Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

There is a growing body of evidence implicating oxidative stress mechanisms and the impaired redox regulation in the pathophysiology of diverse psychiatric disorders (Do et al. 2009; Steckert et al. 2010; Bitanihirwe and Woo 2011; Yao and Keshavan 2011). Oxidative stress defined in accordance with the free radical hypothesis refers to the cytopathological consequences of an imbalance between free radical production on the one side and deficiency of the antioxidant defense system on the other side. Brain cells are particularly vulnerable to oxidative damage due to relatively low to moderate activity of antioxidant enzymes (superoxide dismutase, SOD; catalase, CAT; glutathione peroxidase, GPx) when compared to the liver or kidney, high levels of lipids and polyunsaturated fatty acids, high metal content, and high oxygen utilization (Dringen 2000; Valko et al. 2007). Hence, the free radical-mediated damage of important cellular molecules, like lipids, proteins, and DNA, leading to the impairment of cell function and vitality is currently considered as one of the main mechanisms in the pathophysiology of both neurodegenerative and psychiatric disorders (Bains and Shaw 1997; Valko et al. 2007; Ng et al. 2008; Steckert et al. 2010; Bitanihirwe and Woo 2011).

On the other hand, as the clinical trials with supplementation of free radical scavenging antioxidants show little benefit in humans, a complementary hypothesis for oxidative stress has been postulated (Jones 2006, 2008). This hypothesis, which is termed the “redox hypothesis” to facilitate its distinction from the “free radical hypothesis,” assumes that disruption of the redox states of thiol systems which normally function in cell signaling and physiological regulations is the most central feature of oxidative stress. Three main thiol systems represented by thiol/disulfide redox couples exist in biological systems: (i) reduced glutathione (GSH) and its disulfide (GSSG), GSH/GSSG couple; (ii) cysteine (Cys) and its disulfide, cystine (CySS), Cys/CySS couple; and (iii) reduced and oxidized thioredoxins (Trx), Trxred/Trxox couple (Kemp et al. 2008). All these systems are responsible for the maintenance of the appropriate redox potential in different cellular compartments (Jones 2006, 2008). The “redox hypothesis” draws attention to non-radical mechanisms of oxidative stress. It has been demonstrated that biological systems generate significantly more non-radical oxidants than free radicals (Jones 2008). Hydrogen peroxide (H2O2), the most abundant non-radical oxidant, is formed in many enzymatic reactions. For instance, H2O2 is the predominant product in the reaction catalyzed by xanthine oxidase although the superoxide radical anion (O2 −•) is also produced. Moreover, O2 −• is converted into oxidant H2O2 in the reaction catalyzed by the antioxidant enzyme SOD. The free radicals NO• and O2 −• react to generate peroxynitrite (ONOO−) that is a powerful oxidizing and nitrating non-radical intermediate (Thomas et al. 2006). Apart from the abovementioned oxidants, biological systems produce other oxidants, such as hydroperoxyfatty acids, aldehydes, quinones, epoxides, and disulfides. All these oxidants, albeit not free radicals, contribute importantly to the regulation of cellular redox state by modulation of the thiol-containing proteins that play important roles in cell-to-cell signaling, macromolecular trafficking, and physiological regulation (Jones 2008). Redox elements represented by sulfhydryl (−SH) residues of cysteine and thioether groups of methionine, found in the active site of many proteins, are susceptible to two-electron oxidants. Hence, the two-electron oxidation can be considered to be a component of oxidative stress that is distinct from free radical-mediated macromolecular damage. The function of these thiol-containing proteins is controlled by thiol antioxidants, GSH, Cys, and Trxred, which are able to prevent the two-electron oxidation. Therefore, the appropriate levels of these antioxidants are of great importance for the maintenance and regulation of the thiol redox status of the cells.

The scope of this chapter is to review the available literature referring to GSH synthesis and its multiple functions in the central nervous system. Particular attention will be focused on GSH deficit and dysregulation of redox state in psychiatric disorders such as schizophrenia and bipolar disorder. Oxidative stress-mediated alterations in psychiatric patients will be compared to those observed in animal models of GSH deficiency. Finally, a treatment restoring the redox balance will be discussed in the context of therapy of psychiatric disorders.

2 Functions of Glutathione

Glutathione (γ-glutamyl-L-cysteinylglycine, GSH), a cysteine-containing tripeptide and the most abundant nonprotein thiol, is present in the mammalian brain with an average concentration of 2–3 mM, but there are marked differences in its content between particular cell types (Dringen 2000). Neurons have much less GSH than glial cells. Among glial cells, microglia contains the highest amount of GSH, while oligodendrocytes, which are affected in schizophrenia, the lowest level. Astrocytes, like microglial cells, are characterized by a relatively high level of GSH (Dringen 2000). Glutathione exists predominantly in the thiol, i.e., reduced (GSH) form (99 %), while the disulfide, i.e., oxidized form (GSSG), represents less than 1 % of the total glutathione pool under physiological conditions.

2.1 The Main Functions of Glutathione

In the mammalian cells, GSH plays a lot of diverse functions including (1) scavenging of free radicals; (2) detoxification of xenobiotics; (3) maintenance of the redox state of proteins; (4) providing a nontoxic storage form of cysteine; (5) modulation of critical cellular processes, such as DNA synthesis and repair, cell proliferation, and redox signaling; (6) regulation of nitric oxide homeostasis; and (7) modulation of the activity of glutamate receptors in the central nervous system (Jánaky et al. 1998, 1999; Oja et al. 2000). GSH serves as an endogenous NO reservoir to form S-nitrosoglutathione (GSNO) (Hogg 2002). In the brain, GSNO can also elicit neuroprotective effect under oxidative stress conditions (Rauhala et al. 1998).

Thus, given many critical processes that are affected by GSH, it is not surprising that disturbances in its homeostasis have been implicated in the etiology and/or progression of a number of human diseases, including neurodegenerative and neuropsychiatric diseases (Ballatori et al. 2009; Do et al. 2009).

2.2 Antioxidant Activity of Glutathione

As an antioxidant, GSH scavenges reactive oxygen species (ROS) generated during electron transport and cellular metabolism of endo- and exogenous compounds. GSH is also involved in the disposal of hydrogen peroxide (H2O2) and hydroperoxides, which are non-radical oxidants produced during different cellular processes, strongly affecting redox state of cells. The detoxification of ROS and peroxides is associated with two types of reaction. Firstly, GSH reacts directly and nonenzymatically with such radicals as superoxide radical anion (O2 −•), nitric oxide (NO•), or hydroxyl radical (HO•) (Hogg 2002; Winterbourn and Metodiewa 1994). The removal of HO• via this route is one of the most important functions of GSH in the nervous system:

$$ 2\;\mathrm{G}\mathrm{S}\mathrm{H}+2\;{\mathrm{H}\mathrm{O}}^{\bullet}\to \mathrm{GSSG}+2\;{\mathrm{H}}_2\mathrm{O} $$

Secondly, GSH is an electron donor in the reduction of peroxides, mainly H2O2 in the reaction catalyzed by glutathione peroxidase (GPx):

$$ {\mathrm{H}}_2{\mathrm{O}}_2+2\;\mathrm{G}\mathrm{S}\mathrm{H}\overset{\mathrm{GPx}}{\to}\mathrm{G}\mathrm{S}\mathrm{S}\mathrm{G}+2{\mathrm{H}}_2\mathrm{O} $$

