sign in sign up
<<
Home , GABAA receptor

UMEÅ UNIVERSITY MEDICAL DISSERTATIONS New Series No 1091 ISSN 0346-6612 IBSN 978-91-7264-269-0 From the Department of Clinical Science, Obstetrics and Gynecology, Umeå University, Umeå, Sweden

Sex and stress steroid modulation of GABA mediated flux in rat CNS

Jessica Strömberg

Umeå 2007 Department of Clinical Science, Obstetrics and Gynecology Umeå Univesity SE-901 85 UMEÅ, Sweden

Cover illustration: Synaptic transmission by Sven Geier

Copyright © Jessica Strömberg ISBN 978-91-7264-269-0

Printed by Print & Media, Umeå University, 2007

2

To my mother

Det som inte går med våld går med mer våld

3 ABSTRACT...... 6

ABBREVIATIONS ...... 7

LIST OF ORIGINAL PAPERS...... 8

INTRODUCTION ...... 9

Neuroactive steroids...... 9 Synthesis...... 10 ...... 12 The clinical relevance of allopregnanolone...... 12 Neuroactive steroids and stress response ...... 13 The hormonal genomic pathway and the non-genomic pathway...... 15

Synaptic transmission...... 16 ...... 17

GABAA receptors ...... 19 GABAA receptor function, and antagonists...... 19 Subunits and distribution...... 21 Kinetic properties of the GABAA receptor...... 21

AIMS ...... 24

MATERIAL AND METHODS ...... 25

Animals...... 25

Substances ...... 25

35 [ S]TBPS binding assay ...... 25

Chloride ion uptake assay...... 26

Patch-clamp recording ...... 27

Statistical Analysis ...... 28

RESULTS AND DISCUSSION ...... 29

4 The allopregnanolone enhanced effect can be reduced by 3β-OH- 5α-pregnane-20-one in a selective way (paper I)...... 29 The effect of 3β-OH-5α-pregnane-20-one on the GABA mediated allopregnanolone enhancement ...... 29 The effect of 3β-OH-5α-pregnane-20-one on GABA and on positive GABAA receptor modulators...... 31 3β-steroids modulate both the GABA effect and the allopregnanolone enhancement on the GABAA receptor (Paper II) ...... 32 GABA and allopregnanolone sensitivity differ between brain regions ..... 33 The effect of 3β-steroids in different brain regions...... 34 The effect of 5β-pregnane-3β, 20(R)-diol on the kinetic properties of the GABAA receptor (Paper III)...... 37 5β-pregnane-3β, 20(R)-diol reduce the allopregnanolone prolongation of

sIPSCs τdeactivation...... 38 The effect of 5β-pregnane-3β, 20(R)-diol on external application of GABA...... 38 The effect of 5β-pregnane-3β, 20(R)-diol in presence of allopregnanolone on GABA application...... 41

Glucocorticoid metabolites modulate the GABAA receptor function (Paper IV) ...... 42 The 5β-isomer of THDOC is not as efficient as the 5α-isomer ...... 43 3α5α/5β-cortisol and 3α5α/5β-cortisone reduce the GABA mediated chloride ion uptake...... 44 The glucocorticoid metabolites interact with allopregnanolone ...... 45

SUMMARY AND CONCLUDING REMARKS ...... 47

CONCLUSIONS...... 49

ACKNOWLEDGEMENTS ...... 50

REFERENCES...... 52

5

Abstract Background: Sex and stress steroids are metabolized to 3α-hydroxy-pregnane-steroid metabolites such as allopregnanolone (Allo) and tetrahydrodeoxycorticosterone (THDOC). Allo and THDOC are neuroactive steroids that are metabolized in the brain and act in brain as potent positive GABAA receptor function modulators. Allo as well as THDOC levels increase during stress. Allo has been associated with a number of symptoms and malfunctions such as impaired memory function and negative mood symptoms in a subgroup of individuals both for animals and humans. Pregnane steroids with 3β-hydroxy-configuration (3β-steroids) have been shown to reduce the Allo enhanced GABA effect. Aims: The aims for the present thesis were to investigate the effect of 3β-steroids on the GABA mediated GABAA receptor function in presence of positive GABAA receptor modulators. Further, the regional variances between the 3β-steroids as well as the mechanism of the effect were studied. Finally, the effect of stress steroid metabolites on the GABAA receptor function was investigated. Results: 3β-OH-5α-pregnane-20-one reduced the Allo enhanced GABA mediated chloride ion uptake into cortical microsacs. The 3β-isomer reduced the efficacy of Allo without shift the concentration response curve. It is therefore suggested that the 3β-isomer has a non-competitive effect. Further, it was shown that the 3β-isomer reduced the Allo effect in a selective way since the 3β-isomer did not interact with other positive modulators or with GABA itself. Five tested 3β-steroids reduced the Allo enhanced GABA mediated chloride ion uptake in cerebral cortex and hippocampus as well as the Allo prolongation on spontaneous inhibitory postsynaptic currents (sIPSCs) in preoptic nucleus. In cerebellum on the other hand the 3β-steroids showed to have weaker or no effect compared to the other tested regions. Interestingly, in absence of Allo, two of the 3β-steroids positively modulated the GABA stimulated GABAA receptor function. In absence of Allo, 5β-pregnane-3β,20(R)-diol increased the desensitization rate of current response. In contrast to the reducing effect on the Allo induced prolongation on sIPSCs, the effect of the 3β-steroid on GABA application, was not altered in presence of Allo. The mechanism of the 3β-steroid is therefore suggested being desensitization dependent in contrast to Allo, which has been suggested to decrease the GABA unbinding rate. In contrast to the enhanced effect of Allo, glucocorticoid metabolites reduced the GABA mediated chloride ion uptake in a concentration dependent way. The results in present thesis indicate that both sex and stress steroid metabolites interact with the GABAA receptor function. The knowledge that diversity of endogenous steroids interact with the GABAA receptor function is of importance for further understanding of different sex and stress steroid related symptoms and syndromes.

Keywords: Allopregnanolone, chloride ion uptake, patch-clamp, 3β-steroids, kinetic parameters, rat brain

6 Abbreviations 3β5β-THDOC 5β-pregnane-3β, 21-diol-20-one 3β-HSD 3β-hydroxysteroid dehydrogenase-isomerase enzyme 3β-steroid 3β-hydroxy configuration of reduced A-ring pregnane steroids 5αR 5α-reductase 11β-HSD 11β-hydroxysteroid dehydrogenase ACTH adrenocorticotropic hormone Allo allopregnanolone Allo-THF 3α-OH, 5α-cortisol, allotetrahydrocortisol ANOVA analysis of variance B bound state of the receptor CRH corticotropin releasing hormone D desensitized state of the receptor Delta amplitude end-to peak amplitude ratio DHDOC 5α-dihydrodeoxycorticosterone DHEA DHP dihydroprogesterone DMSO dimethyl sulfoxide EC20, 50, 75 the concentration of substances that produce 20, 50 or 75 % respectively, of maximal response. Emax the maximal response GABA gamma-aminobutyric acid GABAA, receptor gamma-aminobutyric acid type A receptor GAD glutamic acid decarboxylase GAT GABA transporter HPA hypothalamic-pituitary-adrenal IC50 the concentration of substance that produce half-max. response MPN medial preoptic nucleus O open state of the receptor PMDD premenstrual dysphoric disorder PMS premenstrual syndrome PREG pregnenolone SEM standard error of mean sIPSCs spontaneous inhibitory postsynaptic currents t50 desensitization rate measures as the time elapsed from the peak amplitude to the current had decline 50 % of the amplitude between the peak amplitude and the end amplitude TBPS t-butylbicyclophosphorothionate THDOC tetrahydrodeoxycorticosterone THE 3α-OH, 5β-cortisone, tetrahydrocortisone THF 3α-OH, 5β-cortisol, tetrahydrocortisol τdeactivation time course of deactivation UB unbound state of the receptor VIAAT vesicular inhibitory amino acid transporter

7 List of original papers

This thesis is based on the following papers. These will be referred to in the text by their Roman numerals.

I. Lundgren P., Stromberg J., Backstrom T., Wang M. (2003) Allopregnanolone stimulated GABA-mediated chloride ion flux is inhibited by 3beta-hydroxy-5alpha-pregnane-20-one (isoallopregnanolone). Brain Res 982:45-53

II. Stromberg J., Haage D., Taube M., Backstrom T., Lundgren P. (2006) modulation of allopregnanolone and GABA effect on the GABA-A receptor. Neuroscience 143:73-81

III. Stromberg J., Lundgren P., Taube M., Backstrom T., Wang M., Haage D. The neuroactive steroid 5b-pregnane-3b, 20(R)-diol alters the kinetic properties of the GABAA-receptor differently from . Manuscript

IV. Stromberg J., Backstrom T., Lundgren P. (2005) Rapid non-genomic effect of glucocorticoid metabolites and on the gamma- aminobutyric acid-A receptor. Eur J Neurosci 21:2083-2088

These papers are reproduced with the permission of the copyright holders Blackwell Publishing, Oxford, United Kingdom and Elsevier Ltd, Kidlington, United Kingdom.

8 Introduction Neuroactive steroids are steroids that act on the nervous system and they include those that are synthesized in the brain, those that are metabolized locally from circulating precursors, and synthetic steroids (Gasior et al., 1999; Paul and Purdy, 1992). The scope of the term “neuroactive steroids” was extended during the 1990s when the relation between steroid hormones, their metabolites and several functions in the nervous system was discovered. It was shown that steroid hormones were not only synthesized within the brain but also enzymatically converted into several neuroactive metabolites (for review: Melcangi et al., 1999). The neuroactive steroids, it was realized, could exert their action through genomic steroid receptors such as the receptor (PR) and also through non-genomic receptors such as those that belong to the -gated superfamily, e.g. the GABAA and NMDA receptors (Baulieu et al., 2001).

Gamma-aminobutyric acid (GABA) is the major inhibitory amino acid in the vertebrate . GABA is released by exocytosis from terminal boutons on the presynaptic , and activates GABA receptors on the postsynaptic neuron. Hitherto, three different classes of GABA receptors have been discovered – the two ionotropic GABAA and

GABAC receptors and the metabotropic GABAB B receptor (Barnard et al., 1998). + GABAB B receptors act via a G to increase conductance in K channels (for review: Mehta and Ticku, 1999). The GABAA receptor coupled to the chloride ion channel can be modulated both positively and negatively. The neuroactive progesterone metabolite allopregnanolone, 3α-OH-5α-pregnane-20- one (Allo) is one of the most potent positive GABAA receptor modulators. Allo has been associated with a number of symptoms and disorders such as impaired memory function and negative mood symptoms in women with premenstrual disorder. Allo is not only an active sex steroid metabolite but is also released in response to stress in both men (Droogleever Fortuyn et al., 2004) and women (Girdler et al., 2001).

As the neuroactive steroid Allo is associated with hormonal symptoms, is released in stress and interacts with the ligand-gated GABAA receptor, it was of interest therefore to investigate further the relation between neuroactive steroids and GABAA receptor function. In this thesis, I have examined the effects of sex and stress steroid metabolites on GABAA receptor function in the rat brain. Neuroactive steroids Hormones can be divided into different groups: 1) water-soluble, including a) amino acid derivates and b) peptides and , and 2) fat-soluble, including steroids (Berne et al., 1998). Steroid hormones are synthesized from

9 in the adrenal glands, in the gonads and in the . They play a major role in sexual development and behavior. Some of the steroid hormones can also be synthesized de novo in the brain. Steroid hormones can further be divided into three classes, 1) androgens (e.g. ), 2) estrogens (e.g. estradiol) and 3) progestines (e.g. progesterone) (Falkenstein et al., 2000). Steroid hormones can easily penetrate the and bind to specific intracellular receptors, thereby influencing gene expression. This type of action takes minutes to hours or even days for full expression. However, hormones can also exert their effect within seconds to minutes by binding to specific receptors in the cell membrane (Makara and Haller, 2001). Synthesis The synthesis of steroid hormones and other neuroactive steroids in CNS starts with cholesterol or with steroidal precursors transported to the brain via the bloodstream from peripheral organs. Cholesterol can also be synthesized locally within the brain and the peripheral nerves (Morell and Jurevics, 1996) and in cultured glial cells (Hu et al., 1989; Jung-Testas et al., 1992). The enzymes necessary for the synthesis of neuroactive steroids can be divided into two main groups: The cytochrome P450s and the non-P450 group that are in general either mitochondrial or microsomal (Mellon and Vaudry, 2001). Cytochromes P450 are oxidases, able to catalyze the oxidative conversion of not only many steroids but also of lipids, various xenobiotics and environmental toxins (Miller, 1988).

The first step in the synthesis of all steroid hormones is the conversion of cholesterol to pregnenolone (PREG) by P450scc (Baulieu et al., 2001; Mellon and Vaudry, 2001) (Fig. 1). P450scc is a mitochondrial-specific hydroxylase involved in cholesterol side-chain cleavage (SCC) (Baulieu et al., 2001; Mellon and Vaudry, 2001) that has been found in myelinating glial cells, in isolated oligodendrocytes (Hu et al., 1987) and also in (Baulieu et al., 2001). PREG can then be oxidized to progesterone (PROG) by the 3β-hydroxysteroid dehydrogenase-isomerase enzyme (3β−HSD) located in the endoplasmic reticulum (Mellon and Vaudry, 2001). The enzyme 3β−HSD consists of several isoforms that are present in the rat peripheral nervous system and in most parts of the rat brain where they are found in glial cells and in neurons (Guennoun et al., 1995; Guennoun et al., 1997; Sanne and Krueger, 1995).

Both PREG and PROG can undergo 17α-hydroxylation to 17α- hydroxypregnenolone and 17α-hydroxyprogesterone (Baulieu et al., 2001; Mellon and Vaudry, 2001). The 17α-hydroxylated PREG can then be metabolized to dehydroepiandrosterone (DHEA) and further to androstenedione, testosterone or estradiol. 17α-hydroxyprogesterone undergoes the transformation to androstenedione directly (Mellon and Vaudry, 2001) or may be metabolized to 11-deoxycortisol and further to cortisol (Fig. 1).

10

Figure 1. Diagram of the sterogenetic pathways. The synthesis of allopregnanolone, THDOC, cortisol and corticosterone. Enzymes involoved are indicated at the arrows.

