precipitate an anxiety disorder in susceptible subjects, such as individuals with major depression. However, the clinical importance of this finding remains questionable given the inconsistency of the anxiolytic response to GABA_B positive allosteric modulators in animal models, and the lack of any reported anxiogenic effect of an orthosteric GABA_B receptor antagonist following administration to humans (Froestl et al., 2004).

Other preclinical data supporting the hypothesis that GABA_B receptor blockade may alleviate depression include the finding that GABA_B receptor antagonists increase gene expression and protein levels of nerve growth factor and brain-derived neurotophic factor (BDNF) in various regions of the rat brain (Heese et al., 2000; Enna et al., 2006). The relevance of this discovery to depression is based on reports that various classes of antidepressants, as well as electroconvulsive shock, increase the expression of BDNF in the rat hippocampus, and that BDNF displays antidepressant-like activity when placed directly into this brain region (Duman and Monteggia, 2006). In addition, hippocampal levels of BDNF are decreased in a mouse model of depression (Tsankova et al., 2006). It has been proposed that this effect of antidepressants on BDNF expression induces hippocampal neurogenesis, which is thought to be an important factor in alleviating depression (Miller et al., 2008). However, as noted by Tanti and Belzung (2010), BDNF polymorphisms are associated with a number of conditions, the effect of antidepressants on hippocampal neurogenesis is species-dependent and this response is not observed with all agents that display antidepressant activity in other models. Thus, the predictive value of enhanced BDNF production in assessing antidepressant potential remains uncertain.

Nonetheless, the results from nearly 30 years of research suggest that antidepressants cause an up-regulation of GABA_B receptor number and function secondary to a decrease in GABAergic activity that may result from prolonged activation of monoaminergic systems (Sands et al., 2004). Serotonergic transmission appears to be particularly important in this regard in that GABA_B receptor antagonists no longer display antidepressant properties following administration of para-chlorophenalanine, an inhibitor of tryptophan hydroxylase (Slattery et al., 2005). It has also been reported that antidepressants decrease the function of presynaptic serotonin-3 receptors on GABA neurons, resulting in a decrease in GABA release (Nakagawa and Ishima, 2003). Such findings have led to speculation that depression is characterized by an enhanced GABA_B tone, perhaps as a result of a decrease in serotonergic activity, and that the response to antidepressants is dependent upon a reduction in GABA_B receptor stimulation which, in turn, leads to a supersensitive GABA_B receptor system.

Human studies on the GABA_B system and neuropsychiatric disorders

A number of clinical and post-mortem studies support a causal relationship between the GABAergic system and depression (Cryan and Slattery, 2010) (Table 1). Thus, GABA levels in the cerebral cortex, plasma and CSF are lower than normal in depressed patients, as is the number of GABA neurons in layer II of the orbitofrontal cortex (Petty, 1995; Rajkowska et al., 1999; Sanacora et al., 2000; Krystal et al., 2002). A GABA_B receptor involvement is suggested by the findings of some (Marchesi et al., 1991; O'Flynn and Dinan, 1993; Lucey et al., 1994), but not others (Davis et al., 1997), that the growth hormone response to baclofen is blunted in depressed individuals as compared with controls, suggesting altered GABA_B receptor responsiveness in these patients. Also, an efficacy study with baclofen indicates that this GABA_B receptor agonist worsens symptoms of depression (Post et al., 1991). Although the sample size is too small for drawing firm conclusions from this study, these data are interesting in light of the subsequent preclinical work suggesting that depression may be associated with an overstimulation of the GABA_B system.

Post-mortem studies reveal that those diagnosed with an affective disorder display a decreased expression of cerebellar glutamic acid decarboxylases (GAD) (Fatemi et al., 2005), the enzymes responsible for the synthesis of GABA, and differences in GAD immunohistochemistry in various regions of the cerebral cortex and the hippocampus as compared with controls (Bielau et al., 2007). There have also been reports of differences between depressed and control subjects in the size and density of cerebral cortical GABA neurons (Rajkowska et al., 2007; Maciag et al., 2009). While such studies are important for establishing a GABAergic dysfunction in depression, they do not directly address whether, and to what extent, these changes influence, or are related to, the GABA_B receptor system.