Final products of the GPx-mediated reduction of H2O2 comprise water and glutathione disulfide (GSSG). Also, catalase can reduce H2O2, but the brain has relatively low level of this enzyme as compared with that of GPx (Dringen and Hamprecht 1997). GSSG is then reduced back to GSH by glutathione reductase (GR):

$$ \mathrm{GSSG}+\mathrm{NADPH}+{\mathrm{H}}^{+}\;\overset{\mathrm{GR}}{\to}\;2\;\mathrm{G}\mathrm{S}\mathrm{H}+{\mathrm{NADP}}^{+} $$

This enzyme transfers electrons from nicotinamide adenine dinucleotide phosphate (NADPH) to GSSG, thereby regenerating GSH. Hence, the detoxification of peroxides depends on the availability and regeneration of NADPH. Organic peroxides can be reduced by GPx and glutathione S-transferase (GST). During the reactions catalyzed by GPx and GR, glutathione is not consumed but recycled. The relative ratio of the reduced to oxidized glutathione (GSH/GSSG) serves as an indicator of the cellular redox environment. The maintenance of high GSH/GSSG ratio requires energy expenditure. The adult brain relies almost exclusively on the glucose oxidation to meet its energy requirements. The pentose phosphate pathway is present in the brain, especially in astrocytes, but only 3–5 % of glucose is converted via this pathway, while the rest is oxidized via the tricarboxylic acid cycle. Nevertheless, the pentose phosphate pathway is important in the brain as a means of providing NADPH for the GSSG reduction to GSH (Dringen 2000). Consistently, it has been demonstrated that the pentose phosphate pathway was strongly activated in cultured astrocytes during the detoxification of H2O2 (Ben-Yoseph et al. 1996). On the other hand, glucose deprivation of astrocyte cultures significantly reduced astrocyte ability to remove H2O2 (Dringen and Hamprecht 1997) and prolonged regeneration time of GSH from GSSG after treatment with hydroxyperoxides (Dringen et al 1998).

The third important role of GSH is associated with the maintenance of intracellular redox homeostasis. Protein S-glutathionylation, the reversible formation of mixed disulfides between glutathione and low-pKa cysteinyl residues of proteins, is thought to be a regulatory and antioxidant mechanism (Dalle-Donne et al. 2007; Mieyal et al. 2008). The binding of glutathione molecules to proteins to form mixed disulfides protects protein –SH groups against irreversible oxidation to –SO2H and –SO3H. Hence, protein S-glutathionylation is an important mechanism for a dynamic, posttranslational modification of a variety of regulatory, structural, and metabolic proteins as well as for the regulation of signaling routes and metabolic pathways (Dalle-Donne et al. 2007; Mieyal et al. 2008). This modulation of proteins is not only a cellular response to mild oxidative/nitrosative stress but also occurs under physiological conditions.

2.3 Detoxifying Function of Glutathione

GSH reacts with various endogenous compounds and xenobiotics in the reaction catalyzed by glutathione S-transferase (GST) to form glutathione S-conjugates which are exported to the outside of the cell (Commandeur et al. 1995; Salinas and Wong 1999). There is only one enzyme, γ-glutamyl transpeptidase (γ-GT), localized on the outer side of the plasma membrane of certain cell types that is able to hydrolyze GSH conjugates to cysteinyl-glycine conjugates. The cysteinyl-glycine bond is then cleaved by dipeptidase, resulting in a cysteinyl conjugate that following N-acetylation is further metabolized to mercapturic acid.

Lipid peroxidation induced by ROS leads to the conversion of polyunsaturated fatty acids to highly reactive aldehydes, such as 4-hydroxynoneal (4-HNE), that inactivate proteins required for cell viability (Esterbauer et al. 1991). Therefore, rapid and efficient removal of these compounds is necessary to maintain cell function. GSH can react with 4-HNE via the action of GST to form GSH-HNE conjugates (Xie et al. 1998) which are then exported from the cells via transport-mediated efflux (Berhane et al. 1994). This process plays an important role in cellular detoxification. Furthermore, in the brain, GST detoxifies quinones that are formed during the oxidation of dopamine and other catecholamines (Baez et al. 1997; Dagnino-Subiabre et al. 2000). The latter reactions irreversibly consumes intracellular GSH, and without supplementation of GSH stores, the formation of glutathione S-conjugates can rapidly compromise cellular antioxidant capacity, finally leading to the enhanced production of ROS and disruption of the cellular redox balance.

2.4 Specific Function of Extracellular Glutathione in the Brain

The presence of GSH in the extracellular space has been confirmed using microdialysis method (Yang et al. 1994; Lada and Kennedy 1997). Experiments performed on brain slices demonstrated that GSH was released by depolarization induced by high K+ concentration in a Ca2+ dependent manner, which indicates its origin from a neuronal compartment (Zängerle et al. 1992). The mechanism of this release is unknown; however, the fact that it is Ca2+ dependent suggests that GSH is released by a vesicular mechanism similar to that of classical neurotransmitters or its efflux is under the control of a released neurotransmitter. In the rat brain slices, the most prominent release of GSH was observed in the mesodiencephalon, cortex, hippocampus, and striatum and lowest in the pons-medulla and cerebellum (Zängerle et al. 1992). On the other hand, studies performed in cell culture have shown that astrocytes are mainly involved in GSH release and together with γ-GT affect its extracellular level (Sagara et al. 1996; Dringen et al. 1997). Consistently, it has been calculated that cultured astrocytes release about 10 % of the intracellular pool of GSH within one hour (Dringen et al. 1997). So, the astrocyte-mediated release of GSH is the process consuming the largest amount of this peptide.

Although GSH plays a crucial role in many cellular processes, its extracellular functions are less known. However, in the nervous system, besides these generally known functions, GSH may serve additionally as a modulator of ionotropic glutamate receptors or as a new neurotransmitter (Jánaky et al. 1998, 1999, 2000; Ogita et al. 1998; Pasqualotto et al. 1998). So, it has been demonstrated that GSH binds via γ-glutamyl moiety to ionotropic glutamate receptors, preferentially AMPA and NMDA (Janáky et al. 1999). At micromolar concentrations, GSH displaces excitatory agonists from their binding sites, acting to halt their physiological actions on target neurons (Janáky et al. 1999; Oja et al. 2000). Since AMPA and NMDA receptors are colocalized and cooperate at postsynaptic membranes, the co-release of glutamate and GSH from nerve endings (Hjelle et al. 1998) may have profound consequences in synaptic transmission. According to the model of GSH synaptic actions proposed by Janáky et al. (1999), this peptide may inhibit the fast depolarization evoked by glutamate via AMPA receptors and thus inhibit the voltage-dependent opening of NMDA receptor ionophores. Hence, the co-release of glutamate and GSH would lead to a cascade of events enabling the receptors to be reactivated within the short time (Janáky et al. 1999; Oja et al. 2000). Moreover, GSH (at millimolar concentrations) acting via its cysteinyl thiol group can modulate the redox site of NMDA receptors (Janáky et al. 1999; Oja et al. 2000). As such modulation has been shown to increase NMDA receptor ion channel currents, this action may play a significant role in normal and abnormal synaptic activity. Finally, it has been demonstrated that GSH at nanomolar to micromolar range binds to at least two populations of binding sites that are distinct from any known glutamate receptor subtypes. It is believed that these binding sites represent a unique population of GSH receptors. GSH binds to these receptors via cysteinyl moiety and is not displaceable by glutamatergic agonists or antagonists (Shaw et al. 1996; Janáky et al. 1999, 2000). The application of GSH to cortical slices elicits a fast depolarizing potential that is markedly larger than that produced by NMDA and AMPA (Shaw et al. 1996). The GSH current appears to be linked to sodium ionophores as it was blocked by the absence of sodium ions but not by lowering of calcium or NMDA or AMPA antagonists (Shaw et al. 1996; Janáky et al. 1999, 2000). These reports suggest that GSH receptors may be a key component of cortical excitatory neurotransmission (Shaw et al. 1996).