Progesterone can also enter a number of biochemical pathways via the 5α- reductase-3α-hydroxysteroid dehydrogenase (5αR-3αHSD) enzymatic system to form a variety of neuroactive steroids such as Allo and 3α5α- tetrahydrodeoxycorticosterone (THDOC) (Celotti et al., 1992) (Fig. 1). The 5αR-3αHSD is very adaptable. Steroids with the delta 4-3 keto configuration are first 5α-reduced by 5αR, which may enhance the activity of the given steroid (Melcangi et al., 1999). The configuration is subsequently 3α−hydroxylated by 3α-HSD, which modifies the steroids so that their metabolites can act on receptors other than typical genomic steroid receptors

11 (Melcangi et al., 1999). PROG is metabolized by 5αR into dihydroprogesterone (DHP) and subsequently by 3αHSD into Allo, and deoxycorticosterone (DOC) is metabolized into 5α-dihydrodeoxycorticosterone (DHDOC) and further to THDOC (Melcangi et al., 1999). The 5αR-3αHSD is also involved in metabolism of cortisol and corticosterone into 3α-OH, 5α-cortisol, allotetrehydrocortisol (Allo-THF); 3α-OH, 5β-cortisol, tetrahydrocortisol (THF) and 3α-OH, 5β-cortisone, tetrahydrocortisone (THE) (Tsilchorozidou et al., 2003) (Fig. 2). Allopregnanolone The 3α-hydroxy-A ring-reduced pregnane steroid, Allo is synthesized in the brain (Melcangi et al., 1999), in the human corpus luteum of the ovary (Ottander et al., 2005), and in adrenal cortex (Corpechot et al., 1993; Purdy et al., 1991). The corpus luteum is formed in ovulatory menstrual cycles. The menstrual cycle starts with menstrual bleeding during day 1–5, followed by the follicular phase days 6–12, the ovulatory phase days 13–15, the luteal phase days 16–23 and finally the premenstrual phase days 24–28 (Mumenthaler et al., 2001). In anovulatory cycles, where no corpus luteum is formed, either progesterone or Allo is synthesized (Backstrom et al., 2003). The level of Allo in both brain and plasma correlates well with the level of progesterone (Paul and Purdy, 1992; Wang et al., 1996). In ovulatory menstrual cycles, Allo concentration varies with the different phases. Circulating levels of Allo have been measured and found to be two to four times higher in the luteal phase than in the follicular phase (for review: Havlikova et al., 2006; Wang et al., 2001). The highest concentration of Allo is however reached during pregnancy (Bixo et al., 1997; Paul and Purdy, 1992). In the third trimester of human pregnancy, plasma levels of Allo and were approximately 100 nM (Paul and Purdy, 1992). In both cycling and pregnant rats, Allo levels in the brain are higher than in plasma (Corpechot et al., 1993).

In rat CNS, it is thought that the regulation of Allo is through an elimination pathway of hydroxylation of 3β-OH-5α-pregnane-20-one (Stromstedt et al., 1993). Allo is first converted to DHP catalyzed by 3α-HSD and then to 3β-OH- 5α-pregnane-20-one by 3β-HSD. 3β-OH-5α-pregnane-20-one is then further hydroxylated into more polar metabolites (Stromstedt et al., 1993). Circulating levels of the 3β-OH-5α-pregnane-20-one seem to be approximately half the level of Allo in both the follicular and luteal phases (Havlikova et al., 2006). The clinical relevance of allopregnanolone Allo levels in brain and plasma have been associated with varied modulation of the behavioral and biochemical responses to multiple conditions, e.g. anxiety, depression, stress, convulsions, and sleep (Paul and Purdy, 1992). In fertile

12 women, a variety of symptoms and disorders have been related to the ovulatory menstrual cycle. The most common symptom pattern is the premenstrual dysphoric disorder (PMDD) or the less severe premenstrual syndrome (PMS) (for review: Backstrom et al., 2003). The symptoms begin in the luteal phase, and resolve within a few days after the onset of menstrual bleeding (Sundstrom et al., 1999). The typical symptoms of PMDD and PMS, irritability, anxiety, memory impairment and mood changes, require ovulation and the consequent formation of a corpus luteum (Rapkin et al., 1997; Sundstrom et al., 1999; Wang et al., 2001) and are therefore related to the secretion of progesterone and Allo. The memory impairment has also been shown in rats by Morris Water Maze where the effect of Allo was achieved rapidly after injection (Johansson et al., 2002). Rats showed clear memory impairment when Allo was injected 8 minutes but not 20 minutes before test (Johansson et al., 2002).

In women with partial epilepsy a symptom pattern has also been described (for review: Backstrom et al., 2003). The frequency of epileptic seizures in the luteal phase is low, while seizure frequency increases during menstruation (Backstrom et al., 2003). In pregnancy, where the serum level of Allo rises (Luisi et al., 2000), it is thought that Allo may contribute to the psychological changes seen in pregnancy and postpartum (Majewska et al., 1989). Allo has also been linked to depression, as a difference in the concentration of 3α-reduced neuroactive steroids has been observed in patients suffering from major depression, compared to healthy controls (for review: Eser et al., 2006). In depressed patients, the concentration of Allo is decreased in both plasma (Eser et al., 2006; Romeo et al., 1998) and cerebrospinal fluid (CSF) (Eser et al., 2006; Uzunova et al., 1998) compared to healthy controls. When plasma and CSF levels of Allo decreased, levels of 3β-OH-5α-pregnane-20-one increased (Romeo et al., 1998). Neuroactive steroids and stress response Both Allo and THDOC are released after stress stimuli. The concentration of Allo and THDOC increased approximately 4–8 fold in rat cortex and plasma after exposure to acute stress (Barbaccia et al., 1998; Purdy et al., 1991; Serra et al., 2000). The concentrations of Allo and THDOC peaked 10 to 30 minutes after stress stimuli (Barbaccia et al., 1998; Purdy et al., 1991; Reddy, 2003; Serra et al., 2000). Neither Allo nor THDOC were detectable in plasma before or after stress stimuli in adrenalectomized rats (Purdy et al., 1991). However, levels of Allo but not of THDOC were measurable in the brain of the adrenalectomized rats (Purdy et al., 1991).

13

Figure 2. Pathways of cortisol metabolism. The synthesis of tetrahydrocortisol, allotetrahydrocortisol and tetrahydrocortisone. Enzymes involved are indicated at the arrows.

In contrast to Allo, which can be synthesized both de novo in the brain and in the adrenal cortex during stress, THDOC appears to be derived almost exclusively from the adrenal cortex but can be converted from its precursor in the brain (Reddy, 2003). In rats, plasma levels of Allo and THDOC are normally about 1–5 nM but increase to about 15–30 nM after acute stress stimuli (Barbaccia et al., 1998; Reddy and Rogawski, 2002) and to about 40–60 nM during pregnancy (Concas et al., 1998). Levels of Allo and THDOC in the brain increase from approximately 20 nM in control rats to 100–140 nM after exposure to foot shock stress (Barbaccia et al., 1998).

Stress is often associated with an unpleasant event. Indeed, stressful events are defined as events that promote activity in physiological systems (McEwen, 2002). Stress responses are initiated by activation of the HPA axis that are thought to be beneficial in acute stress (Dubrovsky, 2000) while in chronic stress, continuous activation of the system can lead to pathological states (Reagan and McEwen, 1997). Stressors initiate the release of corticotropin releasing hormone (CRH) from the hypothalamus via e.g. the limbic system (emotion of fear and anxiety) (Herman et al., 2005). CRH stimulates the secretion of adrenocorticotropic hormone (ACTH) from the anterior pituitary, which in turn triggers the release of glucocorticoids from the adrenal cortex. The

14 adrenal cortex also secretes mineralocorticoids and sex hormones (Berne et al., 1998). However, stressors also activate the sympathetic nervous system that induces the secretion of catecholamines from the adrenal medulla (Korte, 2001).

After HPA axis stimulation, glucocorticoids and mineralocorticoids are released (Dubrovsky, 2000). In humans, both corticosterone and cortisol are secreted, but the predominant glucocorticoid is cortisol (Ganong, 1997). The peak concentration of cortisol in plasma is reached after about 30 minutes (Florkowski et al., 1996). Cortisol has a half-life of 60–90 minutes (Ganong, 1997). In humans, about 50% of secreted cortisol is metabolized to Allo-THF, THF and THE (Fukushima et al., 1960) (Fig. 2). The 11β-hydroxysteroid sehydrogenase (11β-HSD), catalyses the interconversion of active cortisol and its inert 11-keto forms, cortisone (Seckl and Walker, 2004) (Fig. 2).

In rats on the other hand, corticosterone is the predominant glucocorticoid but cortisol is also released (Carroll et al., 1975). In rat brain, glucocorticoid levels can reach up to 1 µM with peak concentrations occurring 30 minutes after stress stimuli (Carroll et al., 1975; de Quervain et al., 1998; Makara and Haller, 2001). Glucocorticoids act via genomic intracellular mineralocorticoid and glucocorticoid receptors which are distributed throughout various brain regions (Korte, 2001; Wellman, 2001). However, glucocorticoids have also diverse non- genomic effects (Borski, 2000; Majewska et al., 1985; Makara and Haller, 2001; Wheling, 1997). Cortisol metabolites have shown to reduce the GABA-mediated chloride ion flux and THDOC-enhanced stimulation (Penland and Morrow, 2004). As glucocorticoids and their metabolites activate ligand-gated receptors in rat brain, glucocorticoids can be defined as neuroactive steroids. The hormonal genomic pathway and the non-genomic pathway The physiological actions of steroid hormones, gonadal and adrenal steroids (androgens, estrogens and corticoids) are generally exerted via genomic intracellular receptors and therefore may take several hours or even days (Paul and Purdy, 1992). In contrast, other progesterone metabolites, neuroactive steroids, can produce their effects via non-genomic receptors within minutes or even seconds (Paul and Purdy, 1992). Most steroids are transported in the blood bound to corticosteroid-binding globulin (CBG, transcortin), sex-hormone binding globulin (SHBG) or albumin (Ganong, 1997). Bound steroids appear to be physiologically inactive and act as a circulating reservoir. The proportion of steroids that is free in plasma is only about 2–4% (Ganong, 1997). The ability to penetrate the blood-brain barrier by free diffusion depends on the number of hydrogen bonds the compound forms in aqueous solution. Sex steroid hormones penetrate easily into the brain, in contrast to cortisol which doesn’t (Pardridge and Mietus, 1980b). Steroids bound to albumin seem to be transported into the

15 brain, while CBG- or SHBG-bound steroids do not enter the brain (Mellon and Vaudry, 2001; Pardridge and Mietus, 1980a).

Estrogen and progesterone are two typical steroid hormones that have intracellular receptors. The estrogen receptor consists of two isomers, ERα and ERβ, and the progesterone receptor (PR) consists of two receptors, PR-A and PR-B (Ganong, 1997). When the hormones reach the target cell, they can easily release the carrier protein and diffuse through the cell membrane. Inside the cell, the steroid hormone binds to a specific receptor protein, forming the hormone/receptor complex (Falkenstein et al., 2000). The complex is activated and able to bind to specific regions of the DNA, called the hormone response elements. The genomic effects are then followed – either by genetic transcription or gene repression. The gene transcription produces messenger RNA (mRNA), which then codes for the production of specific proteins while gene repression inhibits the transcription (Falkenstein et al., 2000).

Non-genomic effects are exerted via a different pathway. Non-genomic steroid effects are achieved within seconds or minutes, and invariably within 15 minutes (Makara and Haller, 2001). Non-genomic effects can also be rapidly washed out after steroid removal (Borski, 2000; Makara and Haller, 2001; Wheling, 1997). Blocking genomic receptors makes it possible to investigate whether a specific steroid effect is exerted via non-genomic receptors. If the steroid effect is intact, the action is made via the non-genomic pathway (Makara and Haller, 2001). Cell preparations in which the nuclei are no longer intact can be regarded as forming a non-genomic environment (Makara and Haller, 2001). Synaptic transmission Most neuronal communication involves transmitting electrical signals from one cell to another through chemical . Cell communication in myocardial cells for example, can also be via direct rapid passage of an electrical current, called electrical transmission. These cells are joined together by gap junctions that allow conduction in both directions (Berne et al., 1998). Cells with chemical synapses communicate by releasing transmitters. The transmitters are released from the presynaptic neuron into the extracellular synaptic cleft and bind to receptors on the postsynaptic membrane (Fig. 3). The time it takes for transmitters to bind to the postsynaptic receptor after entering the synaptic cleft is called synaptic delay. Classical transmitters, such as acetylcholine and GABA are synthesized in the cytoplasm of the nerve terminals, while the synaptic vesicles are produced in the cell body and transported through the axon to the nerve terminals (Reimer et al., 1998). The transmitters are packaged in the synaptic vesicles by the Golgi complex (Rothman, 1996). The filled vesicles are then stored in clusters close to the presynaptic membrane. The neurotransmitters are released by exocytosis, triggered by influx of Ca2+ through the

16 presynaptic membrane (Evans and Zamponi, 2006). The Ca2+ ion influx occurs momentarily when voltage-gated calcium channels open as a result of stimulation by action potentials (Ganong, 1997). The entry of Ca2+ ions causes the synaptic vesicles to fuse with the membrane and release their contents into the synaptic cleft. The amount of transmitters in one vesicle is called a quantum (Edwards et al., 1990; Reimer et al., 1998).

The chemical receptors can be divided into two groups, depending on whether they achieve direct gating or indirect gating of the ion channels. In direct gating, the chemical receptor is coupled to an ion channel and is called ionotropic receptor. These receptors mediate fast inhibitory or excitatory with a time course of activation of 1–100 ms (Luddens and Korpi, 1995). When the neurotransmitter binds to the receptor, the receptor undergoes a conformational change that opens the channel. The indirect chemical transmission is slower, about 0.1 s to minutes, since the receptor and the ion channel are separate and the communication is mediated by a second messenger (Luddens and Korpi, 1995). When the receptor has been bound by a neurotransmitter, a GTP-protein (G-protein) is activated. The G-protein activates a second messenger cascade that can for example result in modulation in protein expression or in ion channel activity, such as the GABABB receptor. Neurotransmitters There are several criteria that must apply before a substance can be accepted as a transmitter substance (Berne et al., 1998; Pinel, 2000). 1) The substance has to be found and synthesized in the presynaptic cell. 2) The substance has to be released by the presynaptic cell as a response to a specific stimulus. 3) The postsynaptic cell response upon external application of the substance must be the same as that produced by stimulating the presynaptic cell. 4) The effects of presynaptic stimulation and the external application of the substance should be similarly affected by drugs. 5) After release in the synaptic cleft, the neurotransmitters must undergo deactivation. The mechanisms of deactivation are through reuptake or through deactivation by enzymes. Neurotransmitters can be divided into four classes of small molecules: acetylcholine, biogenic amines, purines and amino acids (Purves et al., 2004). In addition, neurotransmitters can also belong to a class of large molecules, the neuropeptides and to the unconventional group, such as soluble gases. The most common amino acid neurotransmitters are glutamate, aspartate, glycine and gamma aminobutyric acid (GABA). Glutamate and aspartate are excitatory transmitters, while glycine and GABA are inhibitory neurotransmitters in adult mammals. Glutamate and GABA are the most common neurotransmitters in the CNS.

17

Figure 3. GABAergic synaptic transmission. GABA is synthesized from glutamate by glutamic acid decarboxylase (GAD) with the co-factor pyridoxal phosphate. GABA is packed into synaptic vesicles by vesicular inhibitory amino acid transporters (VIAAT) and released when an reaches the postsynaptic terminal. GABA binds to different GABA receptors. Here is the ionotropic GABAA receptor presented. GABA activates the chloride ion channel coupled GABAA receptor by opening the chloride ion channel. Negative charged chloride ions flux into the postsynaptic cell and hyperpolarize it. Released GABA is recycled directly by reuptake into the presynaptic terminal, and indirectly via the glutamine cycle in the presynaptic terminal and in neighboring glial cells. In both cases the uptake is performed by the Na+-dependent plasma membrane transporter, GABA-transporter (GAT). (Figure reproduced with permission by Sinauer Associates Inc.)

The inhibitory neurotransmitter GABA is synthesized from glutamate by the specific glutamic acid decarboxylase (GAD). The synthesis occurs in certain neurons, so called GABAergic neurons for example interneurons, e.g. stellate cells in the neocortex (Brodal, 1992; Peters, 2002), and cerebellar purkinje cells (Brodal, 1992). It has been estimated that GABA is the transmitter in about 30% of all synapses in CNS (Berne et al., 1998). Although GABA is the most common inhibitory transmitter in the vertebrate CNS, GABA acts as an excitatory transmitter both during development (Ben-Ari et al., 1997) and in low invertebrates (Arshavsky et al., 1993).