This issue was addressed directly by a study showing that the cortical silent period, a measure of cortical inhibition thought to be a reflection of GABA_B receptor function, is prolonged in depressed individuals (Levinson et al., 2010). As baclofen administration lengthens the cortical silent period in normal subjects, the finding with depressed patients supports the preclinical work suggesting that this disorder is characterized by an enhancement in GABA_B receptor activity. In contrast, however, earlier binding studies on post-mortem tissue (Cross et al., 1988; Arranz et al., 1992) found no differences in GABA_B receptor number or affinity between controls and suicide subjects in frontal and temporal cortices and hippocampal samples. However, the interpretation of these results is compromised by the fact that these studies were conducted using relatively gross brain regions, which may dilute any changes that occur in highly discreet brain areas, and that ligand binding alone reveals nothing about the functional state of the receptor.

Efforts have been made to determine whether modifications in the expression of GABA_B receptor subunits are associated with neuropsychiatric illness. Inasmuch as both GABA_B1 and GABA_B2 must be present to form a functional receptor, a change in the production of either could signal an alteration in the responsiveness of this system. As detailed previously, laboratory animal studies indicate that antidepressant administration, as well as chronic pain and stress, alters GABA_B subunit expression and receptor function in rat brain and spinal cord (Sands et al., 2003a; 2004; McCarson et al., 2006), demonstrating the utility of analysing subunit expression as an indicator of receptor modifications. As both GABA_B receptor subunits are found throughout the human brain (Billinton et al., 2000; Berthele et al., 2001; Waldvogel et al., 2004), it is likely that alterations in their expression could result in CNS disturbances, the nature of which would depend on the specific brain region involved. Indeed, reports indicate regionally selective changes in GABA_B subunit expression in association with schizophrenia (Mizukami et al., 2002), temporal lobe epilepsy (Furtinger et al., 2003; Princivalle et al., 2003) and autism (Fatemi et al., 2009). Microarray studies of post-mortem brain tissue obtained from depressed and non-depressed suicide victims indicate modifications in the expression of genes responsible for the production of both GABA_A and GABA_B receptors in various regions of the prefrontal cortex and in selected subcortical areas (Klempan et al., 2009; Sequeira et al., 2009). The GABA_B2 subunit expression appeared to be particularly affected, being elevated in the depressed suicide group relative to non-depressed individuals (Sequeira et al., 2009). A full appreciation of the significance of these findings awaits replication of this work and a determination as to whether these alterations in gene expression are indicative of a change in subunit protein. Besides hinting at an association between changes in the production of GABA_B receptor subunits and depression, these studies confirm the importance of examining this issue in well-defined regions, as the gene alterations are not global, but rather circumscribed to rather discreet brain areas.

Human GABA_B receptor subunit expression in depression

To confirm that depression is associated with selective, regionally defined changes in human brain GABA_B receptor subunits, a preliminary study was undertaken with post-mortem brain samples obtained from depressed individuals and control subjects. The primary hypothesis was that regionally selective alterations in brain GABA_B receptor subunit gene expression are a characteristic of this disorder. The brain areas examined were the hippocampus, subgenual cingulate and orbitofrontal cortex, regions implicated in the pathophysiology of depression (Sheline, 1996; Sheline et al., 1999; Mayberg, 2003; Seminowicz et al., 2004; Pittenger and Duman, 2008; Hajszan et al., 2009; Koolschijn et al., 2009; Yucel et al., 2009; Price and Drevets, 2010). Particular emphasis was placed on subsections of the hippocampus as this is a region where the balance between excitatory and inhibitory inputs appears to be particularly critical. Examining the molecular contribution of GABA_B alterations within the extended cortical depression system complements a focus on the reward regions in depression. The entire extended network includes the subgenual cingulate, the dorsolateral prefrontal cortex, orbitofrontal cortex, pregenual anterior cingulate and frontal pole, as well as the amygdala, hippocampus and insula. The regions of greatest importance to major depression are prefrontal cortex, anterior cingulate, amygdala, anterior thalamus and regions within the hippocampus. Inasmuch as these brain areas are richly innervated by monoaminergic and GABAergic neurons, they are most likely to display changes in GABA_B receptor subunit expression if, as indicated by the preclinical studies, there is a functional interplay between these transmitter systems in depression and the response to antidepressants.