3 Synthesis of GSH

3.1 Biosynthesis of GSH in the Brain

GSH is synthesized from its constituent amino acids, i.e., glutamate, cysteine, and glycine, in the cytosol of all mammalian cells by the consecutive action of two enzymes requiring adenosine triphosphate (ATP) as a co-substrate (Dringen 2000; Lu 2009):

$$ \mathrm{L}-\mathrm{glutamate}+\mathrm{L}-\mathrm{cysteine}+\mathrm{A}\mathrm{T}\mathrm{P}\to \gamma -\mathrm{glutamyl}-\mathrm{L}-\mathrm{cysteine}+\mathrm{A}\mathrm{D}\mathrm{P}+\mathrm{Pi} $$
(1)
$$ \gamma -\mathrm{glutamyl}+\mathrm{L}-\mathrm{cysteine}+\mathrm{L}-\mathrm{glycine}+\mathrm{A}\mathrm{T}\mathrm{P}\to \mathrm{G}\mathrm{S}\mathrm{H}+\mathrm{A}\mathrm{D}\mathrm{P}+\mathrm{Pi} $$
(2)

In the first step of GSH biosynthesis, glutamate cysteine ligase (GCL, EC 6.3.2.2; formerly γ-glutamylcysteine synthetase) catalyzes the formation of the dipeptide γ-glutamyl-L-cysteine (γ-GluCys) from glutamate and cysteine (Reaction 1). This reaction exhibits absolute requirement not only for ATP but also for Mg2+ or Mn2+ (Franklin et al. 2009). The second step of GSH biosynthesis is catalyzed by GSH synthase (GS, EC 6.3.2.3; formerly known as GSH synthetase) which ligates glycine to γ-GluCys, thus forming GSH (Reaction 2).

Cysteine is the rate-limiting substrate for GSH synthesis (Dringen et al. 1999). In the brain, mature neurons use exclusively cysteine for GSH synthesis, while astrocytes utilize both cystine and cysteine (Dringen et al. 1999; Dringen 2000). Cystine is taken up by astrocytes and microglial cells via cystine/glutamate exchanger also known as system xc_ (Shih et al. 2006). Only immature neurons express the xc(−) transport system (Murphy et al. 1990), while mature ones do not possess this transporter and therefore are unable to take up cystine for GSH synthesis. In the mature brain, neurons rely mainly on cysteine derived from GSH released by astrocytes into the extracellular space, where GSH is cleaved by sequentially acting the membrane-bound enzymes γ-GT and dipeptidase to constituent amino acids (Dringen et al. 1999; Dringen 2000). In addition to cysteine, neurons can utilize the cysteine-containing dipeptides γ-GluCys or cysteinylglycine (CysGly) for GSH synthesis (Dringen et al. 1999), although it is unclear how these dipeptides are taken up into neurons. Neural uptake of cysteine is mediated primarily by sodium-dependent excitatory amino acid transporter (EAAT) systems, known as excitatory amino acid carrier 1 (EAAC1 also termed EAAT3) (Shanker et al. 2001; Chen and Swanson 2003; Himi et al. 2003). EAAC1-deficient mice showed 30–40 % decreases in brain GSH levels, increased vulnerability to oxidative stress, as well as developed brain atrophy and behavioral abnormalities (Aoyama et al. 2006). The abovementioned data clearly indicate that transport of cysteine by EAAC1 system is also a rate-limiting factor for GSH synthesis in neurons.

3.2 Regulation of GSH Synthesis

Under physiological conditions, the rate of GSH synthesis depends on the expression and catalytic activity of GCL (Dalton et al. 2004; Dickinson et al. 2004). GCL is a heterodimeric protein composed of a heavy or catalytic (GCLC, Mr ~ 73,000) and a light or modifier (GCLM, Mr ~ 30,000) subunit, which are encoded in humans and rodents by different genes (Gipp et al. 1992; Dalton et al. 2004; Franklin et al. 2009). Only GCLC possesses all the enzymatic activity and is a subject of feedback inhibition by the end product GSH (Richman and Meister 1975; Seelig et al. 1984). GCLM is enzymatically inactive but plays an important regulatory function by increasing the V(max) and K(cat) of GCLC, by lowering the K(m) of GCL for glutamate and ATP, and by raising the K(i) for GSH-mediated feedback inhibition of GCL (Chen et al. 2005; Lu 2009; Yang et al. 2007). Thus, the holoenzyme is catalytically more efficient and less susceptible to inhibition by GSH than GCLC alone. GCLC-knockout mice showed embryonic lethality, demonstrating that the gene encoding this subunit was critical for development (Dalton et al. 2000, 2004). In turn, GCLM-knockout (Gclm−/−) mice are viable, but in the absence of this subunit, GCLC activity is insufficient, leading to a decrease of GSH level (Yang et al. 2002; Dalton et al. 2004). Alterations in GCL activity can result from regulation at multiple levels affecting only catalytic (GCLC) or both catalytic and modifier (GCLM) subunits of this enzyme. Many studies have focused on transcriptional regulation of GCL at the promoter level (Lu 2009).

GSH synthase (GS), the second enzyme participating in GSH biosynthesis, is composed of two identical subunits and has a Mr of approximately 118,000 Da. Mapping studies of the GS substrate binding sites indicate that the regions of the active site that bind glycine and the cysteinyl moiety of γ-GluCys are highly specific, while the γ-glutamyl moiety can be replaced by a variety of analogs. However, in contrast to GCL, GS is not subject to feedback inhibition by GSH. Moreover, GS overexpression had no effect on GSH level, whereas GCL overexpression markedly increased GSH level (Grant et al. 1997). Hence GCL, but not GS, is considered to be the rate-limiting enzyme in the GSH synthesis.

4 Oxidative Stress and GSH Synthesis in Schizophrenia and Bipolar Disorder

4.1 Glutathione Deficiency as a Marker of Oxidative Stress in Schizophrenia

Several studies have shown that the level of GSH, the major antioxidant and redox regulator, is decreased in a cerebrospinal fluid and medial frontal cortex (by 27 % and 52 % of control level, respectively) of drug-naïve schizophrenic patients (Do et al. 2000) as well as in the postmortem striatum (by 40 % of control) (Yao et al. 2006) and prefrontal cortex of those treated earlier with antipsychotic drugs (Gawryluk et al. 2011a). Moreover, there was a significant negative correlation between GSH levels and the severity of negative symptoms in schizophrenic patients (Matsuzawa et al 2008). In periphery, significantly lower levels of GSH were found in erythrocytes (Altuntas et al. 2000; Raffa et al. 2009, 2011; Micó et al. 2011) and plasma (Zhang et al. 2007; Dietrich-Muszalska et al. 2009; Raffa et al. 2009) of antipsychotic-free and chronically medicated schizophrenic patients in comparison to healthy control. As indicated by some genetic studies, the GSH deficit in schizophrenia seems to be linked to polymorphisms of genes encoding both catalytic (GCLC) and modifier (GCLM) subunits of glutamate cysteine ligase (GCL), an enzyme responsible for the biosynthesis of this tripeptide (Tosic et al. 2006; Gysin et al. 2007, 2009, 2011).

