Molecular Psychiatry (2003) 8, 721–737 & 2003 Nature Publishing Group All rights reserved 1359-4184/03 $25.00 www.nature.com/mp FEATURE ARTICLE GABAergic dysfunction in mood disorders P Brambilla1, J Perez1, F Barale2, G Schettini3 and JC Soares4,5 1Biological Psychiatry Unit, IRCCS S Giovanni di Dio, Fatebenefratelli, Brescia, Italy; 2Department of Psychiatry, IRCCS S Matteo, University of Pavia, Italy; 3Advanced Biotechnology Center, University of Genova, Italy; 4Department of Psychiatry, The University of Texas Health Sciences Center, San Antonio, TX, USA; 5South Texas VA Health Care System, Audie Murphy Division, San Antonio, TX, USA The authors review the available literature on the preclinical and clinical studies involving GABAergic neurotransmission in mood disorders. c-Aminobutyric acid (GABA) is an inhibitory neurotransmitter present almost exclusively in the central nervous system (CNS), distributed across almost all brain regions, and expressed in interneurons modulating local circuits. The role of GABAergic dysfunction in mood disorders was first proposed 20 years ago. Preclinical studies have suggested that GABA levels may be decreased in animal models of depression, and clinical studies reported low plasma and CSF GABA levels in mood disorder patients. Also, antidepressants, mood stabilizers, electroconvulsive therapy, and GABA agonists have been shown to reverse the depression-like behavior in animal models and to be effective in unipolar and bipolar patients by increasing brain GABAergic activity. The hypothesis of reduced GABAergic activity in mood disorders may complement the monoaminergic and serotonergic theories, proposing that the balance between multiple neurotransmitter systems may be altered in these disorders. However, low GABAergic cortical function may probably be a feature of a subset of mood disorder patients, representing a genetic susceptibility. In this paper, we discuss the status of GABAergic hypothesis of mood disorders and suggest possible directions for future preclinical and clinical research in this area. Molecular Psychiatry (2003) 8, 721–737. doi:10.1038/sj.mp.4001362 Keywords: GABA; bipolar disorder; unipolar disorder; mood disorders; antidepressants; mood stabilizers g-Aminobutyric acid (GABA) is the most abundant In the present paper, we reviewed the physiology of inhibitory neurotransmitter in the mammalian brain, GABAergic transmission in human brain and sum- where it is widely distributed.1 In regions such as the marized the findings from preclinical and clinical cerebral cortex, hippocampus, thalamus, basal gang- studies evaluating GABAergic function in mood lia, cerebellum, hypothalamus, and brainstem, it disorders. We attempted to elucidate available find- represents about one-third of the synapses.2–4 GABA ings for GABAergic dysfunction in an integrated transmission is present in interneurons modulating hypothesis of mood disorder and also discussed local neuronal circuitry, including noradrenergic, potential directions for future research in this area. dopaminergic, and serotonergic neurons. The potential role of GABAergic dysfunction in mood disorders was first proposed by Emrich et al,5 GABAergic pathways in the brain based on the efficacy of valproate in the treatment of bipolar patients. They proposed that valproate, GABA metabolism and uptake through the enhancement of GABA brain concentra- GABA in GABAergic terminals is formed from tion, might compensate for a potential GABAergic glutamate in an enzymatic reaction mediated by deficiency, and formulated the GABA hypothesis of glutamic acid decarboxylase (GAD), using pyridoxal 7,8 mood disorders. After Emrich’s hypothesis, several phosphate as cofactor. After being released into the animal and human studies have evaluated the synapses, GABA is inactivated by reuptake into potential role of GABAergic abnormalities in the presynaptic terminals or into glia cells mediated by 9 pathophysiology of mood disorders.6 GABA transporters (GATs). Specifically, at the pre- sent time, four complementary DNAs (cDNAs) en- coding highly homologous GATs proteins have been cloned (GAT-1, GAT-2, GAT-3, and BGT-1). GAT-1 is considered to be a neuronal transporter, GAT-2 and Correspondence: P Brambilla MD, Biological Psychiatry Unit, GAT-3 are believed to be glial transporters, whereas IRCCS S Giovanni di Dio, Fatebenefratelli, via Pilastroni 4, 25125 Brescia, Italy. E-mail: [email protected] the role of BGT-1 in brain GABA uptake is un- known.10 Precisely, GAT-1 is the most copiously Received 9 January 2003; revised 11 April 2003; accepted 16 April expressed GAT in the CNS and is mainly localized 2003 into presynaptic axon terminal and into few GABA abnormalities in mood disorders P Brambilla et al 722 astrocytic processes. GAT-2 