18 GABAA receptors

GABAA receptor function, agonists and antagonists

The GABAA receptor is a member of the ligand-gated ion channel (LGICs) superfamily, composed of five subunits that are arranged together to form an ion channel (Chebib and Johnston, 1999; Macdonald and Olsen, 1994) (Fig. 4). Each subunit has four transmembrane domains (Macdonald and Olsen, 1994). The ion channel is formed by subunits arranged in such way that the second transmembrane domains form the wall of the channel pore. When GABA binds to the receptor, the neuronal membrane conductance for chloride ions increases (Sieghart, 1992). After the ion channel has opened, the ion flux depends on the gradient across the membrane. The chloride ion concentration inside neuronal cells is normally low. Therefore, the gradient across the cell membrane will give rise to movement of the chloride ions into the cell. The increased chloride ion influx results in a membrane hyperpolarization, and neuronal excitability is reduced (Sieghart, 1995). However, the GABA activity can also lead to an excitation by inhibition of inhibitory neurons, so called disinhibition. GABAA receptors are activated by GABA and by structural analogues of GABA, e.g. muscimol (Macdonald and Olsen, 1994). The GABAA receptor can also be modulated by both agonists and antagonists. The two 3α-hydroxy-A ring- reduced pregnane steroid metabolites, Allo and THDOC, are thought to be among the most potent neuroactive steroids (Majewska, 1992; Paul and Purdy, 1992; Purdy et al., 1990). Allo and THDOC bind with high affinity to the ligand-gated GABAA receptor and act as positive allosteric modulators of the GABAA receptor (Paul and Purdy, 1992). Allo and THDOC increase the frequency and duration of the opening of the chloride ion channel and act in a similar way to and (Lambert et al., 1995; Paul and Purdy, 1992). It has been shown that Allo positively modulates GABAA receptor function even at nanomolar concentrations (Fodor et al., 2005). The synthetic steroid alphaxalone (5α-pregnane-3α-ol-11, 20-dione) is also a potent general anaesthetic steroid that mimics the endogenous progesterone metabolites by modulates the GABAA receptor function (Callachan et al., 1987; Harrison and Simmonds, 1984). Possibly one of the most commonly used drugs, , has a positive modulation of the GABAA receptor’s , and functions (for review: Luddens and Korpi, 1995). The two common drug classes benzodiazepines and barbiturates are agonists that allosterically modulate the binding of GABA (Macdonald and Olsen, 1994). Barbiturates but not benzodiazepines can also directly activate chloride ion influx (Macdonald and Olsen, 1994; Schwartz et al., 1984).

19

Figure 4. A schematic presentation of a GABAA receptor. When GABA binds to the receptor, chloride ions flux through the chloride ion channel. In presence of GABAA receptor agonists the chloride ion flux increase. (Figure reproduced with permission by Sinauer Associates Inc.)

In contrast to the agonists, the antagonist bicucullin reduce the GABAA receptor function in a competitive way while blocks the receptor non- competitively (Bormann, 2000). The convulsant, t- butylbicyclophosphorothionate (TBPS), is thought to bind to the GABAA receptor in the closed state (Gee, 1988; Sieghart, 1992). As the TBPS binds to closed receptors, this binding is inhibited by GABAA receptor agonists such as GABA, muscimol, benzodiazepines and Allo. The positive GABAA receptor modulation by the neuroactive steroids shows marked stereoselectivity since the α-configuration of the hydroxyl group at C-3 of the steroid A ring is an important determinant of potency (Lambert et al., 1995). When the hydroxyl group at C-3 is in the β-configuration the steroids are suggested to be inactive on the GABAA receptor function, for example, the 3β-isomer of alphaxalone and Allo; betaxalone and 3β-OH-5α-pregnane-20-one respectively (Cottrell et al., 1987; Harrison and Simmonds, 1984; Purdy et al., 1990). However, in presence of Allo, 3β-OH-5α-pregnane-20-one has shown to act as an antagonist by reduce the Allo effect (Wang et al., 2001).

20 Subunits and distribution Receptor sensitivity for GABA and for positive or negative modulators is governed by the type and stoichiometry of its subunits (Follesa et al., 2004). Sixteen different subunits of the GABAA receptor have been identified: α1-6, β1-3, γ1-3, δ, ε, π and θ (Barnard, 2001; Korpi and Sinkkonen, 2006; Pirker et al., 2000). The second class of the ligand-gated ion channel, GABAC, is found in the retina and is made up of three additional subunits, ρ1-3 (for review: Enz and Cutting, 1998; Korpi and Sinkkonen, 2006). Despite the marked heterogenecity of the GABAA-receptor, only a limited number of these have been found (Sieghart, 1995; Sieghart et al., 1999; Wafford, 2005). It appears that at least one α, one β and one other such as γ, δ or ε are required to form a fully functional GABAA receptor (Chebib and Johnston, 1999). The subunit composition of the GABAA receptor is heterogeneous and region-specific throughout the CNS. The α1-subunit is the most prominent subunit in the rodent brain, often co- distributed with β2 (Wisden et al., 1992). The γ2 variant is in turn often co- localized with α1 and β2, an arrangement of subunits, α1β2γ2, which represents the most common subtype, making up about 43% of all GABAA receptors in the rat brain (McKernan et al., 2000; Pirker et al., 2000; Whiting, 2003; Wisden et al., 1992). Some of the subunits are more region-specific. The α6 subunit is expressed in cerebellar granule cells (Luddens and Korpi, 1995; Pirker et al., 2000; Wisden et al., 1992), while α2β1(β3) and α5β1(β3) are most abundant in hippocampus (Wisden et al., 1992). However, more than one GABAA receptor composition can be co-localized in a single neuron (Macdonald and Olsen, 1994).

Kinetic properties of the GABAA receptor

GABA binds to the GABAA receptor and thereby regulates the gating, i.e. the opening and closing, of the chloride ion channel (Macdonald and Olsen, 1994). The GABAA receptor has both high and low-affinity GABA binding sites (Macdonald and Olsen, 1994). At the high-affinity binding site nanomolar concentrations of GABA will activate the GABAA receptor, while the low- affinity binding site requires micromolar concentrations. GABAA receptor binding studies have shown that the receptor has distinct binding sites for GABA, benzodiazepines, barbiturates and picrotoxin (Chebib and Johnston, 1999; Macdonald and Olsen, 1994). Barbiturates, positive GABAA receptor modulators, typically exhibit a concentration-response pattern of enhancement of GABAA receptor function at nanomolar concentration but activation of the GABAA receptors at high (micromolar) concentrations (Ikeda et al., 1989; Wikinski et al., 2001). This pattern has also been observed for neuroactive steroids, suggests that some neuroactive steroids have two distinct binding sites on the GABAA receptor, one mediating enhancement and the other mediating activation (Belelli and Lambert, 2005). Recently, two distinct binding sites on

21 the GABAA receptor for Allo and THDOC have been found (Hosie et al., 2006). By comparing wild type and mutated GABAA receptors, it was further shown that Allo and 3α5α-THDOC bind to the same sites on the GABAA receptor (Hosie et al., 2006).

GABAA receptor function can be increased in several ways, by increasing channel conductance, opening frequencies, burst frequencies and burst duration (Macdonald and Olsen, 1994). The negative modulators reduce GABAA receptor function by similar actions (Macdonald and Olsen, 1994). To investigate, compare and to at least some extent understand the mechanisms of substances, different models that are more or less complex have been suggested (Jones and Westbrook, 1995; Macdonald and Olsen, 1994). The model suggests different states of the GABAA receptor. In the present thesis a simplified model that previously has been adapted from Jones and Westbrook 1995 (Haage et al., 2005) has been used to discuss the effect of the neuroactive steroids on the kinetic properties of the GABAA receptor (5). Unbound state (UB) and bound state (B) represent before and after GABA is bound to the receptor. When GABA is bound to the receptor, it can enter a closed, desensitized state (D) or an open state (O). Modulators of the GABAA receptor function affect the different kinetic parameters that drive the receptor to one state or another.

Figure 5. The simplified kinetic model used here, shows four different states of the GABAA receptor, unbound (UB) where no GABA is bound, bound (B) when GABA is bound and the two different states the receptor can enter desensitized (D) and open state (O). In the open state it is possible for chloride ions to flux over the membrane. (Adapted from Haage et al 2005)

Benzodiazepines are thought to enhance GABAA receptor current by increasing open and burst frequencies without altering channel conduction or open duration (for review: Macdonald and Olsen, 1994). This could be explained by an increased association rate or a decreased dissociation rate only at the step from unbounded to the binding site (Macdonald and Olsen, 1994) (Fig. 5). To get this response, benzodiazepines could also reduce the rate of entry into a desensitized state without altering the gating of the bound GABAA receptor channel (Macdonald and Olsen, 1994). Barbiturates act like benzodiazepines by enhancing GABAA receptor function and also by activating the GABAA receptor channel in the absence of GABA (Macdonald and Olsen, 1994). The enhancing

22 effect of barbiturates is thought to depend on altering the gating of the channel once GABA is bound (Macdonald and Olsen, 1994). On the other hand, some steroids seem to enhance GABAA receptor current by increasing the channel opening frequency and mean open time of the GABA-evoked openings (Macdonald and Olsen, 1994). The mechanism for the prolongation is similar to that described for barbiturates, suggesting that steroids and barbiturates may modulate the GABAA receptor through at least one common mechanism (Macdonald and Olsen, 1994). However, the potent positive GABAA Allo has been suggested to decrease the rate from bound state to unbound state (Haage et al., 2005) that is in line with the mechanism of benzodiazepines.

23 Aims The 3α-hydroxy-A-ring reduced pregnane steroids have been shown to positively modulate the GABAA receptor function while the 3β-steroids are inactive. Previously studies have further shown that the 3β-isomer of Allo can act as an antagonist in presence of Allo. Different hypothesis have been made if the effect of the 3β-steroids are dependent of the degree of GABAA receptor activity or selectively on the Allo enhancement. It was therefore of interest to further study the effect of the 3β-steroids on the GABAA receptor function.

The aims of the thesis were to:

ƒ Investigate how 3β-OH-5α-pregnane-20-one interacts with different

GABAA receptor agonists and to see if 3β-OH-5α-pregnane-20-one selectively reduces the Allopregnanolone enhanced GABA mediated chloride ion uptake

ƒ Investigate the effect of five different 3β-steroids on GABA and Allopregnanolone effect in different brain regions by both chloride ion uptake and by patch clamp technique

ƒ Study the effect of the 3β-steroid, 5β-pregnane-3β, 20(R)-diol and Pregnenolone sulfate of the kinetic parameters of the current response to GABA application and on sIPSCs both in the presence and in the absence of Allopregnanolone

ƒ Investigate the effects of glucocorticoid metabolites on the chloride ion uptake in presence of Allopregnanolone and/or of GABA and to study if 3β-OH-5α-pregnane-20-one interacts with the glucocorticoid metabolites. Further, to investigate if the efficacy and the potency of the 5β-isomer of THDOC are similar to THDOC and Allopregnanolone

24 Material and Methods In the present thesis sex and stress steroids were investigated, in vitro, on the GABAA receptor function in rat brain. To answer the raised questions, three different methods were used; the radio labeled TBPS binding assay, the radio labeled chloride ion uptake assay and the patch-clamp technique. Animals When the effect of sex and stress steroids is investigated it is important that the endogenous levels of these are as similar as possible between the rats. Therefore, in order to avoid fluctuating steroid levels due to the estrus cycle, non stressed male rats were used in all experiments.

Adult Wistar rats, 240 – 320 g, were used for the chloride ion uptake assay as well as for the TBPS binding study. Sprague-Dawleys rats 50 – 120 g were used for the patch-clamp experiments. The difference in rat strains was due to practical reason. Due to a higher level of neuronal survivor and a higher patching success rate usually comparatively young rats are used for patch-clamp experiments. All animals were housed and handled by same procedures. The experimental protocols were approved by the Animal Experiment Ethics Committee in Umeå. Substances Steroids are often low polar compounds and therefore have poor solubility in aqueous solutions. Because of that, solvents such as ethanol and Dimethyl sulfoxide (DMSO) are often needed. However, both ethanol and DMSO modulate the GABAA receptor function. Therefore, in the present thesis the following protocol was used for lipophilic steroid: They were first dissolved in 99.5 % ethanol followed by evaporation at the wells of 96-wells plates, for the chloride ion uptake assay, or in bottles, for the TBPS binding study and the patch-clamp experiments. The steroids were then dissolved in buffer on a shaking table or in the ultra sound bath. This procedure allows the steroids to display a better solubility in aqueous buffer. This may be a result of a less regular crystalline structure is formed, which allows these low polar steroids to dissolve in an aqueous media. [35S]TBPS binding assay

TBPS is a convulsant that inhibits the GABAA receptor function (Majewska, 1992). TBPS bind to receptors that are in the closed state. Positive GABAA receptor modulators inhibit the binding of the TBPS (Majewska, 1992). By the radiolabeled t-butylbicyclophosphorothionate, [35S]TBPS, binding assay, it is possible to investigate the effect of substances on the GABAA receptor function at a population of GABAA receptors in a specific brain region. The TBPS

25 binding study is a well established method to study the GABAA receptor function in a non-genomic situation. Since it is the binding of TBPS with a long incubation time and not the actual chloride ion flux that are studied, the variation between samples are low. The TBPS binding study was used in paper I to study the interaction between Allo and 3β-OH-5α-pregnane-20-one in rat cerebral cortex. The homogenate of the cerebral cortex was prepared by dissection and homogenization in PBS buffer, pH 7.2. The homogenate was then washed twice by centrifugation at 34000 x g for 30 minutes. After the washing step, the pellet was resuspended in PBS. The homogenate was then incubated in a TBPS mixture that also contained steroids and GABA. In this step the TBPS binds to GABAA receptors depending on how the steroids have modulated the receptor. After incubation, the mixture with the homogenate was filtrated through a glass fiber filter. The filter bound radioactivity was measured by liquid scintillation counting. Chloride ion uptake assay In contrast to the TBPS binding study, the chloride ion uptake assay makes it possible to measure the chloride ion influx into brain microsacs via the GABAA receptor. The homogenate use for the chloride ion uptake assay is a crude membrane fraction where microsacs are constructed by fusion of intact pieces of the cell membrane. This cell preparation, with an inactive nucleus, gives a non- genomic situation. In paper I, II and IV, the chloride ion uptake assay was used in different brain regions. The brain region of interest was homogenized in a standard extracellular buffer, pH 7.5. The homogenate was then gentle washed twice by centrifugation at 1000 x g for 15 minutes. The final pellet was dissolved in the buffer and allowed to acquire room temperature for minimum 10 minutes. Thereby the microsacs were achieved. Microsacs contain intact GABAA receptors and have the ability to preserve a (DeLorey and Brown, 1992; Macdonald and Olsen, 1994). Therefore, the response of GABA agonists or antagonists can be measured with radiotracer ion [36Cl-] influx or efflux. The chloride ion uptake assay is a well established method used to investigate negative and positive modulators on the GABAA receptor function. Our group has developed the system by using 96-well micriplates and the cell harvester system. In all chloride ion uptake experiments the 96-wells microplates were prepared with steroids of interest, GABA and the isotope [36Cl-]. The reaction was started by adding the homogenate to the prepared wells. GABA and the tested steroids were bounded to the GABAA receptors and allowed chloride ion flux into the microsacs. The reaction was terminated after about 4-5 seconds by suction of the content in the wells into a cell harvester. At the same time, the wells were washed with cold buffer containing picrotoxin, the GABAA receptor blocker. The microsacs were caught on glass fiber filters under vacuum. The filter-bound radioactivity could then be measured by liquid scintillation spectroscope. By using the cell harvester system