In addition, the activity of glutathione peroxidase (GPx), a key antioxidant enzyme involved in the elimination of hydrogen peroxide and lipid peroxides, was found to be unchanged (Mukerjee et al. 1996; Yao et al. 1998; Evans et al. 2003; Raffa et al. 2009), increased (Kuloglu et al. 2002; Raffa et al. 2011; Micó et al. 2011), or decreased (Abdalla et al. 1986; Altuntas et al. 2000) in erythrocytes of drug-naïve and antipsychotic-free schizophrenic patients when compared to controls. In a majority of chronically medicated schizophrenic patients treated with typical or atypical antipsychotic drugs, GPx activity in erythrocytes was found to be decreased (Altuntas et al. 2000; Ranjekar et al. 2003; Zhang et al. 2006; Ben Othmen et al. 2008; Raffa et al. 2009), while only in a few studies it was unchanged (Reddy et al. 1991; Evans et al. 2003) or increased (Herken et al. 2001; Akyol et al. 2002). Apart from the cytosolic form of GPx, there exists a related enzyme, called human plasma glutathione peroxidase (hpGPx), that is localized exclusively extracellularly. The level of this enzyme was significantly and positively correlated with the psychosis rating scores in schizophrenic patients both on and off haloperidol treatment (Yao et al. 1999). Decreased levels of GPx and glutathione reductase (GR), suggesting attenuated antioxidant functions of these two enzymes, were also found in the caudate region of brains from schizophrenic patients (Yao et al. 2006). Furthermore, a significantly lower level of Mu class of GST isoenzyme in the prefrontal cortex of schizophrenic patients than in nonpsychiatric controls has also been reported (Gawryluk et al. 2011b). The decreased level of GST Mu indicates that these patients had lesser ability to remove xenobiotics and also to detoxify endogenous substances such as quinines and lipid peroxidation products.

All the above-described results clearly indicate that due to a deficit of GSH and the decreased scavenging ability of GSH-related antioxidant enzymes, the redox balance of GSH/GSSG couple in peripheral tissues and in the brain cells of schizophrenic patients has to be shifted in favor of oxidative processes.

4.2 Changes in GSH Redox Status in Bipolar Disorder

Although the exact mechanisms underlying bipolar disorder (BD) are not completely understood, some studies suggest an involvement of oxidative stress and alterations in GSH redox status in the pathophysiology of the disease (Andreazza et al. 2007; Kuloglu et al 2002; Ranjekar et al. 2003; Machado-Vieira et al. 2007). In patients with BD, like in schizophrenic patients, oxidative stress was assessed indirectly by measuring the activities of antioxidant enzymes (glutathione peroxidase, GPx; superoxide dismutase, SOD; and catalase, CAT) in erythrocytes and serum as well as by determination of the content of thiobarbituric acid reactive substances (TBARS) as an index of lipid peroxidation. These studies, although less numerous than in schizophrenia, demonstrated that enzymatic activities of GPx, SOD, and CAT were impaired in BD (Kuloglu et al. 2002; Ranjekar et al. 2003; Ozcan et al. 2004; Andreazza et al. 2007; Kunz et al. 2008; Machado-Vieira et al. 2007) and lipid peroxidation was significantly enhanced when compared to nonpsychiatric control subjects (Kuloglu et al. 2002; Ranjekar et al. 2003; Ozcan et al. 2004; Machado-Vieira et al. 2007; Andreazza et al. 2007, 2008; Kunz et al. 2008). In line with studies on antioxidant enzymes in blood, a postmortem study of the hippocampus from BD patients demonstrated a lowered gene expression for several antioxidant enzymes in that structure, including GPx, CAT, SOD, and glutathione S-transferase (GST) (Benes et al. 2006). Moreover, recently diminished level of both reduced (GSH) and oxidized (GSSG) forms of glutathione in the prefrontal cortex (Gawryluk et al. 2011a) as well as oxidative damage to mitochondrial proteins (Andreazza et al. 2010) in that structure, increased lipid peroxidation in the cingulate cortex, and RNA oxidation in the hippocampus (Wang et al. 2009; Che et al. 2010) have been reported in postmortem brain from BD patients.

In addition, in patients not medicated with mood stabilizers, the level of class Mu of GST was decreased, while in BD patients treated with these drugs, the level of this isoenzyme was not different from that observed in controls (Gawryluk et al. 2011b). The latter effect is consistent with previous studies which demonstrated that mood stabilizers such as lithium, lamotrigine, and olanzapine increased GST expression and activity in primary cultured rat cerebral cortical cells (Shao et al. 2008; Bakare et al. 2009). Furthermore, chronic treatment with lithium and valproate increased the glutamate cysteine ligase (GCL) expression and GSH levels in these cultures (Cui et al. 2007). Since GST conjugates GSH with a variety of oxidized compounds to form nontoxic products, it has been suggested that lithium and valproate selectively target GST isoenzymes in order to produce neuroprotective effects against oxidative stress (Cui et al. 2007; Shao et al. 2008).

Summing up, all these studies suggest that oxidative stress and disturbances in GSH homeostasis play a significant role in the etiology of other psychiatric illnesses besides schizophrenia, e.g., BD, and the drugs increasing GSH content exert beneficial therapeutic effects in the treatment of this disease.

5 GSH-Deficient Animal Models of Schizophrenia

In general, the lowering of GSH level may result either from genetically determined alterations in the activities of GSH synthesizing enzymes (Tosic et al. 2006; Gysin et al. 2007, 2009) or from the limited availability of cysteine, a substrate for GSH synthesis (Dringen 2000). In experimental animals, the tissue GSH content can be decreased by inhibition of GCL enzymatic activity using specific compounds (Broquist 1992), by modulation of genes encoding catalytic or modifier subunits of this enzyme (Yang et al. 2002; Dalton et al. 2000, 2004), or by GSH depletion evoked by different endogenous or exogenous compounds that oxidize or conjugate the thiol group of this tripeptide (Masukawa et al. 1989). All these models have been used to check whether GSH deficit can lead to behavioral, morphological, and biochemical anomalies similar to those observed in schizophrenic patients.

5.1 Behavioral Effects of Glutathione Deficiency in Animals

The consequences of brain GSH deficit on cognitive functions were examined in the abovementioned animal models. It has been demonstrated that the GSH deficit induced in adult rats by intracerebra, chronic administration of L-buthionine-(S,R)-sulfoximine (BSO) combined with intracerebral injection of dopamine (DA), induced psychomotor (Shukitt-Hale et al. 1997) and spatial memory deficits in the water maze test (Shukitt-Hale et al. 1998).