is primarily present in the leptomeninges and in ependymal and choroid plexus cells and, to a minor extent, in neuronal and non- neuronal elements. GAT-3 is localized exclusively to distal astrocytic processes, although a neuronal localization has been reported in some brain regions such as the retina.10 GATs are regulated by several factors including GABA itself, brain-derived neuro- trophic factor (BDNF), and hormones. The different response of GATs to the composition of extracellular environment, the different regulation of their activity and/or expression, and the possibility of reversing the direction of GABA transport, confer to the GABA transport system considerable flexibility for the fine regulation of GABA levels under physiological and pathological conditions.11 GABA that is taken up by astrocytes is not immediately available for synaptic transmission, because it is metabolized to succinic semialdehyde (SSA) by GABA-transaminase (GABA-T), which uses pyridoxal phosphate. Then, succinic semialdehyde is oxidized either by succinic semialdehyde dehydro- Figure 1 GABA metabolism and uptake in human brain. Glutamate is the precursor of free GABA in GABAergic genase (SSA-DH) to succinic acid (SA), which re- terminals and comes from two different sources (Kreb’s enters the Kreb’s cycle and then is transformed into cycle in glia cells and glutamine in nerve terminals). Then glutamate, or by aldehyde reductase to g-hydroxybu- the enzyme glutamic acid decarboxylase (GAD) forms tyrate. Glutamate in astrocytes cannot be converted GABA from glutamate. After being released into the into GABA due to the absence of GAD and is synapses, GABA is inactivated by reuptake mediated by transformed by glutamine synthetase into glutamine, GABA transporters (GATs) into presynaptic terminals or which is then transferred to axon terminals by into glia cells where it is metabolized by GABA transami- specific transporters. In nerve terminals, glutamine nase (GABA-T). is then converted into glutamate by the enzyme glutaminase, and, finally, GAD forms GABA from 7,8 glutamate closing the cycle (Figure 1). On the ylation sites, and a short C terminus. GABAA receptor contrary, GABA that is taken up by neuronal usually contains a, b, and g subunits with variable transporters is readily available for further release, combinations, which may be relevant to pharmacolo- because it either undergoes the same transformation gical differences observed between drugs and may as in astrocytes (with the notable difference that nerve modulate receptor activity. In mammalian brain, endings contain GAD and can resynthesize GABA) or a1b2g2 is the major GABAA receptor subunit. During is recycled directly into synaptic vesicles. GAD is neurotransmission, GABA acts postsynaptically localized only in GABAergic presynaptic terminals, through allosteric interaction with GABAA receptors lacks in glial cells, and two forms have been and allows the chloride (ClÀ) ion channel opening, 12 À15,16 discovered so far (GAD65 and GAD67). Glutamine increasing the conductance of Cl . Once GABAA synthetase is present only in glia, whereas GABA-T receptors are activated, hyperpolarization of the and SSA-DH are found in neuronal and glial neuronal membrane is established, reducing the cell mitochondria.4 excitability and leading to the inhibitory actions of GABA. However, in the presence of chronic GABA GABA receptors administration, ClÀ currents gradually decrease, as GABAergic receptors are composed by two main per a concentration-dependent GABAA response. types with different distribution on the surface of GABAA receptors have several binding sites for neurons, GABAA and GABAB receptors. different ligands, such as muscimol (GABA agonist), GABAA receptors are ionotropic and mostly post- bicuculline (GABA antagonist), benzodiazepines synaptic receptors mainly located at the apical (BZDs), barbiturates, ethanol, anticonvulsants, neuro- dendrite of the neurons, causing the fast inhibitory steroids, steroid anesthetics, and volatile general postsynaptic potential (IPSP).13 They are hetero- anesthetics.17–19 These are allosteric agents, leading oligomeric membrane proteins organized in a chan- to increased GABA affinity and increased frequency nel, composed of five subunits belonging to several of chloride channel opening. Specifically, BZDs bind different classes with multiple variants (a1–a6; b1–b4, to subunit a and increase the affinity of the receptor g1–g3, d, e, y, p, and r1–r3).14 Each subunit has a large for GABA.20,21 In addition, it has been shown that N terminus, four hydrophobic transmembrane do- phosphorylation and dephosphorylation processes mains, an intracellular loop containing protein kinase might regulate GABAA receptor function. For in- A, protein