26 it was possible to receive up to 400 samples in one experiment depending of the volume of the homogenate. For smaller brain region such as amygdala and preoptic area, it should be necessary to pool these regions from many animals to get enough volume of the homogenate for one experiment. In the chloride ion uptake assay where the chloride ions flux into the microsacs in the range of seconds the variation between samples are higher compared to the TBPS binding study. In the present thesis each sample was therefore performed in three, four or nine replicates depending on the amount of tissue that was obtained from each individual. Further, the homogenate without adding any substance gives a quantifiable basal uptake. In each experiment a number of samples where homogenate with only isotope added were included. Data are expressed as the net uptake; the uptake involving the substances minus the basal uptake in the homogenate (without any substance added). Patch-clamp recording When neurons are acutely dissociated by using the vibro dissociating technique presynaptic terminals are often coisolated. Thereby, the spontaneous release of GABA from these presynaptic terminals which underlies spontaneous inhibitory postsynaptic currents (sIPSC) may be studied using the patch-clamp technique. In the patch-clamp technique the current or the potential over the postsynaptic cell membrane is measured. When the patch-clamp technique is performed under voltage clamp conditions the membrane potential is held at a fixed level by a feedback amplifier system. The measured parameter is the current needed to maintain this potential. In contrast to the chloride ion uptake assay where a bulk response from a large number of neurons is investigated the patch-clamp technique measure the electrical response from a single neuron or even a single channel. This gives the possibility to study cells from a very small and specific part of the brain. In paper II and III, the GABAA receptor function was investigated by the patch-clamp technique under voltage clamp condition on neurons from medial preoptic nucleus (MPN). A block of the rat brain containing the MPN was cut out and sliced by a vibroslicer. The neurons were then isolated by vibrodissociation with a thin glass rod (Vorobjev, 1991). By the use of mechanical dissociation, instead of enzymes, neurons on which responses from sIPSCs and externally applied GABA could be measured were obtained. In order to preserve the internal cellular components of the cell while allowing control over internal Na+, K+, Cl- concentrations the ampothericin B perforated patch technique was used (Horn and Marty, 1988). In the perforated patch technique the pipette containing intracellular solution is attached to the cell by gentle suction without disruption of the cellmembrane. The amphotericin B is then perforating the cell membrane by incorporation of small pores permeable only to monovalent ions. A steady holding potential of –17 mV was used in all experiments, after compensation for the liquid-junction potential (Neher, 1992).

27 Extracellular solutions with or without test substances were applied by a gravity- fed perfusion system controlled by solenoid valves. For the investigation of sIPSCs, the current was recorded in EC solution for 30- 60 s followed by a 60 – 300 s test period with 3β-steroids and/or Allo. To allow for near steady state conditions the first 30 s interval was not used in the analysis. When applying external GABA in absence or presence of steroids, the applications were 2 s long and given every 30 s with four replications. In contrast to the sIPSCs the current responses from the externally applied GABA varied between the neurons. Therefore all data from external applications are presented relative to a control. Statistical Analysis The concentration-response data in all papers were fitted to the sigmoid function using the Hill equation. In paper I, the curve fitting were done with Sigma Plot n n n Software, using the Hill equation as follows: E = E0 + Emax x C / (EC 50 + C ), where C is the concentration of test substance, EC50 is the concentration of test substance that produced half-maximum potentiation, n is the Hill coefficient, - - Emax is the maximum Cl uptake response and E0 is the estimated Cl uptake response in the absence of test substance. For paper II – IV, the data analysis of concentration-response data was fitted to the sigmoid concentration-response equation: Y = Minimal response + (Maximal response – Minimal response)/(1+10^((LogEC50–X)*Hill Slope)) with GraphPad Prism. In paper I and paper IV, analysis of variance (ANOVA) with repeated measures followed by post hoc least significant difference test (LSDT) was used to evaluate the Cl- uptake parameters. In paper II and III, ANOVA with the Kruskal Wallis non-parametric test was used to evaluate the Cl- uptake and the patch-clamp parameters. When the initial ANOVA test was significant, a post hoc test, the Mann Whitney test with Bonferroni adjustment, was performed. The non-parametric Mann-Whitney test with Bonferroni adjustment was used to compare the effect of single concentrations of substances with the control value. In all papers, all values in the text are displayed as mean ± SEM. The P-value of < 0.05 was used for significance. The SPSS statistical package version 10.0 was used for all ANOVA and Mann-Whitney tests.

28 Results and Discussion The allopregnanolone enhanced effect can be reduced by 3β-OH-5α-pregnane-20-one in a selective way (paper I) Allo and THDOC have been showed being extremely potent as sedative- agents by modulating the GABAA receptor function (Paul and Purdy, 1992). Furthermore, Allo has been associated with symptoms related to the menstrual cycle such as PMDD and PMS symptoms (Backstrom et al., 2003). Both Allo and THDOC are increased during stress (Barbaccia et al., 1998; Purdy et al., 1991; Serra et al., 2000). The pregnane steroids having 3β-configuration of the hydroxyl group of the C-3 of the steroid A-ring has been suggested to be inactive on the GABAA receptor function (Cottrell et al., 1987; Harrison and Simmonds, 1984; Purdy et al., 1990). However, in presence of Allo, the 3β- isomer 3β-OH-5α-pregnane-20-one has been shown to reduce the Allo effect (Wang et al., 2000). Therefore, in paper I, the effect of 3β-OH-5α-pregnane-20- one and the interaction with Allo were further investigated both by TBPS binding and by GABA mediated chloride ion uptake in rat cerebral cortex. By the chloride ion uptake assay, the effect of 3β-OH-5α-pregnane-20-one was also studied in presence of the common positive GABAA receptor function modulators; Allo, () and () as well as in presence of both high and low concentration of GABA.

The effect of 3β-OH-5α-pregnane-20-one on the GABA mediated allopregnanolone enhancement

GABA as well as GABA agonists such as muscimol stimulate the GABAA receptor function in a concentration dependent way, which has been shown by different methods (Hawkinson et al., 1994; Luu et al., 1987). In paper I, the chloride ion uptake assay was used to evaluate the EC50 value of GABA, 0 – 1000 µM, in rat cerebral cortex. By Hill equation a best fit curve from the concentration dependent increase was drawn and the EC50 value of 8.95 µM was calculated. To be able to measure both the increasing and the decreasing effects of GABA, 10 µM GABA was used for all further chloride ion uptake and TBPS binding studies in paper I. Neuroactive steroids have shown a correlation between the inhibition of TBPS and enhancement of GABA evoked current (Hawkinson et al., 1994). Allo is suggested to be particularly potent by enhancing the GABA response already at the concentration range of 10-30 nM (Callachan et al., 1987; Paul and Purdy, 1992). Allo has been shown to inhibit the TBPS binding by maximal 87 % and obtains an IC50 value of 37 nM (Hawkinson et al., 1994). Allo, in the concentration interval 0.1 – 1000 nM, was shown to reduced the TBPS binding

29 in a concentration dependent way (IC50 = 60 nM, paper I). When the interaction between Allo 0.1 – 1000 nM and 30 µM 3β-OH-5α-pregnane-20-one, was investigated no effect of 3β-OH-5α-pregnane-20-one was observed. Previous binding studies have shown that 3β-OH-5α-pregnane-20-one reduces the Allo enhancement of [3H]flunitrazepam binding (Prince and Simmonds, 1993). This indicates that the interaction can be measured both by chloride ion flux assay and by [3H]flunitrazepam binding studies but not with the TBPS binding study. Therefore, the interaction studies were investigated with the chloride ion uptake assay.

Allo enhanced the GABA mediated chloride ion uptake in a concentration response way that was shown in paper I. The maximal enhancement of Allo was almost 170 % of control and the EC50 value about 200 nM. When applied 3β-OH-5α-pregnane-20-one, 3 – 30 µM, in presence of different concentrations of Allo; 50 – 3000 nM the interaction could be measured by the chloride ion uptake (Fig. 6). 3β-OH-5α-pregnane-20-one reduced the Allo enhancement in a concentration dependent way (IC50 = 12.25 µM in presence of 1 µM Allo). Previously, 3β-OH-5α-pregnane-20-one has been suggested to act in a non-competitive way (Wang et al., 2002; Wang et al., 2000). This was in line with the results in paper I, where the 3β-OH-5α-pregnane-20-one reduced the efficacy of Allo without alter the concentration response curve of Allo. Maximal reducing effect of 133 % was seen by 30 µM 3β-OH-5α-pregnane-20- one on 1 µM Allo (Fig. 6). This gives a concentration ratio between :antagonist 1:30. This ratio was used for all further interaction studies throughout the thesis.

Figure 6. The GABA mediated Allo enhanced chloride ion uptake is reduced by 3β-OH- 5α-pregnane-20-one in a concentration dependent way. All values are presented as mean percentage ± SEM of control, the response of 10 µM GABA set as 100 %. Allo at the concentration of 3 µM, enhanced the GABA mediated chloride ion uptake by maximal 170 %. The maximal reducing effect of 3β-OH-5α-pregnane- 20-one was achieved at 30 µM on 1 µM Allo enhanced effect.

30 The effect of 3β-OH-5α-pregnane-20-one on GABA and on positive GABAA receptor modulators In paper I, it was shown that 3β-OH-5α-pregnane-20-one reduced the effect on Allo enhancement without shift the concentration response curve of Allo. The effect of the 3β-steroid was higher in high concentration of Allo, up to 1 µM, indicating that the reducing effect may be dependent of the degree of the GABAA receptor activation. However, Wang and coworkers (2000), showed that 3β-OH-5α-pregnane-20-one reduced the Allo enhancement of the muscimol effect, without interacting with muscimol in absence of Allo and therefore suggested that 3β-OH-5α-pregnane-20-one selectively reduced the Allo effect. The investigation of the interaction between 3β-OH-5α-pregnane-20-one and other positive GABAA receptor function modulators, flunitrazepam (benzodiazepine) and pentobarbital (barbiturate) was therefore proceeding. Benzodiazepines have been shown to exert a biphasic concentration dependent response on chloride ion flux (Ikeda et al., 1989; Schwartz et al., 1986; Wikinski et al., 2001). Benzodiazepines up to micromolar concentration increase the chloride ion uptake while at higher concentrations the ion flux was decreased (Wikinski et al., 2001). Both barbiturates and ethanol stimulate the chloride ion uptake in the same pattern as benzodiazepines but are less potent (Morrow and Paul, 1988). In paper I, it was shown that the benzodiazepine, flunitrazepam, 50 – 1000 nM, and the barbiturate, pentobarbital, 10 – 300 µM, increased the chloride ion flux in a concentration dependent way. Flunitrazepam was about 400 times as potent as pentobarbital (EC50 = 0.09 µM and EC50 = 41.66 µM respectively) but the efficacy was less compared to pentobarbital, maximal response of approximately 40 % and 130 % respectively. To investigate if 3β- OH-5α-pregnane-20-one had a more general or a selective reducing effect, the 3β-steroid in the concentration interval 1 – 30 µM was studied in presence of flunitrazepam 50 – 1000 nM or pentobarbital 10 – 300 µM as well as in presence of GABA, 1 – 1000 µM. It was shown that 3β-OH-5α-pregnane-20- one had no effect on flunitrazepam or pentobarbital enhanced or on GABA mediated chloride ion uptake (paper I, Fig. 7). By voltage clamp experiments in Xenopus oocytes, 3β-OH-5α-pregnane-20-one reduced the currents in response to 20 µM GABA (Wang et al., 2002). Further, the efficiency of 3β-steroids has been suggested being correlated with the degree of GABAA receptor (Eisenman et al., 2003; Wang et al., 2002). 3β-OH-5α-pregnane-20-one has been shown to be inactive in absence of GABA, both in [3H]flunitrazepam binding studies (Prince and Simmonds, 1992) and in population spike studies (Wang et al., 2000). The inactive effect of 3β-OH-5α-pregnane-20-one on the GABAA receptor function in absence of GABA was also shown in paper I by the chloride ion uptake assay. As a conclusion, the results in paper I indicate that 3β-OH-5α-pregnane-20-one selectively reduce the Allo enhanced GABA mediated chloride ion uptake in a non-competitive way.

31

Figure 7. Flunitrazepam and pentobarbital enhanced the GABA mediated chloride ion uptake in a concentration dependent way. All values are presented as mean percentage ± SEM of control, the response of 10 µM GABA set as 100 %. 3β-OH-5α-pregnane-20-one (1-30 µM) had no effect either on the flunitrazepam enhancement or the pentobarbital enhancement.

3β-steroids modulate both the GABA effect and the allopregnanolone enhancement on the GABAA receptor (Paper II)

The effect of 3β-steroids on the GABAA receptor function has been investigated in several studies with diverse results (Le Foll et al., 1997; Peters et al., 1988; Shen et al., 2000; Wang et al., 2002; Wang et al., 2000). 3β-steroids have been suggested to selectively reduce the Allo effect (Wang et al., 2000) as well as suggested being dependent on the degree of GABAA receptor activity (Eisenman et al., 2003; Wang et al., 2002). 3β-OH-5α-pregnane-20-one has previously been shown to reduce the Allo enhanced GABA mediated chloride ion uptake in a selectively way, without any effect on other positive GABAA receptor modulators or on GABA alone (paper I). Except for 3β-OH-5α-pregnane-20- one, which is quite well studied (Prince and Simmonds, 1993; Wang et al., 2002; Wang et al., 2000), 5β-pregnane-3β, 21-diol-20-one (3β5β-THDOC) and 3β-OH-5β-pregnane-20-one have been shown to reduce the Allo enhanced effect in a non-competitive manner by voltage clamp studied in Xenopus oocyte (Wang et al., 2002). 3β-OH-5β-pregnane-20-one has also been investigated by the [3H]flunitrazepam binding, where the 3β-steroid reduced the Allo enhanced effect on the [3H]flunitrazepam binding (Prince and Simmonds, 1993) without

32 having any effect in absence of Allo (Turner et al., 1989). Both in vivo and in vitro studies of 5α-pregnane-3β, 20(R)-diol showed that the 3β-steroid reduced the Allo enhanced GABA mediated chloride ion uptake and the learning impairment induced in rats by Allo (Turkmen et al., 2004). It was therefore of interest to further study different 3β- steroids on the GABA mediated GABAA receptor function in absence and presence of Allo. The chloride ion uptake assay and the patch-clamp technique were used in paper II to further study the effect of five different 3β-hydroxy steroids on the GABA effect both in absence and in presence of Allo in different brain regions. In addition to the four 3β-steroids that earlier have been studied, 3β-OH-5α-pregnane-20-one; 5α-pregnane-3β, 20(R)-diol; 3β-OH-5β-pregnane-20-one and 3β5β-THDOC, the 3β5β-steroid 5β-pregnane-3β, 20(S)-diol where also investigated. Since the subunit composition of the GABAA receptor is region-specific distributed throughout the CNS (for review: Korpi et al., 2002; Korpi and Sinkkonen, 2006; Pirker et al., 2000; Wisden et al., 1992) and the efficiency of GABA and neuroactive steroids differ between the regions (Luu et al., 1987; Wilson and Biscardi, 1997), it was of interest to study if the effect of 3β-steroids varied between brain regions. The chloride ion uptake was used to study the effects in cerebral cortex, hippocampus and cerebellum and the patch-clamp technique to study sIPSCs from MPN neurons. The MPN was not studied with the chloride ion uptake assay since it is a small region and therefore should require many animals to achieve volume of homogenate enough to one experiment. GABA and allopregnanolone sensitivity differ between brain regions

The different GABAA receptor subunits can be widespread presented in most brain areas or be more region specific. The α1β2γ2 subunit composition is present in most brain regions and common in both cerebral cortex and hippocampus while the α5β3γ2/γ3 is more region specific and most prominent in hippocampus (McKernan et al., 2000). The receptor composition α6β2γ2 is also a region specific subunit composition, localized in cerebellum (McKernan et al., 2000). The different receptor subunit compositions have been suggested to correlate with diverse properties depending on localization (for review: Korpi and Sinkkonen, 2006). It is therefore most likely that the response to GABA and neuroactive steroids differ between the regions. In paper II the chloride ion uptake of 1 – 1000 µM GABA stimulation was investigated in microsac preparations from rat cerebral cortex, hippocampus and cerebellum. Both the efficacy and the potency of GABA was similar in cerebral cortex and hippocampus where the maximal chloride ion uptake was about 250 pmol/mg protein and the EC50 values were 21 µM and 24 µM respectively. In cerebellum on the other hand the efficacy was lower, the maximal response reached approximately 100 pmol/mg protein. Similar pattern where shown in muscimol

33 stimulated chloride ion uptake (Luu et al., 1987). Further, in cerebellum the potency of GABA was about three times higher than in cerebral cortex and hippocampus (EC50 = 8 µM). This reflecting the molecular heterogeneity of the GABAA receptor that has also been shown by variation of the TBPS binding between different brain regions (Luddens and Korpi, 1995). In studies where human embryonic kidney cells (HEK-293 cells) were transfected with GABAA receptor α6β2γ2 or α1β2γ2 subtype, it was observed that the α6β2γ2 subtype was more sensitive to GABA and Allo than α1β2γ2 subtype (Korpi and Luddens, 1993). This indicates that the sensitivity for both GABA and Allo is higher in cerebellum than in cerebral cortex and hippocampus and in line with the results in paper II.