According to the neurodevelopmental theory of schizophrenia, unknown gestational or perinatal events can impair brain development, leading to the establishment of an abnormal cerebral connectivity and detrimental effects of which may appear in adolescence, hence the consequences of GSH deficit during development were also studied in animal models (Rougemont et al. 2002; Castagné et al. 2004a, b; Cabungcal et al. 2007). It was found that rats treated with BSO between postnatal days 5 and 16 developed a strong ~50 % GSH deficit in various brain structures including the cortex (Rougemont et al. 2002). It is worth to mention that in the rat, the peak GSH concentration on postnatal day 7 is critical for brain development, as it occurs during a period of intensive synaptogenesis and may play the neuroprotective role during that time (Nanda et al. 1996). Rodents may compensate for GSH deficit by increasing ascorbic acid synthesis, as demonstrated in some studies on the Osteogenic Disorder Shionogi (ODS) mutant rats, which, like humans, cannot synthesize ascorbic acid (Castagné et al. 2004a, b; Cabungcal et al. 2007). Moreover, as the dysfunction of dopaminergic system is associated with schizophrenia, ODS rats were treated in the early postnatal period (between days 5 and 16) with BSO or the dopamine uptake inhibitor GBR 12909, alone and in combination, and later on in juvenile and adult rats, the object recognition test was performed (Castagné et al. 2004b). The object recognition test is based on the spontaneous tendency of rats to investigate objects and to favor novel objects versus familiar ones. In the latter study, it has been demonstrated that ODS rats receiving BSO and GBR 12909 failed to discriminate between familiar and novel objects, while ODS rats treated with either BSO or GBR 12909 alone had normal behavior in this test (Castagné et al. 2004b). Since after the combined treatment these rats preserved basic motor and sensory skills, the alterations observed in the test of object recognition can be attributed to cognitive impairment. The fact that ODS rats treated with BSO and GBR 12909 did not investigate more intensively the novel object than the familiar one suggests that increased dopaminergic tone coinciding with GSH deficiency during development can result in the long-term cognitive deficit observed in adult rats (Castagné et al. 2004b). Hence, the observed disturbances in the object recognition test in ODS rats treated with BSO and GBR 12909 are in line with the decreased object recognition capacity of schizophrenic patients as compared to healthy control subjects (Danion et al. 1999; Doniger et al. 2001; Heckers et al. 2000).

On the other hand, in the genetic model of GSH deficit, that is, in GCLM-knockout (Gclm−/−) mice, some subtle alterations in behavior of animals were observed (Steullet et al. 2010). In particular, Gclm−/− mice (4–6 months old) displayed an increased novelty-induced exploration, altered behavior during the object recognition task, altered emotion and stress-related behaviors, and lower response to delayed fear conditioning but had intact spatial learning and spatial memory (Steullet et al. 2010). The authors of the latter study revealed that genetically compromised GSH synthesis in Gclm−/− mice affected the morphological and functional integrity of hippocampal parvalbumin-immunoreactive (PV-ir) fast-spiking interneurons (FSIs), which are known to be altered in schizophrenia. In that study, it was demonstrated that the decreased GSH level in Gclm−/− mice caused a selective reduction of PV-ir interneurons in CA3 and dentate gyrus of the ventral hippocampus (VH) but not the dorsal hippocampus (DH) and a concomitant reduction of β/γ oscillations (Steullet et al. 2010). According to those authors, the altered behavior of Gclm−/− mice was associated with the functional disruption of the VH. Therefore, the hippocampus-dependent behaviors, known to implicate differentially the VH and DH, observed in Gclm−/− mice were discussed on the background of other literature data, in comparative manner. Mice with functional disruption of PV-ir FSI in the whole hippocampus had deficit in recognition of novel spatial arrangement of familiar objects and in novel object recognition (Fuchs et al. 2007). Gclm−/− mice that had dysfunctional only the VH recognized changes in spatial arrangement of objects, a task that requires functional DH (Gaskin et al. 2009), but explored novel and familiar objects with the same intensity. Thus, Gclm−/− mice recognized spatially displaced objects but showed altered behavior during an object recognition task. Moreover, mice with functional disruption of PV-ir FSI in the whole hippocampus had also impaired spatial working memory (Fuchs et al. 2007), while Gclm−/− mice that had normally functioning DH did not show such deficit. Consequently, mice with functional disruption of PV-ir FSI in the whole hippocampus were hypoactive, while Gclm−/− mice showed potent novelty-induced exploration. Steullet et al. (2010) suggest that hyperactivity in Gclm−/− mice could be induced by a decreased GABA inhibition in the VH. Finally, altered emotion and stress-related behaviors in Gclm−/− mice were in line with specific disruption of the VH but not the DH. It should be mentioned here that there is a growing evidence of structural and functional anomalies of the anterior hippocampus, a region of human brain that corresponds to the VH in rodents, in schizophrenic patients (Goldman and Mitchell 2004).

Moreover, in the most recent study, Kulak et al. (2012) using a different package of behavioral tests demonstrated that Gclm−/− mice displayed hyperlocomotion in the open field and forced swimming test but normal activity in home cage, suggesting that hyperlocomotion was selective to environmental novelty and mildly stressful situations. In the study by Kulak et al. (2012), similarly as in that performed by Steullet et al. (2010), spatial working memory in Gclm−/− mice remained unaffected. Gclm−/− mice showed a potentiated hyperlocomotor response to an acute amphetamine injection, impaired sensorimotor gating (prepulse inhibition), and altered social behavior when compared to wild-type mice (Kulak et al. 2012).

Altogether, the experimental data from different animal models of GSH deficiency reported above show that low level of this antioxidant and redox regulator can induce behavioral alterations that are relevant to those observed in schizophrenia.

5.2 Biochemical Consequences of Glutathione Deficiency in In Vitro and In Vivo Models

Glutathione deficiency seems to be related to the dysfunction of central dopaminergic, glutamatergic, and GABAergic neurotransmissions that are known to be implicated in the pathogenesis of schizophrenia (Laruelle et al. 2003; Moghaddam 2003; Cabungcal et al. 2006).

5.2.1 The Effect of Glutathione Deficit on Dopaminergic System Function

According to the dopaminergic hypothesis of the disease, a decrease in dopamine (DA) release in the prefrontal cortex (PFC) has been associated with negative symptoms, particularly cognitive deficits, while disinhibition of DA release in the dorsal striatum (nucleus accumbens) with the manifestation of positive symptoms, such as delusions and hallucinations. Evidence for dopaminergic dysfunction in schizophrenia is in majority indirect and is mainly based on the fact that most antipsychotic drugs being antagonists of DA D2 receptors alleviate some symptoms of this disease (Seeman et al. 1975; Kane and Marder 1993), while DA-releasing stimulants, such as amphetamine, induce psychosis (Janowsky and Risch 1979). Although the mechanisms underlying dopaminergic dysfunction in schizophrenia have not been elucidated yet, the hypothesis of GSH deficiency creates possibility to explain, at least, some of its aspects.