Because major depression is a neuroanatomically complex condition, it is critical to examine in discrete brain regions the possible molecular and neurochemical changes associated with this condition to reduce the possibility of overlooking a meaningful alteration because of a diluting effect of adjacent tissue. Changes in the volume, function and interaction among cortical-limbic brain areas are particularly evident in depression (Sheline, 1996; Sheline et al., 1999; Mayberg, 2003; Seminowicz et al., 2004; Pittenger and Duman, 2008). Within the hippocampus, alterations in the morphology of the dentate gyrus (DG), CA1 and CA3 regions occur in association with stress, depression and antidepressant therapy (Malberg et al., 2000; Pittenger and Duman, 2008; Hajszan et al., 2009). Likewise, depression-related volume changes occur in the subgenual cingulate cortex (Yucel et al., 2009) and the orbital frontal cortex (Koolschijn et al., 2009). Thus, these brain regions, along with cerebellar cortex, an area not believe to contribute to the symptoms of this disorder, were selected for studying possible changes in GABA_B subunit expression in depression.

Human brain tissue from depression and control cases was obtained from the Dallas Brain Collection (Stan et al., 2006). The tissue was collected only after acquiring consent from the next of kin along with permission to review medical records and to conduct a telephone interview with a primary caregiver. All clinical information on each case was evaluated by at least three research psychiatrists and diagnoses were made using DSM-IV criteria. Blood screens for drugs of abuse, alcohol and prescription medications were conducted on each subject. Cases were excluded when there was a known history of neurological disorders or of an axis I psychiatric condition other than major depression.

The human post-mortem material was obtained from 12 cases diagnosed with major depressive disorder and 12 control subjects. The two groups were matched as closely as possible for age, brain pH, post-mortem interval (PMI), and RNA integrity number (RIN), an indicator of human post-mortem tissue RNA quality (Stan et al., 2006) (Table 2). While half the members of the depression group were suicides, only two subjects in this total cohort had detectible blood levels of medication at the time of death (Table 2). The groups were analysed in a paired design. The hippocampus, anterior cingulate and orbitofrontal cortices were selected for study because of the in vivo imaging data suggesting their involvement in the clinical manifestations of depression and the response to antidepressant treatments. The tissue samples were dissected from the anterior cingulate (BA24) and orbitofrontal cortex (BA11), as well as hippocampal subfields and cerebellum. Other than for the hippocampus, the samples were frozen immediately in a mixture of dry ice and isopentane (1:1, v : v), pulverized on dry ice and stored at −80°C until analysed.

Table 2.