Due to the findings that GABA was more potent in cerebellum than in cerebral cortex and hippocampus, the EC20 value for each region was used as the GABAA receptor activator in further interaction studies. The EC20 for GABA in cerebellum was about three times lower (3 µM) than in cerebral cortex and hippocampus (10 µM). Allo in the concentration interval 0.1 – 3 µM for studies in cerebral cortex and hippocampus and 0.03 – 3 µM in cerebellum showed a concentration dependent enhancement of the GABA response. The Allo enhanced effect in different brain region showed the same pattern as for GABA response. In cerebral cortex and hippocampus the Allo enhancement was similar both in efficacy and in potency (Emax = 307 % and 273 % respectively, EC50 = 0.46 µM and 0.41 µM respectively). In cerebellum on the other hand, the effect of Allo reached lower maximal effect (Emax = 156 %) but the potency was higher than in cerebral cortex and hippocampus (EC50 = 0.2 µM). Differences between brain regions have been shown earlier by Wilson and Biscardi 1997. However, in their work, Wilson and Biscardi showed a significantly higher efficacy of both Allo and THDOC in hippocampus compared to cerebral cortex. In cerebellum on the other hand Wilson and Biscardi found that the efficacy of Allo and THDOC was lower compared to both cerebral cortex and hippocampus which was in line with the result in paper II. The difference in the results in cerebral cortex and hippocampus can be due to e.g. different rat strain, differences in the method or in the statistic analysis. In paper II, the non- parametric Mann-Whitney test with Bonferroni adjustment was used. This method is restrictive where the differences needed to be obvious without large variation to get significance.

The effect of 3β-steroids in different brain regions Previously, in paper I, it was found that the maximal reducing effect of 3β-OH- 5α-pregnane-20-one on Allo enhanced GABA mediated chloride ion uptake was in the relationship of [Allo]:[3β-steroid] 1:30. This relationship where therefore used for the interaction studies in paper II. Since the efficacy as well as the potency of Allo differed between the brain regions, the concentration of Allo

34 that gave the EC75 value was used. In cerebral cortex and hippocampus Allo EC75 were 1 µM while in cerebellum the EC75 was 0.3 µM. By the patch-clamp technique it has previously been shown that Allo at the concentration of 100 nM strongly prolongs the time course of the deactivation of sIPSCs (Haage et al., 2005; Haage and Johansson, 1999). Therefore, 100 nM Allo was used in the patch-clamp experiments in paper II. The relationship between the concentration of Allo and the concentration of the 3β-steroids were in the patch-clamp studies as for the chloride ion uptake 1:30. In paper II it was found that the five 3β- steroids reduced the Allo enhanced GABA mediated chloride ion uptake into both cerebral cortex and hippocampus. However the efficacy differed between the 3β-steroids but not for single 3β-steroid between cerebral cortex and hippocampus. 3β-OH-5β-pregnane-20-one was significantly less efficient in both cerebral cortex and hippocampus compared to 3β-OH-5α-pregnane-20-one, 5β-pregnane-3β, 20 (S)-diol and 3β5β-THDOC. The difference between the less effective and the most effective 3β-steroid was about 30 % where 3β-OH-5β- pregnane-20-one reduced the Allo effect by about 10 % while 3β-OH-5β- pregnane-20-one reduced the Allo effect in cerebral cortex and in hippocampus with approximately 40 % (Fig. 8A). 5α-pregnane-3β, 20(R)-diol has previously been shown to reduce the Allo enhanced GABA mediated chloride ion uptake in both cerebral cortex and in hippocampus where the effect appeared to be both more efficient and more potent in hippocampus than in cerebral cortex (Turkmen et al., 2004). The same pattern was shown in paper II where 5α- pregnane-3β, 20(R)-diol reduced the Allo enhancement with approximately 20 % in cerebral cortex and 30 % in hippocampus. On the other hand, in cerebellum there was only weak effect of 3β-OH-5α-pregnane-20-one while all other 3β- steroids tested did not modulate the GABAA receptor function (Fig. 8A). By the patch-clamp technique the time course of the deactivation of sIPSCs was measured in control condition (about 33 ms) and in presence of Allo (about 200 ms). The 3β-steroids showed a general reducing effect, about 25 %, of the Allo prolonged sIPSCs τdeactivation in MPN (Fig. 8B). This indicates that the effect of the 3β-steroids as for GABA and positive GABAA receptor modulators is more or less selective for different GABAA receptor subtypes at least subtypes typical for cerebellum.

The effect of the 3β-steroids was also investigated in absence of Allo on GABA mediated chloride ion flux into cortical microsacs and on sIPSCs currents from MPN neurons. Surprisingly, 5β-pregnane-3β, 20(S)-diol and 3β-OH-5β- pregnane-20-one enhanced the GABA (EC20) effect as well as prolonged the τdeactivation of sIPSCs currents (Fig. 9). This enhanced effect has previously been shown for 3β-OH-5β-pregnane-20-one in both the Xenopus oocyte (Wang et al., 2002) and in cultured bovine adrenomedullary chromaffin cells (Peters et al., 1988).

35

Figure 8. The regional variance in the effect of 3β-steroids on GABA mediated Allo enhancement. (A) 30µM 3β-steroids reduced 1µM Allo induced enhancement of GABA mediated chloride ion uptake in a similar way in cerebral cortex and in hippocampus. In cerebellum only one out of five 3β-steroids (9 µM) reduced the Allo (300 nM) enhancement. All values are presented as mean percentage ± SEM of control, response of 1µM Allo in presence of 10 µM GABA set as 100 %. (B) In MPN there was a general reducing effect of the 3β-steroids (3 µM) on the Allo induced prolongation of GABA mediated sIPSCs τdeactivation. All values are presented as mean percentage ± SEM of control, response of sIPSCs τdeactivation in presence of 100 nM Allo set as 100 %.

Since 5β-pregnane-3β, 20(S)-diol and 3β-OH-5β-pregnane-20-one have dualistic properties depending on Allo is present or not it is indicated that the 3β-steroids may have different mechanism. 5β-pregnane-3β, 20(S)-diol and 3β- OH-5β-pregnane-20-one have the 5β-configuration in common. However, the dualistic properties cannot only be due to the 5β-configuration since 3β5β- THDOC not enhances the GABA effect.

36

Figure 9. In absence of Allo, two of the 3β-steroids have an enhancing effect on the (A) GABA mediated chloride ion uptake as well as (B) prolonged the sIPSCs τdeactivation. All values are presented as mean ± SEM % of control, chloride ion uptake in presence of 10 µM or sIPSCs τdeactivation under control condition, set as 100 %.

The effect of 5β-pregnane-3β, 20(R)-diol on the kinetic properties of the GABAA receptor (Paper III) The mechanism behind the effect of 3β-hydroxy steroids is not completely clear. In the literature two main hypothesis of the mechanism of the 3β-steroids are described. 3β-steroids are suggested to act either similar to pregnenolone sulfate (PS) and therefore its effect is determined by the degree of GABAA-receptor activity (Wang et al., 2002), or specifically against Allo (Wang et al., 2000). The results in paper I and paper II have been in line with the hypothesis that there is an Allo specific mechanism with the addition that some 3β-steroids, in the absence of Allo, actually appeared to act as positive GABAA-receptor modulators. No results indicated that 3β-steroids act as negative modulators if Allo is absent. In paper III, this question was further addressed by investigating the kinetic changes induced by 5β-pregnane-3β, 20(R)-diol at the GABAA- receptor. This was studied by patch-clamp experiments carried out on neurons from medial preoptic nucleus (MPN) of male rat.

37 5β-pregnane-3β, 20(R)-diol reduce the allopregnanolone prolongation of sIPSCs τdeactivation Previously studies have shown that Allo prolongs the time course of the sIPSCs deactivation at MPN neurons without altering the amplitude (Haage et al., 2002). In paper II it was shown that 3β-steroids reduce the effect of Allo on sIPSCs but two of the 3β-steroid tested also gave a prolongation of the sIPSCs decay time in absence of Allo. In line with the previous study, in paper III, it was shown that 5β-pregnane-3β, 20(R)-diol reduced the Allo induced prolongation of sIPSCs τdeactivation by approximately 25 % but had no effect in absence of Allo (Fig. 10). Based on the results from sIPSCs recordings in paper II and paper III, it appears that 3β-OH steroids do not have an inhibitory effect by them self but instead selectively reduces the effect of Allo.

Figure 10. Typical trace of sIPSCs under control condition (Ctrl) and in presence of 5β-pregnane-3β, 20(R)-diol (3 µM), Allo (100 nM) and 5β-pregnane-3β, 20(R)-diol in presence of Allo. The 3β-steroid reduced the Allo induced sIPSCs τdeactivation without any effect in the absence of Allo.

The effect of 5β-pregnane-3β, 20(R)-diol on external application of GABA For synaptically released GABA the exposure time is considered to be less than 5 ms (Jones and Westbrook, 1995; Mody and Pearce, 2004). Because of that no desensitization phase of the current response is usually detectable. However, the kinetics of desensitization is suggested to have a large impact on the time course of sIPSCs (Jones and Westbrook, 1995; Zhu and Vicini, 1997). Therefore, in order to study the effect of 5β-pregnane-3β, 20(R)-diol on the kinetic properties of the GABAA receptor in detail, in paper III, an experimental protocol in which GABA was externally applied for 2 s was used. An advantage of this is that the desensitization kinetics can easily be detected. The effect on different parameters were investigated and compared to the effect of PS. PS was used as a reference due to its well characterized kinetic effect to increase the

38 desensitization rate and because 3β-OH steroids have been suggested to share that property with PS (Wang et al., 2002). The different parameters investigated were; peak amplitude, the time course of deactivation τdeactivation and the desensitization. To describe the desensitization, two parameters were used, the delta amplitude (end-to peak amplitude ratio) and t50, the desensitization rate measures as the time elapsed from the peak amplitude to the current had decline 50 % of the amplitude between the peak amplitude and the end amplitude. In the concentration interval 0.3 – 3 µM, neither 5β-pregnane-3β, 20(R)-diol nor PS modulated the peak amplitude evoked by 2 s long 100 µM GABA applications (Fig. 11). At high steroid concentrations, 10 µM, both 5β-pregnane-3β, 20(R)- diol and PS decreased the peak amplitude. However, the decrease in the peak amplitude did neither differ between applications of 10 µM and 100 µM GABA nor between the two steroids.

Figure. 11. The peak amplitude evoked by 100 µM GABA applications was not altered by either 5β-pregnane-3β, 20(R)-diol or PS in the concentration interval 0.3 – 3 µM. All values are presented as mean ± SEM % of control, the peak amplitud of 100 µM GABA application, set as 100 %.

Both 5β-pregnane-3β, 20(R)-diol and PS 0.3 – 10 µM increased the desensitization rate of 100 µM GABA application in a concentration dependent way, a result that it is not in accordance with the results obtained in paper II, where no inhibitory effect was seen in the absence of Allo. The results from the GABA applications showed that the effect on the t50 was almost identical for 5β- pregnane-3β, 20 (R)-diol and PS both in efficacy and in potency (Emax approximately 30 % and 20 % respectively; IC50 value: 2.61 µM and 2.35 µM respectively, Fig. 12A). However, the effect of 5β-pregnane-3β, 20(R)-diol and PS on the delta amplitude was different. Both 5β-pregnane-3β, 20(R)-diol and PS decreased the delta amplitude in a concentration dependent way but the effect differed both in efficacy and in potency (Fig. 12B). 5β-pregnane-3β, 20(R)-diol at 10 µM decreased the delta amplitude by about 60 % while the delta amplitude at 10 µM PS was almost zero. Further, the potency of PS was 7 times higher than for 5β-pregnane-3β, 20(R)-diol (IC50 value: 0.5 µM and 3.48

39 µM respectively). These results show that the 3β-steroid increases the desensitization rate but do so differently to PS. But most important, the results show that also in the absence of Allo at least one 3β-steroid may clearly exert an inhibitory effect.

Figure 12. 5β-pregnane-3β, 20(R)-diol and PS in the concentration interval 0.3 – 10 µM increase the desensitization rate of 100 µM GABA application by (A) decrease the t50 almost identical. (B) The delta amplitude was also decreased by both tested steroids in a concentration dependent way but PS was more efficient and more potent. All values are presented as mean ± SEM % of control, the response of 100 µM GABA application, set as 100 %.

In similarity to Allo, although to less extent 5β-pregnane-3β, 20(R)-diol increased the τdeactivation while PS had the opposite effect (paper III, Fig. 13). The effect by 5β-pregnane-3β, 20(R)-diol on the τdeactivation seems to be activity dependent. The effect is more prominent at 100 μM GABA than 10 μM GABA.

An increase in τdeactivation as a result of an accumulation of receptors at desensitized states has previously been described (Jones and Westbrook, 1995; Zhu and Vicini, 1997). Therefore, the observation that 5β-pregnane-3β, 20(R)- diol did not alter the τdeactivation of sIPSCs but prolonged the τdeactivation of externally applied of GABA indicate that the modulation is through a desensitization dependent mechanism. This is different from Allo which has been suggested by Haage and coworkers, 2005, to decrease the GABA unbinding rate and therefore affects sIPSCs and longer GABA applications equally (Fig. 5).

40 Figure 13. 5β-pregnane-3β, 20(R)-diol and PS had opposite effect on the τdeactivation of the 100 µM GABA applications. 5β- pregnane-3β, 20(R)-diol increased the τdeactivation while PS decreased it. At higher concentration of PS the τdeactivation was undetectable. All values are presented as mean ± SEM % of control, the τdeactivation of 100 µM GABA application, set as 100 %.