5.2.1.1 The Potential Role of Glutathione Deficiency in the Loss of Dendritic Spines

Metabolism of DA is closely linked with both intracellular and extracellular levels of GSH. DA is a major source of ROS in the mammalian brain, as an excess of this neurotransmitter can easily auto-oxidize to produce DA quinones that have potent oxidizing properties (Baez et al. 1997; Dagnino-Subiabre et al. 2000). Moreover, DA, via monoamine oxidase activity or DA quinones, through redox cycling, can induce the formation of H2O2 and O2 •−, which are known to cause lipid peroxidation, DNA modification, and protein oxidation (Baez et al. 1997; Bains and Shaw 1997; Rabinovic and Hastings 1998). Under physiological conditions, GSH detoxified DA quinones and H2O2 via reactions catalyzed by GST and GPx, respectively (see Sects. 2.2 and 2.3). However, under conditions of GSH deficiency, an excess of reactive intermediates of DA can disrupt cellular functions (Grima et al. 2003; Hastings et al 1996; Hirrlinger et al. 2002a), and this could contribute to the pathogenesis of schizophrenia. Although schizophrenia is not considered to be a neurodegenerative disease (Harrison 1997; Stevens and Casanova 1988; Garey 2010), low level of GSH in the PFC (Do et al. 2000; Gawryluk et al. 2011a) can sensitize neurons to DA-mediated dendritic degeneration (Hastings et al. 1996; Rabinovic and Hastings 1998; Grima et al. 2003). Hence, peroxidation reactions can lead locally to microlesions, affecting the synaptic contacts on dendritic spines of cortical pyramidal neurons, where excitatory glutamatergic terminals converge with dopaminergic ones (Goldman-Rakic et al. 1989). This may lead to the reduction of neuropil, mainly dendritic spines density, reported in postmortem histological studies of the PFC of schizophrenic patients (Selemon and Goldman-Rakic 1999; Glanz and Lewis 2000; Garey 2010; Glausier and Lewis 2012). As a degeneration of spines with their synaptic contacts may lead to abnormal cortico-cortical connectivity, these postmortem findings imply that neuronal integrity is compromised in schizophrenia. Thus, abnormal connectivity in the PFC may be responsible for part of symptoms, particularly those involving cognitive and perceptive functions (Garey 2010).

In line with the “reduced neuropil hypothesis” by Selemon and Goldman-Rakic (1999), it has been demonstrated that the application of DA under conditions of GSH deficiency into cultures of primary mouse cortical neurons induced a significant decrease in the number of neuronal processes which are considered to be spines analogous (Grima et al. 2003). Also, a deficit in brain GSH combined with a DA uptake inhibition during rat postnatal development caused a decrease in the number of dendritic spines of pyramidal neurons in the PFC (Do et al. 2004). So, morphological changes found in in vitro and in vivo studies could be related to morphological alteration reported earlier in the PFC of schizophrenic patients. Moreover, abnormal connectivity in the PFC was suggested by in vivo nuclear magnetic resonance imaging (NMR) studies in which it was found that N-acetyl aspartate (NAA), a marker of neuronal integrity, was decreased in schizophrenic patients (Callicott et al. 2000; Deicken et al. 2000; Yamasue et al. 2002). Interestingly, GSH deficiency decreases NAA level in the rat brain (Heales et al. 1995; Jain et al. 1991). Thus, it is likely that GSH deficiency and neuronal impairment are functionally linked in schizophrenia.

5.2.1.2 The Potential Role of Glutathione Deficiency in the Amphetamine-Induced DA Release in Subcortical Regions of the Brain

Besides the possible role of GSH deficiency in the loss of dendritic spines in the PFC (Grima et al. 2003; Do et al. 2004), it is believed that pathological low level of this antioxidant could cause disturbances in the dopaminergic neurotransmission (Jacobsen et al. 2005). The effect of a short-lasting GSH deficiency induced by the GSH synthesis blocker, BSO, on extracellular DA level in the nucleus accumbens of mice receiving amphetamine was investigated using a microdialysis method (Jacobsen et al. 2005). The latter study revealed that extracellular DA release after amphetamine (5 mg/kg) was increased twofold in the nucleus accumbens of GSH-deficient mice as compared to control mice with normal GSH level (Jacobsen et al. 2005). GSH deficiency per se did not change basal extracellular level of DA in the examined brain structure. These results indicate that GSH-deficient mice may experience accumbal hyperdopaminergia when DA transmission is considerably enhanced. The exacerbated amphetamine-induced DA release in the mouse model of GSH deficiency is consistent with the elevated amphetamine-induced DA release in the striatum of schizophrenic patients demonstrated by means of a single photon emission computed tomography (SPECT) and positron-emission tomography (PET) methods (Laruelle et al. 1996; Breier et al. 1997). Moreover, Laruelle et al. (1996) have shown that the elevated amphetamine effect in schizophrenic patients was associated with emergence or worsening of positive psychotic symptoms. These results suggest that psychotic symptoms in schizophrenia are related with exaggerated stimulation of dopaminergic transmission.

The mechanism underlying the interplay between GSH and DA after amphetamine administration in conditions of GSH deficiency is difficult to explain. However, based on previous reports on the general properties of GSH and DA, plausible explanations for these findings can be suggested. It is well known that GSH directly conjugates DA in vitro (Baez et al. 1997; Dagnino-Subiabre et al. 2000) and in vivo (Rabinovic and Hastings 1998). Since extracellular concentration of GSH is relatively high (1–2 μM; Lada and Kennedy 1997) and astrocytes permanently release this antioxidant into extracellular space (Dringen et al. 1997; Hirrlinger et al. 2002b), it is likely that GSH could scavenge the released DA. The amount of GSH released by astrocytes is correlated with the intracellular GSH content (Sagara et al. 1996). Therefore, deficit in intracellular GSH may lead to a concomitant decrease in extracellular pool of this antioxidant. Since GSH directly affects DA transmission by extracellular conjugation, deficit in its extracellular level could impair this mechanism that keeps the released DA under control. However, whether DA-GSH conjugation has functional significance remains to be established.

5.2.1.3 The Potential Role of Glutathione Deficiency in the DA-Mediated Modulation of Glutamatergic Transmission in the Prefrontal Cortex

GSH plays an important role in the redox control of various signal transduction pathways and gene expression (Sen 2000; Esposito et al. 2004). Thus, GSH deficit can alter the function of redox-sensitive proteins implicated in neurotransmission and synaptic plasticity such as NMDA receptors (Köhr et al. 1994; Choi et al. 2001), GABAA receptors (Amato et al. 1999), and ryanodine receptors (Bull et al. 2003) as well as calcium-activated K+ channels (DiChiara and Reinhart 1997) and L-type calcium channels (Campbell et al. 1996). These redox-sensitive proteins could affect neurotransmitter systems, i.e., dopaminergic, glutamatergic, and GABAergic, that are known to be dysfunctional in schizophrenia. In this section, a potential role of GSH deficiency in the DA-mediated modulation of glutamatergic transmission in the prefrontal cortex (PFC) will be discussed.

In the PFC, DA plays an important role in cognitive functions including working memory, reword, and attention. DA-containing neurons are localized in the ventral tegmental area and project to the PFC. A critical function of DA in this brain region is a modulation of glutamatergic transmission. NMDA responses are known to be modulated by DA via the activation of D1 and D2 receptors through multiple pathways acting on different targets, including NMDA receptors and voltage-gated calcium channels (Tseng and O’Donnell 2004). In brain slices obtained from developmentally matured rats, it was demonstrated that the application of NMDA and D1 agonist SKF38393 induced concentration-dependent excitability increases measured by whole-cell patch clamp technique, whereas the application of the D2 receptor agonist quinpirole induced concentration-dependent excitability decrease. The NMDA-mediated responses were potentiated by a D1 agonist while they were attenuated by a D2 agonist (Tseng and O’Donnell 2004).