Characteristics of depressed and control group subjects

Controls							Depression
Case	Age (years)	PMI (hours)	RIN	Gender	Race	Cause of death	Case	Age (years)	PMI (hours)	RIN	Gender	Race	Meds	Cause of death
C1	48	20	6.8	M	C	HT CVD	D1	46	11	8.3	F	C	yes	HT CVD
C2	60	11	9.3	M	AA	HT CVD	D2	54	6	8.9	M	C	no	AS CVD
C3	20	21	8.2	M	C	Blunt force injury	D3	35	9	7.4	M	C	no	suicide
C4	31	16	8.1	M	C	HT CVD	D4	33	18	8.5	M	C	no	suicide
C5	60	20	8.5	M	C	Myocardial infarction	D5	61	20	8.6	M	C	no	suicide
C6	43	15	6.1	M	C	HT CVD	D6	50	23	8.8	M	C	no	suicide
C7	65	11	8	F	C	Aortic dissection	D7	57	16	6.6	F	C	no	suicide
C8	83	13	8.5	F	C	Sharp force injury	D8	59	10	9	F	C	no	HT CVD
C9	34	23	6.2	M	C	HT CVD	D9	40	18	6.9	M	C	no	suicide
C10	48	15	9.5	M	C	HT CVD	D10	41	13	8	F	AA	yes	Diabetic ketoacidosis
C11	63	14	7.7	M	C	AS CVD	D11	65	14	7.6	M	C	no	HT CVD
C12	19	20	9	M	C	Gunshot wound to chest	D12	26	19	8.1	F	C	no	Accidental drowning
av	47.83	16.58	7.99	2F/10M	11C/1AA		av	47.25	14.75	8.06	5F/7M	11C/1AA
SD	19.45	4.08	1.12				SD	12.39	5.10	0.79

PMI, post-mortem interval; RIN, RNA integrity number; meds, antidepressant medications at time of death; HT CVD, hypertensive cardiovascular disease; AS CVD, atherosclerotic cardiovascular disease; F, female; M, male; C, Caucasian; AA, African American; av, average; SD, standard deviation.

The entire hippocampus was removed from the fresh brain, embedded longitudinally into a mold with Histomer polymer (Histotech, Frederiksberg, Denmark). Tissue blocks were then taken at 5 mm intervals and frozen immediately in a mixture of dry ice and isopentane (1:1, v : v). Blocks from the mid-level of the hippocampus were used for the study. Four samples, each 300 µm thick, were cryostat sectioned at −20°C, then stored at −80°C. Nissl staining of 14 µm sections adjacent to the samples was used to determine orientation.

In each of the 300 µm sections, the parahippocampal gyrus was first dissected away from the hippocampus proper, then a series of cuts was made to isolate the CA3, CA1, subiculum and DG (Figure 1).

Figure 1.

Representation of the coordinates used to dissect the dentate gyrus (DG), CA1 and CA3 regions of the hippocampus.

The GABA_B receptor subunit expression assays were performed blind to diagnosis (depressed or control) using paired samples from the two groups of subjects. Sample pairing was performed by someone not involved with the biochemical assay to ensure that tissues tested on any given day included an equal number of samples from the same brain regions of depressed and control individuals.

Modified versions of human GABA_B1a and GABA_B2 expression vectors provided by Dr Klemens Kaupmann (Novartis, Basel, Switzerland) were used for probe synthesis. The probe sequences were bases 276-613 of human GABA_B1a cDNA (Accession Number AJ225028) and bases 2746-3188 of human GABA_B2cDNA (Accession Number BC035071.2). Human β-actin mRNA expression was quantified with probes generated using a pGEM-T vector containing bases 374-1093 (Accession Number NM_001101) of the human β-actin sequence (Nandan and Reiner, 1997).

Total RNA was isolated from tissue samples using a rapid-guanidinium method (Chomczynski and Sacchi, 1987) and assayed separately for GABA_B1a, GABA_B2 and β-actin mRNAs using solution hybridization–nuclease protection assays (McCarson and Krause, 1994).

The primary analyses were designed to test whether regionally selective alterations occur in GABA_B receptor subunit expression in the hippocampus, anterior cingulate and orbitofrontal cortex in cases of depression compared with controls. Secondary analyses were conducted to explore the influence of age and gender on receptor subunit expression. An independent samples t-test was employed for comparing these results given the relatively large sample size, the continuity and normal distribution of the gene expression values, the convergence of measures of central tendency, and the similar variance in each data set. Spearman rank correlations were used to assess possible correlations between mRNA levels with RIN and PMI. Unpaired t-tests were conducted to verify that the two diagnostic groups were matched on demographic variables, age, RIN and PMI. In all analyses, differences were considered statistically significant with P≤ 0.05.