The effect of 5β-pregnane-3β, 20(R)-diol in presence of allopregnanolone on GABA application Since the results in paper I and paper II are in line with the hypothesis that 3β- steroids act selectively against Allo it was of interest to further investigate the effect of 5β-pregnane-3β, 20(R)-diol on GABA applications in the presence of Allo. In order with previous report (Haage and Johansson, 1999), on 100 µM GABA application, Allo strongly prolonged the τdeactivation and decreased the peak amplitude while the desensitization rate was not modulated (paper III, Fig. 14). However, in paper III, it was shown that the effect of 5β-pregnane-3β, 20(R)-diol and PS on GABA applications was not altered in presence of Allo. The effect of these two substances was almost identical to their action in the absence of Allo. This indicates that there is no interaction between 5β-pregnane- 3β, 20(R)-diol and Allo on 2 s GABA applications. Interestingly, 5β-pregnane- 3β, 20(R)-diol did not reduce the prolonged deactivation time as it did for sIPSCs. This adds further pieces of evidence to idea the effect of 5β-pregnane- 3β, 20(R)-diol on the deactivation rate is dependent on the GABAA-receptor activity and does not interfere with the GABA unbinding rate as suggested for Allo (Fig. 5).

41 A

Figure 14. Typical trace for 100 µM GABA application in presence of tested steroids. (A) 10 µM 5β- B pregnane-3β, 20(R)-diol and 10 µM PS strongly reduce the deactivation rate. (B) 100 nM Allo prolongs the τdeactivation of 100 µM GABA applications. However, in presence of Allo the effect of 5β-pregnane-3β, 20(R)- diol and PS was not different from that seen with 100 µM GABA only. All values are presented as mean ± SEM % of control, set as 100 %.

Glucocorticoid metabolites modulate the GABAA receptor function (Paper IV) The 3α-hydroxy configuration of reduced A-ring pregnane steroids has shown being important for the potent positive modulation of the GABAA receptor function (Paul and Purdy, 1992). The 5α-configuration on the other hand, does not seem to be as important. For example, there is no difference in effect between Allo and the 5β-isomer of Allo, 3α-OH-5β-pregnane-20-one (Paul and Purdy, 1992). However, the 5β-isomer of THDOC, 3α, 21-dihydroxy-5α- pregnane-20-one is less efficient than 5α-THDOC (Paul and Purdy, 1992). In paper IV, the efficacy as well as the potency of Allo, THDOC and 3α5β- THDOC was compared by the chloride ion uptake into rat cortical microsacs.

42 Allo and THDOC are, like glucocorticoids, secreted by the adrenal cortex after HPA-axis activation by acute stress stimuli (Purdy et al., 1991). As a response to acute stress, the level of both Allo and THDOC are also increased in the brain (Purdy et al., 1991). In human, about 50 % of the cortisol is metabolized to Allo – THF, THF and THE (Fukushima et al 1960), by the 5α/βR-3αHSD enzymatic pathway (Tsilchorozidou et al., 2003). The 5αR – 3αHSD enzymatic pathway is also involved in the conversion of PROG to Allo and in conversion of DOC to THDOC (Melcangi et al., 1999) see also Fig. 1. Thereby the glucocorticoid metabolites are structurally similar to the potent positive modulators of the GABAA receptor. In addition to the structurally similarities, the glucocorticoids have been shown to act in a non-genomic way (Borski, 2000; Makara and Haller, 2001; Wheling, 1997). In paper IV, it was investigated if the glucocorticoid metabolites modulate the GABAA receptor function and if the metabolites do interact with the GABAA receptor agonist allopregnanolone. The 5β-isomer of THDOC is not as efficient as the 5α-isomer

Steroids with the 5β-pregnane- structure are suggested to modulate the GABAA receptor in some cases as efficient and in some with less efficacy than the 5α- compounds (Paul and Purdy, 1992). By the GABA mediated chloride ion uptake assay, in paper IV, the 5β-isomer of THDOC was compared with Allo and THDOC. 3α5α-THDOC, Allo and 3α5β-THDOC in the concentration interval 0.1 – 10 µM was studied in presence of 10 µM GABA. Previously, Allo and THDOC have been showed to increase the chloride ion uptake by approximately 200 – 300 % while the 3α5β-THDOC is about 60 % less efficient (Paul and Purdy, 1992). 3α5α-THDOC and Allo was shown in paper IV to modulate the GABAA receptor almost identical by enhancing the GABA mediated chloride ion uptake by approximately 130 %. The 3α5β-THDOC increase the GABA effect but the efficacy was not as prominent as for the 5α-compounds, by an increase of approximately 66 %. However, the potency was almost equal for

Allo, 3α5α-THDOC and 3α5β-THDOC (EC50 values: 0.17 µM; 0.2 µM and 0.24 µM respectively). The 5β-configuration seems therefore to be involved in at least the efficacy of the steroid effect on GABAA receptor function but not in the potency of the DOC metabolites. Allo and THDOC that have the planar configuration with α-configuration share the same binding site on the GABAA receptor. The 5β-configuration change the distance between the 3α-hydroxy group of the A-ring and the group in 20 position and form a chair formed substance (Purdy et al., 1991). This may change the binding to the GABAA receptor and thereby make difference in efficacy.

43 3α5α/5β-cortisol and 3α5α/5β-cortisone reduce the GABA mediated chloride ion uptake

Previously, exposure to acute stress have been shown to modulate the GABAA receptor function by increasing the TBPS binding (Concas et al., 1988; Foddi et al., 1997) and decrease the [3H]-GABA binding (Concas et al., 1985). Acute stress also decrease the muscimol stimulated chloride ion flux into rat cerebral cortex by about 25 % (Drugan et al., 1989). In vitro experiments have shown that cortisol and the cortisol metabolites Allo – THF and THF at high concentrations (3 – 100 µM) reduce the muscimol stimulated chloride ion flux in rat cerebral cortex (Penland and Morrow, 2004). Since the maximal glucocorticoid concentration in brain is suggested to be about 1 µM, it was of interest to study the metabolites at lower concentrations (Makara and Haller, 2001). In paper IV, the glucocorticoid metabolites Allo – THF, THF and THE as well as 3α-OH, 5α-cortisone were investigated in the concentration interval of 0.1 – 3 µM, on GABA mediated chloride ion flux. In normal physiological conditions, the level of cortisol in human brain is 70 ng/g (Carroll et al 1975). The corticosterone level in plasma increase from about 70 ng/ml in naive rats to approximately 400 ng/ml immediately after foot shock, equivalent with 200 nM in normal state and 1.2 µM after stress (Barbaccia et al., 1997). In paper IV the glucocorticoid metabolites were thereby tested within physiological levels. In line with previous studies all glucocorticoid metabolites reduced the GABA effect in a concentration response way with a maximal response of about 40 - 50 % (Fig. 15). Muscimol stimulated chloride ion flux was in a previous study shown to be completely blocked by 100 µM THF but did not modulate the ion flux at the concentration of 1 µM (Penland and Morrow, 2004). Cortisol and Allo – THF showed the same pattern by reducing the chloride ion flux at 10 µM but not at 1 µM (Penland and Morrow, 2004). The differences between the result that was found in paper IV and the results found by Penland and Morrow (2004), on muscimol stimulated chloride ion uptake can be due to the method. The glucocorticoid metabolites are difficult to solve in buffer. Therefore in paper IV, they were first dissolved in ethanol, evaporated to dryness in a bottle with wide bottom. Then they were dissolved in buffer with ultrasound and thereby obtain a solution free from solvents such as DMSO and ethanol. In the experiment with muscimol stimulated chloride ion uptake, the procedure of the glucocorticoids are not explained (Penland and Morrow, 2004). However, in both studies the glucocorticoid metabolites had no effect on the function in absence of GABA or muscimol respectively (Penland and Morrow, 2004), paper IV). It has been shown that the corticosterone level in rats increase immediately after exposure to acute stress (Barbaccia et al., 1997). Since it was suggested in paper IV that Allo – THF and THF significantly reduced the chloride ion uptake already at 100 nM, the results indicate that the glucocorticoid secreted in acute stress reach levels that can modulate the GABAA receptor function. Previously, it has been shown that when adrenalectomy rats were exposed to foot shock

44 there were no increase in corticosterone level, indicating that even the immediately secretion of corticosterone is a response of the activation of the HPA-axis (Barbaccia et al., 1997).

Figure 15. Both the cortisol metabolites (A), and the cortisone metabolites (B), in the concentration interval 0.1 – 3 µM, reduced the 10 µM GABA mediated chloride ion uptake in a concentration dependent way. All values are presented as mean ± SEM % of contro, the response of 10 µM GABA, set as 100 %. The glucocorticoid metabolites interact with allopregnanolone In acute stress, not only glucocorticoids but also Allo and THDOC are released with the maximal level reached 10 – 20 minutes after the stress stimuli (Purdy et al., 1991; Reddy, 2003; Serra et al., 2000). The increase of Allo after acute stress has been suggested to counteract the decrease in GABAergic function (Barbaccia et al., 1998). Therefore, it was of interest to further investigate if the glucocorticoid metabolites interact with Allo on the GABAA receptor function. Previously, cortisol and the metabolites Allo – THF and THF at 10 µM are suggested to reduce the THDOC enhanced effect while at 1 µM had no effect on the THDOC enhancement (Penland and Morrow, 2004). In opposite, in paper IV, an interaction between glucocorticoid metabolites and Allo was shown. The GABA reduced effect shown by Allo – THF and 3α - OH, 5α-corticosterone at 0.3 – 3 µM, was completely blocked in presence of 1 µM Allo (Fig. 16). Further, Allo – THF, but not 3α-OH, 5α-cortisone slightly enhanced the Allo enhanced effect (Fig 16). Similar mechanism have been shown in rats exposed to acute stress, where the increased effect by glucocorticoids on the TBPS binding was blocked by injection of the positive GABAA receptor modulator (Foddi et al., 1997). 3β-OH-5α-pregnane-20-one has been shown, in paper I and II, to selectively reduce the Allo enhanced effect. Therefore, in paper IV, it was further investigated if 3β-OH-5α-pregnane-20-one could reduce the Allo enhanced effect. 3β-OH-5α-pregnane-20-one at high concentration, 30 µM, was shown to completely block the Allo – THF effect on the Allo enhancement (Fig. 16A) indicating that the enhancement is due to Allo. Even though, the Allo enhanced effect was blocked, it was not possible to observe the

45 reducing effect of GABA by the Allo – THF. This indicates that the reducing effect on the chloride ion uptake by glucocorticoid metabolites cannot be detected in presence of Allo. However, the interaction between 3β-OH-5α- pregnane-20-one and Allo-THF as negative GABAA receptor modulator is not studied, this effect may also depend on the effect of 3β-OH-5α-pregnane-20- one.

Figure 16. The reducing effect of cortisol and cortisone metabolites on GABA mediated chloride ion uptake was blocked in presence of 1 µM Allo. (A) The cortisol metabolite enhanced the Allo effect. The enhanced effect was completely blocked by 3β-OH-5α-pregnane-20-one. All values are presented as mean ± SEM % of control, the response of 1µM Allo in presence of 10 µM GABA and 3β-OH-5α-pregnane-20-one 3, 10 or 30 µM respectively, set as 100 %. (B) In presence of the cortisone metabolite and Allo no enhancement was seen. All values are presented as mean ± SEM % of control, the response of 1µM Allo in presence of 10 µM GABA, set as 100 %.

46 Summary and concluding remarks One part of the present thesis was about the effect of the 3β-isomer of Allo, 3β- OH-5α-pregnane-20-one. Based on the obtained results it was shown that 3β- OH-5α-pregnane-20-one counteracted the effect of Allo in chloride ion uptake experiments as well as patch-clamp recordings of sIPSCs. In the chloride ion uptake assay, the 3β-isomer reduced the efficacy of Allo without shift the concentration response curve. Further, the 3β-isomer did not interact with other positive modulators or with GABA itself. These results indicate that the 3β- isomer act in a selectively non-competitive way on the Allo enhanced effect.

Another question in this thesis was whether a difference in the effect of 3β- steroids existed between brain regions or not. Five different 3β-steroids, including the 3β-isomer of Allo, reduced the Allo enhanced GABA effect in cerebral cortex, hippocampus and preoptic nucleus. On the other hand, in cerebellum the effect of the 3β-steroids was weaker compared to the other tested brain regions. From these results it was concluded that there is indeed a regional difference. This regional difference was also found for GABA and Allo, where the effect was more potent in cerebellum than in cerebral cortex and hippocampus. Interestingly, in absence of Allo, two of the 3β-steroids positively modulated the GABA stimulated GABAA receptor function. This indicates that the α-configuration is not crucial for the positive modulation of the GABAA receptor. Further, it suggests that the 3β-steroids have multiple kinetic properties and that they may also have an effect by them self. The results from the two first papers mainly support the hypothesis that the 3β-steroids act selectively against Allo (Wang et al., 2000). However, the finding that some 3β-steroids actually can induce a GABAA-receptor potentiation them self, together with the suggested hypothesis that 3β-steroids are activity dependent GABAA receptor antagonists (Wang et al., 2002) challenge this picture. Therefore, to achieve a better understanding for the action of the 3β-steroids in detail, the effect of 5β- pregnane-3β, 20(R)-diol, was investigated on the kinetic parameters of the current response to GABA applications in absence or presence of Allo (paper III). Those recordings clearly showed that 5β-pregnane-3β, 20(R)-diol increased the desensitization rate of GABA evoked currents in a concentration dependent way also in absence of Allo. Additionally, this effect was also seen for a number of other 3β-steroids (unpublished data). However, when the 5β-pregnane-3β, 20(R)-diol was tested on sIPSCs it did not show any effect by itself but reduced the effect of Allo. Obviously, there is a discrepancy between the results from applied GABA and those from sIPSCs and chloride uptake experiments. It might be that an increase in the desensitization rate is the main effect of many 3β- steroids but the effect is more visible at high levels of receptor activity such as in the presence of positive GABAA-receptor modulators like Allo. Furthermore,

47 an alteration of the kinetics of desensitization may explain the decrease in the deactivation rate (Zhu and Vicini, 1997) which was seen for some of the tested 3β-steroids. On the other hand, the effect of 3β-steroids on Allo could not be achieved by other means to increase the GABAA-receptor activity such as, high concentration of GABA, barbiturates or benzodiazepines.

Another observation that supports the idea that 3β-steroids alter the kinetics of desensitization is the interaction study between Allo and the 3β-isomer on TBPS binding where the reducing effect of the 3β-isomer was not detected (paper I). According to the “desensitization hypothesis” the 3β-steroids are suggested to increase the desensitization rate and thereby also increase the accumulation of receptors at desensitized state. However, the GABA is still bound to the receptor and therefore no effect on the TBPS binding is detected due to the fact that TBPS binds to the unbound receptor.

In the present thesis it was shown that the steroids with similar structures which involved the 3α-hydroxy group of, Allo, THDOC, 3α5β-THDOC and the glucocorticoid metabolites Allo-THF, THF and THE all modulated the receptor. In contrast to the enhanced effect of Allo and the two isomers of THDOC, the glucocorticoid metabolites reduced the GABA mediated effect. Interestingly, both the positive and the negative modulation occurred in a concentration dependent way (paper IV). Further, the negative modulation of the glucocorticoid metabolites was blocked in presence of Allo. This indicates that in a stress situation the glucocorticoid metabolites can interact with the GABAA receptor function as well as with other GABAA receptor modulators.