The hypofunction of NMDA receptors reported in schizophrenia (Laruelle et al. 2003; Moghaddam 2003; Labrie and Roder 2010) could be evoked by pathologically low level of GSH in the brain. Therefore, it was interesting to check whether GSH deficiency could change DA receptor-mediated signaling in the PFC. Consistently, Steullet et al. (2008) examined, in cultures of embryonic cortical mouse neurons treated with BSO, how GSH deficiency influenced intracellular pathways implicated in DA signaling, namely, DA modulation of calcium responses to NMDA. In this study, it was shown that in the absence of DA, calcium responses evoked by NMDA were significantly larger in GSH-deficient neurons than in control ones (Steullet et al. 2008). In further experiments, it was established that the increased calcium responses to NMDA were due to the increased function of both L-type calcium channels and ryanodine receptors (RyRs), but not NMDA receptors. So, caffeine, an agonist of RyRs, induced significantly larger calcium release from internal stores in BSO-treated neurons than in control ones, confirming in this way the enhanced function of RyRs in the GSH-deficient neurons (Steullet et al. 2008). Moreover, the increase in the function of RyRs in neurons with low GSH content was in line with the finding that RyRs were redox sensitive, with oxidative conditions enhancing their function (Bull et al. 2003).

DA administration decreased calcium responses evoked by NMDA in GSH-deficient neurons but enhanced them in control ones. To exclude unspecific effects of BSO, it was evidenced that replenishing GSH levels by administration a membrane-permeable GSH analog abolished DA-mediated decrease in calcium responses to NMDA in the cultured GSH-deficient neurons. Since the blockade of DA D2 receptor by sulpiride caused a significant increase in calcium responses in GSH-deficient neurons but not in control ones, it was concluded that DA acting via DA D2 receptors decreased calcium responses evoked by NMDA under conditions of GSH deficiency (Steullet et al. 2008). In contrast, the blockade of DA D1 receptors with SCH23290 did not have any significant effect on DA-mediated modulation of calcium responses in GSH-deficient neurons while tending to decrease them in control neurons. The latter effect suggested that the activation of DA D1 receptors was involved in DA-induced increase of calcium responses evoked by NMDA only in control neurons.

So, the above-described results showed that a GSH deficit changed DA modulation of calcium responses evoked by NMDA. In cultured cortical neurons, NMDA-evoked calcium responses resulted from an initial calcium influx via the activation of NMDA receptors followed by secondary calcium influxes, through voltage-gated calcium channels (L-type channels) and calcium release from intracellular stores via the activation of RyRs (Hayashi et al. 1997). Steullet et al. (2008) examined which of these calcium sources contributing to the total response evoked by NMDA were altered by DA modulation under conditions of GSH deficiency. Consequently, it was demonstrated that DA decreased the calcium influx through L-type channels in GSH-deficient neurons but enhanced it in control ones. Such an effect of a GSH deficit on L-type channels was also observed when these channels were activated by either KCl or by specific agonist, BAY-K8644. DA is known to either increase the function of L-type channels via the activation of DA D1 receptors or decrease it via the activation of DA D2 receptors (Tseng and O’Donnell 2004). Hence, the results presented by Steullet et al. (2008) suggested that a GSH deficit strengthened DA D2 receptor-mediated decrease in the function of L-type calcium channels, occluding DA D1 receptor-mediated increase in the function of these channels.

Regarding internal source of calcium, Steullet et al. (2008) have found that the alteration of DA signaling in GSH-deficient neurons required the redox-sensitive RyRs. Because of the enhanced function of RyRs under condition of oxidative stress, DA evoked a larger release of calcium from intracellular stores in neurons containing a low level of GSH than in normal control. This, in turn, promoted in GSH-deficient neurons a decrease in the function of L-type channels via DA D2 receptor-mediated calcium-dependent pathway, whereas in control neurons, the function of these channels was enhanced. So, the deficit of GSH affected DA modulation of L-type channels but not the other calcium sources implicated in the responses to NMDA. As a consequence of the specific alteration of DA modulation of L-type channels, DA decreased NMDA responses in GSH-deficient neurons but increased them in normal neurons. So, GSH deficit, as that observed in some groups of schizophrenic patients (Do et al 2000; Yao et al. 2006; Gawryluk et al. 2011a), could play a significant role in the pathophysiology of this disease via dysregulation of dopaminergic and glutamatergic transmissions.

5.2.2 The Effect of Glutathione Deficiency on Glutamatergic System Function

In addition to dopaminergic abnormalities, also glutamatergic dysfunction has been associated with the pathogenesis of schizophrenia (Laruelle et al. 2003; Moghaddam 2003; Labrie and Roder 2010). Consistent with this view, reduced NMDA receptor function has been proposed as a cause of schizophrenia, because noncompetitive NMDA receptor antagonists like phencyclidine and ketamine induce psychotic and cognitive symptoms in healthy humans (Moghaddam 2003; Kantrowitz and Javitt 2010) and exacerbate symptoms in schizophrenic patients (Javitt and Zukin 1991; Krystal et al. 1994, 2005). Moreover, a loss of dendritic spines from cortical and hippocampal pyramidal neurons may be combined with the glutamatergic hypothesis of schizophrenia as NMDA receptors are present on their dendrites and probably dendritic spines. It is well documented that the vast majority of excitatory synapses (80–95 %) in the central nervous system are formed onto dendritic spines (Wilson 2007) and, as such, the spines perform a significant role in regulating neuronal excitability. In mature neuronal systems, pharmacological blockade of AMPA receptors or surgical deafferentation of glutamatergic inputs resulted in decreased spine density (Smart and Halpain 2000; Jacobs et al. 2003). Additionally, two recent studies showed that a constitutive reduction in NMDA receptor activity results in decreased spine density and cortical volume in the PFC and sensory cortex (Balu et al. 2012; DeVito et al. 2011). Although the loss of dendritic spines in schizophrenia was reported earlier (Glausier and Lewis 2012), the cause of NMDA receptor hypofunction in this disease has not been established as yet. However, an increasing number of experimental data suggest that GSH deficit may be an important factor contributing to this phenomenon (Steullet et al. 2006, 2010; Do et al. 2009). The potential role of DA in the loss of dendritic spines in condition of GSH deficiency has been presented in section “The Potential Role of Glutathione Deficiency in the Amphetamine-Induced DA Release in Subcortical Regions of the Brain”, whereas the role of this antioxidant in the regulation of NMDA and AMPA receptor functions is presented in Sect. 2.4.