Both GABA_B1aand GABA_B2 receptor subunit gene expression were detectable in all brain regions examined (Table 3). Of the hippocampal subfields studied, only the DG displayed a significant difference between depressed and control groups. The expression of both subunit genes in the DG differed significantly between the depressed subjects and controls, with a 30% decrease in GABA_B1a(t= 2.18, df₁₉, P= 0.04) and a 50% increase in GABA_B2gene expression (t = 2.21, df₂₀, P= 0.04) in this brain region (Table 3). The latter result confirms an earlier finding from a microarray study indicating that GABA_B2 subunit expression is elevated in brain tissue from depressed suicide subjects as compared with non-depressed individuals (Sequeira et al., 2009). Differential modifications in the expression of GABA_B1 and GABA_B2 subunit gene expression have been found in rat brain (Sands et al., 2003b) and spinal cord (McCarson et al., 2006) tissue. It is unknown whether these changes are occurring in the same or different cellular elements. Regardless, an increase or decrease in gene expression of either one or both subunits in a single cell is likely to cause, and reflect, an alteration in GABA_B receptor function. The results also indicated a significant negative correlation between GABA_B1asubunit gene expression and age (R =−0.43, P= 0.04) in the CA3 region of the hippocampus. Covarying for age did not alter the absence of a group difference in GABA_B1a expression in CA3. Comparisons between groups revealed a significant decrease in GABA_B1a subunit expression in the CA3 subfield in depressed male subjects compared with control males (t = 2.55, df₁₁, P= 0.03; Figure 2).

Table 3.

GABA_B1a and GABA_B2 subunit gene expressions in various brain regions of depressed and control subjects

	GABAB1a		GABAB2
Brain region	Depressed	Control	Depressed	Control
CA1	17 ± 4	23 ± 3	Not measured
CA3	86 ± 18	83 ± 8	Not measured
Dentate gyrus	25 ± 2*	37 ± 5	17 ± 2*	11 ± 1
Subgenual cingulate cortex	14 ± 2	16 ± 2	13 ± 2	17 ± 3
Orbital frontal cortex	9 ± 2	7 ± 1	64 ± 13	36 ± 6
Cerebellum	4 ± 1	4 ± 1	19 ± 6	23 ± 5

Values are the mean pg subunit specific mRNA/ng β-actin ± SEM.

^*P < 0.05 compared with corresponding control, two-tailed Student's t-test.

Figure 2.

GABA_B1a subunit gene expression in the CA3 region of the hippocampus of depressed (n= 6) and control (n= 7) male (blue) and depressed (n= 4) and control (n= 2) female (red) subjects. Horizontal lines indicate the means for each group. The level of significance for the difference between means for the male subjects is P= 0.03, as determined by an independent samples t-test.

No significant differences were noted between depressed and control groups in GABA_B1aor GABA_B2subunit gene expression in the orbital frontal and anterior cingulate cortices or cerebellum (Table 3). The GABA_B2 receptor subunit expression in the orbital frontal cortex was, however, nearly twice as high in male depressed subjects as in male controls (t = 2.3, P= 0.04; Figure 3), and there was a strong trend towards a significant increase in the expression of the GABA_B2 subunit (P < 0.06) in this brain region of depressed subjects compared with controls (Cohen's d effect size = 0.83) (Table 3). While there were no significant correlations between age, RIN or PMI and GABA_B2 receptor subunit gene expression for these three brain regions, and for GABA_B1aexpression in the orbital frontal cortex, a correlation was noted between GABA_B1aexpression and PMI (R = 0.41, P= 0.05) and RIN (R = 0.46, P= 0.02) for the anterior cingulate cortex and cerebellum respectively.

Figure 3.