The results in the present thesis shows that the configurations of the C-3 hydroxy position of the steroid A-ring is not crucial for the positive modulation of the GABAA receptor function. The 3α-hydroxy glucocorticoid metabolites can be negative modulators as well as the 3β-hydroxy metabolites can be positive modulators. This indicates a complex situation where both sex and stress steroid metabolites interact with the GABAA receptor function. Allo and other GABAA receptor modulators have been shown to give negative mood changes in a subgroup of individuals both animals and humans (Masia et al., 2000; Miczek et al., 1997; Miczek et al., 2003; Wenzel et al., 2002). 3β-steroids reduced the Allo effect, at least in some way, in a selective way. 3β-steroids may therefore be of interest to use as therapeutics against the negative symptoms induced by fluctuated GABAA receptor function. The knowledge that diversity of endogenous steroids interact with the GABAA receptor function is of importance for further understanding of different sex and stress steroid related symptoms and syndromes.

48 Conclusions ƒ 3β-OH-5α-pregnane-20-one specifically reduces the Allo enhancement in a non-competitive way

ƒ 3β-OH-5α-pregnane-20-one do not modulates the GABAA receptor in absence or presence of GABA

ƒ 5β-pregnane-3β, 20(S)-diol and 3β-OH-5β-pregnane-20-one have dualistic properties and act both as negative as well as positive GABAA receptor modulators

ƒ 3β-steroids generally reduce the Allo-stimulated GABA effect in cortex, hippocampus and preoptic area

ƒ 5β-pregnane-3β, 20(R)-diol and PS act by increasing the desensitization rate almost identically to each other while they affect the Δ-amplitude and deactivation rate differently

ƒ The effect of 5β-pregnane-3β,20(R)-diol on the deactivation rate in presence of Allo is dependent on GABA exposure time

ƒ Glucocorticoid metabolites have a directly antagonistic effect on the GABAA receptor

ƒ In presence of Allo the reducing effect of Allo-THF and 3α5α-cortisone is completely blocked

ƒ Allo-THF but not 3α5α-cortison has a small enhancing effect on Allo stimulated GABA effect which could be blocked by the Allo specific antagonist 5α-pregnane-3β-OH-20-one

ƒ The 5β-isomer of THDOC has a weaker agonistic effect on the GABA mediated chloride ion uptake compare to 3α5α-THDOC and Allo

49 Acknowledgements

I wish to express my sincerely gratitude to all of you who have helped me through this work in one way or another. I am especially grateful to:

Dr. Per Lundgren, my supervisor, for introduce me into the exceptional world of science. Thank you for always taking your time for me, whatever the reason was. For your never-ending patience and for being the structure in my chaos!

Professor Torbjörn Bäckström, my co-supervisor, for accept me as a student at the first place. For your enthusiasm, for all discussions and especially for the way you have encouraged me to think and to argue. I will always remember your stories.

Dr. David Haage, my co-supervisor, for introduce me into the field of patch- clamp technique. For all discussions about science and research but also about life in general.

Professor Ann Lalos, head of the Department of Obstetrics and Gyneacology, for support and personal care.

Dr. Inga-Maj Johansson, thank you for fruitful discussions and for introducing V75 to our Fridays. Monica Isaksson, for your positive mind and care and for always being helpful on the lab. I admire your network of knowledge. Dr. Gianna Ragagnin, for friendship and for chemical discussions, they are always fruitful. Dr. Mingde Wang for valuable comments and discussions. Dr. Magdalena Taube, thank you for your friendship and for motivating discussions.

Agneta Andersson, you were one of the first people who made me feel welcome. Thank you for all discussions and for your sensitivity about my wellbeing. I wish you all luck in future.

Elisabeth Zingmark, thank you for your laboratory helps, for your friendship and for the talks about life in general and stuff in specific. I admire you positivism.

Tobias Näslund, for support with economy and for all chats about worldly matter.

50 Carina Jonsson and Birgit Ericson thank you for all support with administration and for all talks about everything and nothing.

All people at the Department of Obstetrics and Gynecology, for making the conferences, the Wednesday-lunch meetings and the “enhetsdagar” to a pleasure.

Friends, PhD-students and former PhD-students at UNC, Mozibur Rahman, Magnus Löfgren, Sahruh Turkmen, Di Zhu, Vita Birzniece thank you for all chats and friendship, and a special thanks to Charlotte Lindblad for discussions, friendship and for your help.

Thank you, Jenny Åström and Britt-Inger Gladzci for taken good care of the rats. You have always given me your time even with very short notice. I have really appreciated that.

Maria Nordendahl, for all discussion about the life as a PhD student and for the long discussions about statistics. Thank you for your friendship, including all dog talks both in the forest and at dinners. Thank you Conny Nordendahl for the computer support and for all mockery.

Carina Pettersson, thank you for being such good friend. For always listen and always caring. Looking forward to weekends with good wine and lot of important pedigree discussions!

Mats Lyngmark, for be the one you are. Thank you for your patience in my bad days and for being the person who barks at me, when it is necessary!

My mother Margareta Strömberg, thank you for everything, for your faith in me and for always being there.

My two sisters, Madelene and Emma I love you.

51 References

Arshavsky, Y. I., Deliagina, T. G., Gamkrelidze, G. N., Orlovsky, G. N., Panchin, Y. V., Popova, L. B., and Shupliakov, O. V., 1993, Pharmacologically induced elements of the hunting and feeding behavior in the pteropod mollusk Clione limacina. I. Effects of GABA, J Neurophysiol 69(2):512-21.

Backstrom, T., Andersson, A., Andree, L., Birzniece, V., Bixo, M., Bjorn, I., Haage, D., Isaksson, M., Johansson, I. M., Lindblad, C., Lundgren, P., Nyberg, S., Odmark, I. S., Stromberg, J., Sundstrom-Poromaa, I., Turkmen, S., Wahlstrom, G., Wang, M., Wihlback, A. C., Zhu, D., and Zingmark, E., 2003, Pathogenesis in menstrual cycle-linked CNS disorders, Ann N Y Acad Sci 1007:42-53.

Barbaccia, M. L., Concas, A., Serra, M., and Biggio, G., 1998, Stress and neurosteroids in adult and aged rats, Exp Gerontol 33(7-8):697-712.

Barbaccia, M. L., Roscetti, G., Trabucchi, M., Purdy, R. H., Mostallino, M. C., Concas, A., and Biggio, G., 1997, The effects of inhibitors of GABAergic transmission and stress on brain and plasma allopregnanolone concentrations, Br J Pharmacol 120(8):1582-8.

Barnard, E. A., 2001, The molecular architecture of GABAA receptors, in: Pharmacology of GABA and Glycin Transmission (H. Mohler, ed.), Springer, Berlin, pp. 79-99.

Barnard, E. A., Skolnick, P., Olsen, R. W., Mohler, H., Sieghart, W., Biggio, G., Braestrup, C., Bateson, A. N., and Langer, S. Z., 1998, International Union of Pharmacology. XV. Subtypes of gamma-aminobutyric acidA receptors: classification on the basis of subunit structure and receptor function, Pharmacol Rev 50(2):291-313.

Baulieu, E. E., Robel, P., and Schumacher, M., 2001, Neurosteroids: beginning of the story, Int Rev Neurobiol 46:1-32.

Belelli, D., and Lambert, J. J., 2005, Neurosteroids: endogenous regulators of the GABA(A) receptor, Nat Rev Neurosci 6(7):565-75.

Ben-Ari, Y., Khazipov, R., Leinekugel, X., Caillard, O., and Gaiarsa, J. L., 1997, GABAA, NMDA and AMPA receptors: a developmentally regulated 'menage a trois', Trends Neurosci 20(11):523-9.

Berne, R., Levy, M., Koeppen, B., and Stanton, B., 1998, Physiology, Mosby, St. Louis, Missouri, USA, pp. 938-957.

52 Bixo, M., Andersson, A., Winblad, B., Purdy, R. H., and Backstrom, T., 1997, Progesterone, 5alpha-pregnane-3,20-dione and 3alpha-hydroxy-5alpha- pregnane-20-one in specific regions of the human female brain in different endocrine states, Brain Res 764(1-2):173-8.

Bormann, J., 2000, The "ABC" of GABA receptors, TiPS 21:16-19.

Borski, R. J., 2000, Nongenomic membrane actions of glucocorticoids in vertebrates, Trends Endocrinol Metab 11(10):427-36.

Brodal, P., 1992, The Central Nervous System structure and function, Oxford University Press, Inc, New York.

Callachan, H., Cottrell, G. A., Hather, N. Y., Lambert, J. J., Nooney, J. M., and Peters, J. A., 1987, Modulation of the GABAA receptor by progesterone metabolites, Proc R Soc Lond B Biol Sci 231(1264):359-69.

Carroll, B. J., Heath, B., and Jarrett, D. B., 1975, Corticosteroids in brain tissue, Endocrinology 97:290-300.

Celotti, F., Melcangi, R. C., and Martini, L., 1992, The 5 alpha-reductase in the brain: molecular aspects and relation to brain function, Front Neuroendocrinol 13(2):163-215.

Chebib, M., and Johnston, G. A., 1999, The 'ABC' of GABA receptors: a brief review, Clin Exp Pharmacol Physiol 26(11):937-40.

Concas, A., Corda, M. G., and Biggio, G., 1985, Involvement of benzodiazepine recognition sites in the foot shock-induced decrease of low affinity GABA receptors in the rat cerebral cortex, Brain Res 341(1):50-6.

Concas, A., Mostallino, M. C., Porcu, P., Follesa, P., Barbaccia, M. L., Trabucchi, M., Purdy, R. H., Grisenti, P., and Biggio, G., 1998, Role of brain allopregnanolone in the plasticity of gamma-aminobutyric acid type A receptor in rat brain during pregnancy and after delivery, Proc Natl Acad Sci U S A 95(22):13284-9.

Concas, A., Serra, M., Atsoggiu, T., and Biggio, G., 1988, Foot-shock stress and beta-carbolines increase t-[35S]butylbicyclophosphorothionate binding in the rat cerebral cortex, an effect opposite to and gamma-aminobutyric acid mimetics, J Neurochem 51(6):1868-76.

Corpechot, C., Young, J., Calvel, M., Wehrey, C., Veltz, J. N., Touyer, G., Mouren, M., Prasad, V. V., Banner, C., Sjovall, J., and et al., 1993, Neurosteroids: 3 alpha-hydroxy-5 alpha-pregnan-20-one and its

53 precursors in the brain, plasma, and steroidogenic glands of male and female rats, Endocrinology 133(3):1003-9.

Cottrell, G. A., Lambert, J. J., and Peters, J. A., 1987, Modulation of GABAA receptor activity by alphaxalone, Br J Pharmacol 90(3):491-500. de Quervain, D. J., Roozendaal, B., and McGaugh, J. L., 1998, Stress and glucocorticoids impair retrieval of long-term spatial memory, Nature 394(6695):787-90.

DeLorey, T. M., and Brown, G. B., 1992, gamma-Aminobutyric acidA receptor pharmacology in rat cerebral cortical synaptoneurosomes, J Neurochem 58(6):2162-9.

Droogleever Fortuyn, H. A., van Broekhoven, F., Span, P. N., Backstrom, T., Zitman, F. G., and Verkes, R. J., 2004, Effects of PhD examination stress on allopregnanolone and cortisol plasma levels and peripheral benzodiazepine receptor density, Psychoneuroendocrinology 29(10):1341-4.

Drugan, R. C., Morrow, A. L., Weizman, R., Weizman, A., Deutsch, S. I., Crawley, J. N., and Paul, S. M., 1989, Stress-induced behavioral depression in the rat is associated with a decrease in GABA receptor- mediated chloride ion flux and brain benzodiazepine receptor occupancy, Brain Res 487(1):45-51.

Dubrovsky, B., 2000, The specificity of stress responses to different nocuous stimuli: neurosteroids and depression, Brain Res Bull 51(6):443-55.

Edwards, F., Konnerth, A., and Sakmann, B., 1990, Quantal analysis of inhibitory synaptic transmission in the dentate gyrus of rat hippocampal slices: a patch-clamp study, Journal of Physiology 430:213-249.

Eisenman, L. N., He, Y., Fields, C., Zorumski, C. F., and Mennerick, S., 2003, Activation-dependent properties of pregnenolone sulfate inhibition of GABAA receptor-mediated current, J. Physiol. 550.3:679-691.

Enz, R., and Cutting, G. R., 1998, Molecular composition of GABAC receptors, Vision Res 38(10):1431-41.

Eser, D., Romeo, E., Baghai, T. C., di Michele, F., Schule, C., Pasini, A., Zwanzger, P., Padberg, F., and Rupprecht, R., 2006, Neuroactive steroids as modulators of depression and anxiety, Neuroscience 138(3):1041-8.

54 Evans, R. M., and Zamponi, G. W., 2006, Presynaptic Ca2+ channels-- integration centers for neuronal signaling pathways, Trends Neurosci 29(11):617-24.

Falkenstein, E., Tillmann, H. C., Christ, M., Feuring, M., and Wehling, M., 2000, Multiple actions of steroid hormones--a focus on rapid, nongenomic effects, Pharmacol Rev 52(4):513-56.

Florkowski, C. M., Crozier, I. G., Nightingale, S., Evans, M. J., Ellis, M. J., Joyce, P., and Donald, R. A., 1996, Plasma cortisol, PRL, ACTH, AVP and corticotrophin releasing hormone responses to direct current cardioversion and electroconvulsive therapy, Clin Endocrinol (Oxf) 44(2):163-8.

Foddi, M. C., Cinquanta, M., and Mennini, T., 1997, Effect of psychological stress on [35S]TBPS binding in rat brain, Pharmacol Biochem Behav 58(2):373-7.

Fodor, L., Biro, T., and Maksay, G., 2005, Nanomolar allopregnanolone potentiates rat cerebellar GABAA receptors, Neurosci Lett 383(1-2):127- 30.

Follesa, P., Biggio, F., Caria, S., Gorini, G., and Biggio, G., 2004, Modulation of GABA(A) receptor gene expression by allopregnanolone and ethanol, Eur J Pharmacol 500(1-3):413-25.

Fukushima, D. K., Bradlow, H. L., Hellman, L., Zumoff, B., and Gallagher, T. F., 1960, Metabolic transformation of hydrocortisone-4-C14 in normal men, J Biol Chem 235:2246-52.

Ganong, W., 1997, Review of Medical Physiology (W. Ganong, ed.), Appleton & Lange, San Francisco.

Gasior, M., Carter, R. B., and Witkin, J. M., 1999, Neuroactive steroids: potential therapeutic use in neurological and psychiatric disorders, Trends Pharmacol Sci 20(3):107-12.

Gee, K. W., 1988, Steroid modulation of the GABA/benzodiazepine receptor- linked chloride ionophore, Mol Neurobiol 2(4):291-317.

Girdler, S. S., Straneva, P. A., Light, K. C., Pedersen, C. A., and Morrow, A. L., 2001, Allopregnanolone levels and reactivity to mental stress in premenstrual dysphoric disorder, Biol Psychiatry 49(9):788-97.

Guennoun, R., Fiddes, R. J., Gouezou, M., Lombes, M., and Baulieu, E. E., 1995, A key enzyme in the biosynthesis of neurosteroids, 3 beta-

55 hydroxysteroid dehydrogenase/delta 5-delta 4-isomerase (3 beta-HSD), is expressed in rat brain, Brain Res Mol Brain Res 30(2):287-300.

Guennoun, R., Schumacher, M., Robert, F., Delespierre, B., Gouezou, M., Eychenne, B., Akwa, Y., Robel, P., and Baulieu, E. E., 1997, Neurosteroids: expression of functional 3beta-hydroxysteroid dehydrogenase by rat sensory neurons and Schwann cells, Eur J Neurosci 9(11):2236-47.