GSH can affect NMDA receptor function via binding to its regulatory, redox-sensitive site and to glutamate recognition site. Redox sites of NMDA receptor are unusually sensitive to the oxidizing potential of the extracellular environment (Aizenman et al. 1989). Hence, oxidizing agents diminish NMDA receptor function, while reducing compounds, including GSH, enhance it (Köhr et al. 1994; Choi et al. 2001). Since GSH is the main regulator of the brain redox systems, it was assumed that GSH deficiency of the same magnitude as in schizophrenic patients (about 50 % of the control level) could lead to the dysfunction of NMDA receptors. A low GSH level could also alter NMDA receptor function via non-redox mechanisms because GSH can bind via its γ-glutamyl moiety to the glutamate recognition sites of NMDA and AMPA receptors and in this way modulate their function (Varga et al. 1997; see Sect. 2.4). To check experimentally whether GSH deficit could be a causal factor for NMDA hypofunction reported in schizophrenia, Steullet et al. (2006) examined in the CA1 region of the rat hippocampus how GSH deficit, induced by BSO administration, altered basal neurotransmission, cell excitability, and short-term and long-term plasticity. Using electrophysiological techniques, it was demonstrated that in hippocampal slices with low GSH level, the basal excitatory synaptic transmission that mostly depends on the AMPA receptor activation was not changed but NMDA receptor function was markedly depressed (Steullet et al. 2006). An extracellular level of GSH depends on its intracellular content and on the rate of GSH release from the glial compartment (Sagara et al. 1996; Hirrlinger et al. 2002b). Therefore, a deficit in intracellular GSH may result in a concomitant decrease in the extracellular content of this antioxidant. Consequently, the GSH deficit could lead to an excessive oxidation of the extracellular redox-sensitive site of NMDA receptors and to subsequent attenuation of their function. Results by Steullet et al. (2006) partially confirmed this view, as DTNB (5,5’-dithiobis(2-nitrobensoic acid)), a membrane-impermeable thiol-oxidizing compound, diminished pharmacologically isolated NMDA receptor-mediated field excitatory postsynaptic potentials (fEPSPs) in control, but not in BSO-treated slices. In turn, TCEP (tris(carboxyethyl)phosphine hydrochloride), a membrane-impermeable disulfide-reducing agent, increased NMDA responses more distinctly in BSO-treated slices than in control. These data indicate that under experimental conditions, the extracellular redox sites of NMDA receptors were fully oxidized in BSO-treated slices but were partially reduced in the control ones. Hence, the hypofunction of NMDA receptors under conditions of GSH deficit can be explained at least in part by an excessive oxidation of the extracellular redox-sensitive sites of NMDA receptors. In the above-reported study, it was also found that NMDA receptor-dependent long-term potentiation induced by high-frequency stimulation was impaired in GSH-depleted slices. The impairment of such synaptic plasticity could have adverse effects on normal brain functioning, including cognitive processing.

5.2.3 The Effect of Glutathione Deficiency on GABAergic System Function

An alteration of the GABAergic system in the prefrontal cortex (Lewis et al. 2005) and hippocampus (Zhang and Reynolds 2002) is a characteristic feature of the pathology of schizophrenia. Postmortem studies of these brain tissues have provided strong evidence that the GABAergic system is impaired in schizophrenia. These studies showed decreases in the concentration of cortical GABA (Perry et al 1979); in the activity of glutamic acid decarboxylase 67 (GAD-67), the GABA-synthesizing enzyme (Akbarian et al 1995; Hashimoto et al. 2003); and in the content of the calcium-binding protein parvalbumin (PV) in the fast-spiking interneurons (FSIs) of the prefrontal cortex and hippocampus (Hashimoto et al. 2003; Reynolds et al. 2004; Torrey et al. 2005). The existence of GABAergic deficit in schizophrenia was supported by in vivo studies using noninvasive methods. GABA measured in the human brain by magnetic resonance spectroscopy was shown to be decreased in schizophrenic patients (Rosso et al. 2006). Moreover, GABAergic inhibitory activity, as measured by transcranial magnetic stimulation (Daskalakis et al. 2002), was reduced.

NMDA receptor hypofunction could contribute to these abnormalities in the GABAergic system because the administration of NMDA receptor antagonists can cause the loss of parvalbumin and GAD-67 (Keilhoff et al. 2004; Kinney et al. 2006), alter GABA-mediated inhibitory control of cortical neurons (Homayoun and Moghaddam 2007), and disrupt the development of GABAergic neurons (Abekawa et al. 2007). Hence, the hypofunction of NMDA receptors induced by GSH deficiency (Sect. 5.2.2) could also affect the functioning of the GABAergic system in an indirect way. In line with this assumption, in the pharmacological (ODS rats treated with BSO + GBR during early postnatal development; see Sect. 5.1) and genetic (Gclm−/− mice) models of GSH deficiency, it was demonstrated that low level of this antioxidant caused a selective decrease of PV-ir interneurons in the rat prefrontal cortex (Cabungcal et al. 2006) and in the mouse dorsal hippocampus (Steullet et al. 2010). In the latter structure, a concomitant reduction of γ oscillations was also documented. Interestingly, γ oscillations were reduced in schizophrenic patients during impaired performance in cognitive tasks (Cho et al. 2006; Uhlhaas et al. 2008).

The decline of PV-ir FSIs has functional consequences because the activity of cortical pyramidal neurons is regulated by FSIs. These interneurons are necessary for the generation of γ neuronal synchrony that facilitates information processing and transfer within and between brain regions during cognitive tasks (Bartos et al. 2007; Sohal et al. 2009). Chronic GSH deficit in Gclm−/− mice affected the structural and functional integrity of PV-ir FSIs (Steullet et al. 2010), impairing information processing in the VH and leading to specific behavioral alterations, such as enhanced novelty-induced exploration and inadequate responses to stress described in more detail in Sect. 5.1. In conclusion, the alterations observed in the GABAergic system in animal models of GSH deficiency are consistent with that found in schizophrenic patients. Therefore, the participation of GSH deficiency in the pathogenesis of schizophrenia seems to be more and more convincing.

6 N-Acetylcysteine in the Treatment of Psychiatric Disorders

Considering GSH deficiency in the context of characteristic symptoms of schizophrenia, Matsuzawa et al. (2008) described the existence of a negative correlation between the brain GSH levels and the severity of negative symptoms of this disease. This observation suggested that agents increasing GSH levels could be potential therapeutic drugs for the treatment of negative symptoms of schizophrenia (Matsuzawa et al. 2008). The best thiol compound that fulfills such criterion seems to be N-acetylcysteine (NAC), as it acts as a precursor for GSH synthesis by supplying cysteine. NAC has been shown to penetrate successfully the blood-brain barrier and raise brain GSH levels in animal models (Farr et al. 2003). It enters the cell readily (Mazor et al 1996) and is then deacetylated to form L-cysteine. In addition to providing cysteine for GSH production, NAC acts as a direct antioxidant, although with less potency than that of GSH (Aruoma et al. 1989; Hussain et al. 1996).

For more than 30 years, NAC has been used for the treatment of paracetamol overdose, but now it is widely used as a mucolytic agent and in the treatment of HIV infection. As more information comes to light about NAC mode of action, its clinical applications are extending. Currently, potential application of NAC in the treatment of psychiatric disorders particularly in schizophrenia and bipolar disorder is being considered. Recently, in the double-blind, placebo-controlled study, it has been demonstrated that NAC addition (1 g twice daily over 24-week period) to antipsychotic therapy alleviated the negative symptoms, measured on the Positive and Negative Syndrom Scale. Furthermore, improvements in global functioning and reduction of abnormal movements, particularly akathisia, were also found in patients with chronic schizophrenia (Berk et al. 2008a). In addition, NAC relieved the depressive symptoms of bipolar disorder (BD) patients (Berk et al. 2008b). In another clinical study, Lavoie et al. (2008) reported that NAC application (1 g two times daily for 60 days) in schizophrenic patients mitigated an impaired mismatch negativity, which is an auditory evoked potential component related to NMDA receptor function. The abovementioned studies suggest that NAC has the potential to become a therapeutic drug for negative symptoms in schizophrenia and depressive symptoms in BD. These findings are particularly interesting because currently used antipsychotic drugs are rather ineffective against negative symptoms of schizophrenia.