GABA_B2 subunit gene expression in the orbital frontal cortex of depressed (n= 7) and control (n= 9) male (blue) and depressed (n= 5) and control (n= 2) female (red) subjects. Horizontal lines indicate the means for each group. The level of significance for the difference between means for the male subjects is P= 0.04, as determined by an independent samples t-test.

Because major depression has been associated with changes in the volume of some brain regions (Rajkowska, 2000; Rajkowska et al., 2007; Maciag et al., 2009), it is possible that measurement of beta-actin gene expression may not be an appropriate reference for control as the quantity of this marker could differ between the two groups as a result of cell loss. However, analysis of the beta-actin gene revealed no significant differences in expression in the brain areas examined in the depressed and control groups (data not shown). This finding indicates that despite any loss of volume or decrease in cell number, the expression of this marker relative to total RNA levels remains unchanged in depression. Accordingly, beta-actin appears to be an appropriate control gene for normalizing the levels of GABA_B receptor subunit gene expression under these circumstances.

These results suggest a decrease in GABA_B1a and an increase in GABA_B2 subunit expression in the DG of depressed individuals as compared with controls. Because dysfunction in this brain region has been previously linked with depression, this change in GABA_B receptor subunit expression could be associated with the illness itself. Because of the size of the tissue samples, it was not possible to analyse GABA_B2subunit expression in the CA1 and CA3 regions of the hippocampus, leaving open the possibility of a depression-related modification in the expression of this GABA_B receptor subunit in these areas. Nonetheless, the fact that no significant changes in GABA_B1a subunit gene expression were noted in the CA1 or CA3 regions of the hippocampus, or in GABA_B1a and GABA_B2expression in the subgenual cingulate cortex, two areas also thought to be involved in the depression brain circuit, nor in the cerebellum, an area outside this circuit (Pittenger and Duman, 2008), suggests that the changes observed in the DG are selective and not a reflection of a generalized abnormality in GABA_Bsubunit gene expression as a result of the disorder, drug treatment or death.

The finding of an apparent decrease in GABA_B1a subunit expression in the CA3 region of male subjects, and a doubling of the GABA₂subunit expression in the orbital frontal cortex of these individuals as compared with male controls, suggests that the GABA_B system may be modified in these depression circuit brain areas as well (Figures 2 and 3). Inasmuch as the GABA_B1 subunit gene expression data for the CA3 and orbital frontal cortex nearly attained statistical significance when comparing all samples (Table 3), it seems probable these areas are affected, with the lack of statistical significance for the present findings possibly being due to the influence of variation because of the sample size. Given the small number of female subjects, it is impossible to determine whether these changes are related to gender (Figures 2 and 3). Nevertheless, the modifications noted in the DG alone demonstrate that the GABA_Breceptor system is altered in a critical brain region associated with major depressive illness.

It is also noteworthy that a negative correlation was found between GABA_B1a subunit gene expression in the CA3 region and age, while no significant correlations were detected between GABA_B receptor subunit gene expression and drug history or post-mortem delay. It could be speculated that this age-related decline in GABA_B1a expression in the CA3 may contribute to the increase in susceptibility to depression in the elderly.

From these data, it is impossible to know whether the changes in GABA_B receptor subunit expression lead to a change in the production of subunit protein or receptor function. However, numerous studies have indicated that alterations in GABA_B subunit expression are usually accompanied by a change in receptor sensitivity (Sands et al., 2003a; 2004; McCarson et al., 2005; 2006; Merlo et al., 2007). As the production of GABA_B2 subunits appears to be the rate-limiting step in the formation of functional GABA_B receptors (Thuault et al., 2004), these results, like previous work in laboratory animals, suggest that the system is up-regulated in depressed subjects in the DG, and possibly the CA3 region of the hippocampus and in the orbital frontal cortex. As a change in GABA_B receptor activity alters the expression of brain-derived and glial cell line-derived neurotrophic factors (Heese et al., 2000; Enna et al., 2006; Fiorentino et al., 2009), it is possible that alterations in GABA_B receptor expression and function influences hippocampal neurogenesis, which may be a component of the clinical response to antidepressants (Pittenger and Duman, 2008).

As these data were obtained from a predominantly antidepressant-free cohort of cases (Table 2), the results suggest that the changes in GABA_B receptor subunit expression may be part of the molecular phenotype of this psychiatric condition. It is, however, possible that they are long-lasting responses to prior antidepressant treatment. This is a critical issue as it has been reported that the GABA_B receptor response to antidepressants is state-dependent (McCarson et al., 2005), making it impossible to predict the effect of these drugs on human brain GABA_B receptors without knowing whether this system is modified by the condition itself. As discussed previously, an increase in GABA_B function could be a response to a persistent antidepressant-induced reduction in GABAergic tone, or it might be an evidence of a disease-related supersensitive GABA_B receptor system. The latter possibility is consistent with the discovery that GABA_B receptor antagonists display an antidepressant profile in animal models of this condition (Nakagawa et al., 1999; Froestl et al., 2004; Slattery et al., 2005; Nowak et al., 2006).

The findings that baclofen, a GABA_B receptor agonist, worsens the symptoms of depression (Post et al., 1991) and, like depression, prolongs the cortical silent period in humans (Levinson et al., 2010), and that mice lacking functional GABA_B receptors behave as though they have been administered an antidepressant, all support the notion that an overactive GABA system contributes to the symptoms of this disorder. They also argue strongly against the idea that the antidepressant response to GABA_B receptor antagonists might be due to an enhancement in GABA release secondary to the blockade of GABA_B autoreceptors.

These data confirm and extend previous studies (Klempan et al., 2009; Sequeira et al., 2009) indicating a direct relationship between a modification in GABA_Breceptor subunit gene expression and major depression, with receptor subunit changes being most evident in the DG. Given the proposed relationship between the DG and affective illness (Malberg et al., 2000), these findings suggest a direct link between modifications in human brain GABA_B receptor subunit gene expression and depression, and provide insights into the molecular mechanisms that may be responsible, at least in part, for some of the neurochemical and behavioural changes associated with this condition. These discoveries support the notion that GABAergic medications, in particular GABA_B receptor antagonists, may represent a novel approach for the treatment of this disorder.

While it has been some time since the development of orally active GABA_B receptor antagonists (Froestl et al., 1995), only one of these phosphinic acid GABA analogues has been examined clinically (Froestl et al., 2004). In this study, SGS742 progressed through Phase II clinical trials as a potential treatment for cognitive deficits. Although no serious side effects were noted at the doses tested, and some benefits were reported for patients diagnosed with mild cognitive impairment, clinical trials were halted because the efficacy was insufficient to warrant commercial development. Given the difficulties associated with demonstrating clinical antidepressant activity, and the low affinity of SG742 for the GABA_B receptor site, no effort has yet been made to test the hypothesis that GABA_B receptor antagonists are antidepressants. Proof of principle must await the development of more potent, orally active and pharmacokinetically appropriate members of this class.

To fully exploit these preliminary findings on receptor subunit expression in post mortem brain-tissue, future work should focus on determining the functional correlate of these changes and on whether these alterations are drug induced or part of the pathophysiological process. Ultimately, the relationship between GABA_B receptors and depression can only be conclusively tested by a thorough clinical assessment of the antidepressant properties of GABA_B receptor antagonists.

Glossary

Abbreviations

BDNF

brain-derived neurotophic factor

GABA

γ-aminobutyric acid

GABA_B1a

GABA_B1b, GABA_B2, γ-aminobutyric acid_B receptor subunits

PMI

postmortem interval

RGS

regulators of G-protein signaling

RIN

RNA integrity number

Conflict of interest

The authors declare no competing financial interests in relation to the work described in this report.