Haage, D., Backstrom, T., and Johansson, S., 2005, Interaction between allopregnanolone and pregnenolone sulfate in modulating GABA- mediated synaptic currents in neurons from the rat medial preoptic nucleus, Brain Res 1033(1):58-67.

Haage, D., Druzin, M., and Johansson, S., 2002, Allopregnanolone modulates spontaneous GABA release via presynaptic Cl- permeability in rat preoptic nerve terminals, Brain Res 958(2):405-13.

Haage, D., and Johansson, S., 1999, Neurosteroid modulation of synaptic and GABA-evoked currents in neurons from the rat medial preoptic nucleus, J Neurophysiol 82(1):143-51.

Harrison, N. L., and Simmonds, M. A., 1984, Modulation of the GABA receptor complex by a steroid anaesthetic, Brain Res 323(2):287-92.

Hawkinson, J. E., Kimbrough, C. L., Belelli, D., Lambert, J. J., Purdy, R. H., and Lan, N. C., 1994, Correlation of neuroactive steroid modulation of [35S]t-butylbicyclophosphorothionate and [3H]flunitrazepam binding and gamma-aminobutyric acidA receptor function, Mol Pharmacol 46(5):977- 85.

Havlikova, H., Hill, M., Kancheva, L., Vrbikova, J., Pouzar, V., Cerny, I., Kancheva, R., and Starka, L., 2006, Serum profiles of free and conjugated neuroactive pregnanolone isomers in nonpregnant women of fertile age, J Clin Endocrinol Metab 91(8):3092-9.

Herman, J. P., Ostrander, M. M., Mueller, N. K., and Figueiredo, H., 2005, Limbic system mechanisms of stress regulation: hypothalamo-pituitary- adrenocortical axis, Prog Neuropsychopharmacol Biol Psychiatry 29(8):1201-13.

Horn, R., and Marty, A., 1988, Muscarinic activation of ionic currents measured by a new wole-cell recording method, J. Gen. Physiol. 92:145-159.

56 Hosie, A. M., Wilkins, M. E., da Silva, H. M., and Smart, T. G., 2006, Endogenous neurosteroids regulate GABAA receptors through two discrete transmembrane sites, Nature 444(7118):486-9.

Hu, Z. Y., Bourreau, E., Jung-Testas, I., Robel, P., and Baulieu, E. E., 1987, Neurosteroids: oligodendrocyte mitochondria convert cholesterol to pregnenolone, Proc Natl Acad Sci U S A 84(23):8215-9.

Hu, Z. Y., Jung-Testas, I., Robel, P., and Baulieu, E. E., 1989, Neurosteroids: steroidogenesis in primary cultures of rat glial cells after release of aminoglutethimide blockade, Biochem Biophys Res Commun 161(2):917- 22.

Ikeda, M., Weber, K. H., Bechtel, W. D., Malatynska, E., and Yamamura, H. I., 1989, Relative efficacies of 1,4-diazepines on GABA-stimulated chloride influx in rat brain vesicles, Life Sci 45(4):349-58.

Johansson, I. M., Birzniece, V., Lindblad, C., Olsson, T., and Backstrom, T., 2002, Allopregnanolone inhibits learning in the Morris water maze, Brain Res 934(2):125-31.

Jones, M. V., and Westbrook, G. L., 1995, Desensitized states prolong GABAA channel responses to brief agonist pulses, Neuron 15(1):181-91.

Jung-Testas, I., Weintraub, H., Dupuis, D., Eychenne, B., Baulieu, E. E., and Robel, P., 1992, Low density lipoprotein-receptors in primary cultures of rat glial cells, J Steroid Biochem Mol Biol 42(6):597-605.

Korpi, E. R., Grunder, G., and Luddens, H., 2002, Drug interactions at GABA(A) receptors, Prog Neurobiol 67(2):113-59.

Korpi, E. R., and Luddens, H., 1993, Regional gamma-aminobutyric acid sensitivity of t-butylbicyclophosphoro[35S]thionate binding depends on gamma-aminobutyric acidA receptor alpha subunit, Mol Pharmacol 44(1):87-92.

Korpi, E. R., and Sinkkonen, S. T., 2006, GABA(A) receptor subtypes as targets for neuropsychiatric drug development, Pharmacol Ther 109(1-2):12-32.

Korte, S. M., 2001, Corticosteroids in relation to fear, anxiety and psychopathology, Neurosci Biobehav Rev 25(2):117-42.

Lambert, J. J., Belelli, D., Hill-Venning, C., and Peters, J. A., 1995, Neurosteroids and GABAA receptor function, Trends Pharmacol Sci 16(9):295-303.

57 Le Foll, F., Louiset, E., Castel, H., Vaudry, H., and Cazin, L., 1997, Electrophysiological effects of various neuroactive steroids on the GABA(A) receptor in pituitary melanotrope cells, Eur J Pharmacol 331(2-3):303-11.

Luddens, H., and Korpi, E. R., 1995, Biological function of GABAA/benzodiazepine receptor heterogeneity, J Psychiatr Res 29(2):77-94.

Luisi, S., Petraglia, F., Benedetto, C., Nappi, R. E., Bernardi, F., Fadalti, M., Reis, F. M., Luisi, M., and Genazzani, A. R., 2000, Serum allopregnanolone levels in pregnant women: changes during pregnancy, at delivery, and in hypertensive patients, J Clin Endocrinol Metab 85(7):2429-33.

Luu, M. D., Morrow, A. L., Paul, S. M., and Schwartz, R. D., 1987, Characterization of GABAA receptor-mediated 36chloride uptake in rat brain synaptoneurosomes, Life Sci 41(10):1277-87.

Macdonald, R. L., and Olsen, R. W., 1994, GABAA receptor channels, Annu Rev Neurosci 17:569-602.

Majewska, M. D., 1992, Neurosteroids: endogenous bimodal modulators of the GABAA receptor. Mechanism of action and physiological significance, Prog Neurobiol 38(4):379-95.

Majewska, M. D., Bisserbe, J.-C., and Eskay, R. L., 1985, Glucocorticoids are modulators of GABAA recepros in brain, Brain Res 339:178-182.

Majewska, M. D., Ford-Rice, F., and Falkay, G., 1989, Pregnancy-induced alterations of GABAA receptor sensitivity in maternal brain: an antecedent of post-partum 'blues'?, Brain Res 482(2):397-401.

Makara, G. B., and Haller, J., 2001, Non-genomic effects of glucocorticoids in the neural system. Evidence, mechanisms and implications, Prog Neurobiol 65(4):367-90.

Masia, S. L., Perrine, K., Westbrook, L., Alper, K., and Devinsky, O., 2000, Emotional outbursts and post-traumatic stress disorder during intracarotid procedure, Neurology 54(8):1691-3.

McEwen, B. S., 2002, Protective and damaging effects of stress mediators: the good and bad sides of the response to stress, Metabolism 51(6 Suppl 1):2- 4.

58 McKernan, R. M., Rosahl, T. W., Reynolds, D. S., Sur, C., Wafford, K. A., Atack, J. R., Farrar, S., Myers, J., Cook, G., Ferris, P., Garrett, L., Bristow, L., Marshall, G., Macaulay, A., Brown, N., Howell, O., Moore, K. W., Carling, R. W., Street, L. J., Castro, J. L., Ragan, C. I., Dawson, G. R., and Whiting, P. J., 2000, Sedative but not anxiolytic properties of benzodiazepines are mediated by the GABA(A) receptor alpha1 subtype, Nat Neurosci 3(6):587-92.

Mehta, A. K., and Ticku, M. K., 1999, An update on GABAA receptors, Brain Res Brain Res Rev 29(2-3):196-217.

Melcangi, R. C., Magnaghi, V., and Martini, L., 1999, Steroid metabolism and effects in central and peripheral glial cells, J Neurobiol 40(4):471-83.

Mellon, S. H., and Vaudry, H., 2001, Biosynthesis of neurosteroids and regulation of their synthesis, Int Rev Neurobiol 46:33-78.

Miczek, K. A., DeBold, J. F., van Erp, A. M., and Tornatzky, W., 1997, , GABAA-benzodiazepine receptor complex, and aggression, Recent Dev Alcohol 13:139-71.

Miczek, K. A., Fish, E. W., and De Bold, J. F., 2003, Neurosteroids, GABAA receptors, and escalated aggressive behavior, Horm Behav 44(3):242-57.

Miller, W. L., 1988, Molecular biology of steroid hormone synthesis, Endocr Rev 9(3):295-318.

Mody, I., and Pearce, R. A., 2004, Diversity of inhibitory neurotransmission through GABA(A) receptors, Trends Neurosci 27(9):569-75.

Morell, P., and Jurevics, H., 1996, Origin of cholesterol in myelin, Neurochem Res 21(4):463-70.

Morrow, A. L., and Paul, S. M., 1988, Benzodiazepine enhancement of gamma- aminobutyric acid-mediated chloride ion flux in rat brain synaptoneurosomes, J Neurochem 50(1):302-6.

Mumenthaler, M. S., O'Hara, R., Taylor, J. L., Friedman, L., and Yesavage, J. A., 2001, Relationship between variations in estradiol and progesterone levels across the menstrual cycle and human performance, Psychopharmacology (Berl) 155(2):198-203.

Neher, E., 1992, Correction for liquid junction potentials in patch clamp experiments, Methods Enzymol. 207:123-131.

59 Ottander, U., Poromaa, I. S., Bjurulf, E., Skytt, A., Backstrom, T., and Olofsson, J. I., 2005, Allopregnanolone and pregnanolone are produced by the human corpus luteum, Mol Cell Endocrinol 239(1-2):37-44.

Pardridge, W. M., and Mietus, L. J., 1980a, Effects of progesterone-binding globulin versus a progesterone antiserum on steroid hormone transport through the blood-brain barrier, Endocrinology 106(4):1137-41.

Pardridge, W. M., and Mietus, L. J., 1980b, Transport of thyroid and steroid hormones through the blood-brain barrier of the newborn rabbit: primary role of protein-bound hormone, Endocrinology 107(6):1705-10.

Paul, S. M., and Purdy, R. H., 1992, Neuroactive steroids, Faseb J 6(6):2311-22.

Penland, S. N., and Morrow, A. L., 2004, 3alpha,5beta-Reduced cortisol exhibits antagonist properties on cerebral cortical GABA(A) receptors, Eur J Pharmacol 506(2):129-32.

Peters, A., 2002, Examining neocortical circuits: some background and facts, J Neurocytol 31(3-5):183-93.

Peters, J. A., Kirkness, E. F., Callachan, H., Lambert, J. J., and Turner, A. J., 1988, Modulation of the GABAA receptor by depressant barbiturates and pregnane steroids, Br J Pharmacol 94(4):1257-69.

Pinel, J., 2000, Biophysiology, Allyn & Bacon.

Pirker, S., Schwarzer, C., Wieselthaler, A., Sieghart, W., and Sperk, G., 2000, GABA(A) receptors: immunocytochemical distribution of 13 subunits in the adult rat brain, Neuroscience 101(4):815-50.

Prince, R. J., and Simmonds, M. A., 1992, 5 beta-pregnan-3 beta-ol-20-one, a specific antagonist at the neurosteroid site of the GABAA receptor- complex, Neurosci Lett 135(2):273-5.

Prince, R. J., and Simmonds, M. A., 1993, Differential antagonism by epipregnanolone of alphaxalone and pregnanolone potentiation of [3H]flunitrazepam binding suggests more than one class of binding site for steroids at GABAA receptors, Neuropharmacology 32(1):59-63.

Purdy, R. H., Morrow, A. L., Blinn, J. R., and Paul, S. M., 1990, Synthesis, metabolism, and pharmacological activity of 3 alpha-hydroxy steroids which potentiate GABA-receptor-mediated chloride ion uptake in rat cerebral cortical synaptoneurosomes, J Med Chem 33(6):1572-81.

60 Purdy, R. H., Morrow, A. L., Moore, P. H., Jr., and Paul, S. M., 1991, Stress- induced elevations of gamma-aminobutyric acid type A receptor-active steroids in the rat brain, Proc Natl Acad Sci U S A 88(10):4553-7.

Purves, D., Augustine, G., Fitzpatrick, D., Hall, W., LaMantia, A.-S., McNamara, J., and Williams, M., 2004, Neuroscience (D. Purves, G. Augustine, D. Fitzpatrick, W. Hall, A.-S. LaMantia, J. McNamara, and M. Williams, eds.), Sinauer Associates, Inc, Sunderland, Massachusetts.

Rapkin, A. J., Morgan, M., Goldman, L., Brann, D. W., Simone, D., and Mahesh, V. B., 1997, Progesterone metabolite allopregnanolone in women with premenstrual syndrome, Obstet Gynecol 90(5):709-14.

Reagan, L. P., and McEwen, B. S., 1997, Controversies surrounding glucocorticoid-mediated cell death in the hippocampus, J Chem Neuroanat 13(3):149-67.

Reddy, D. S., 2003, Is there a physiological role for the neurosteroid THDOC in stress-sensitive conditions?, Trends Pharmacol Sci 24(3):103-6.

Reddy, D. S., and Rogawski, M. A., 2002, Stress-induced deoxycorticosterone- derived neurosteroids modulate GABA(A) receptor function and seizure susceptibility, J Neurosci 22(9):3795-805.

Reimer, R. J., Fon, E. A., and Edwards, R. H., 1998, Vesicular neurotransmitter transport and the presynaptic regulation of quantal size, Curr Opin Neurobiol 8(3):405-12.

Romeo, E., Strohle, A., Spalletta, G., di Michele, F., Hermann, B., Holsboer, F., Pasini, A., and Rupprecht, R., 1998, Effects of antidepressant treatment on neuroactive steroids in major depression, Am J Psychiatry 155(7):910- 3.

Rothman, J., 1996, The protein machinery of vesicle budding and fusion, Protein Science 5:185-194.

Sanne, J. L., and Krueger, K. E., 1995, Expression of cytochrome P450 side- chain cleavage enzyme and 3 beta-hydroxysteroid dehydrogenase in the rat central nervous system: a study by polymerase chain reaction and in situ hybridization, J Neurochem 65(2):528-36.

Schwartz, R. D., Skolnick, P., Hollingsworth, E. B., and Paul, S. M., 1984, Barbiturate and picrotoxin-sensitive chloride efflux in rat cerebral cortical synaptoneurosomes, FEBS Lett 175(1):193-6.

61 Schwartz, R. D., Suzdak, P. D., and Paul, S. M., 1986, gamma-Aminobutyric acid (GABA)- and barbiturate-mediated 36Cl- uptake in rat brain synaptoneurosomes: evidence for rapid desensitization of the GABA receptor-coupled chloride ion channel, Mol Pharmacol 30(5):419-26.

Seckl, J. R., and Walker, B. R., 2004, 11beta-hydroxysteroid dehydrogenase type 1 as a modulator of glucocorticoid action: from metabolism to memory, Trends Endocrinol Metab 15(9):418-24.

Serra, M., Pisu, M. G., Littera, M., Papi, G., Sanna, E., Tuveri, F., Usala, L., Purdy, R. H., and Biggio, G., 2000, Social isolation-induced decreases in both the abundance of neuroactive steroids and GABA(A) receptor function in rat brain, J Neurochem 75(2):732-40.

Shen, W., Mennerick, S., Covey, D. F., and Zorumski, C. F., 2000, Pregnenolone sulfate modulates inhibitory synaptic transmission by enhancing GABA(A) receptor desensitization, J Neurosci 20(10):3571-9.

Sieghart, W., 1992, GABAA receptors: ligand-gated Cl- ion channels modulated by multiple drug-binding sites, Trends Pharmacol Sci 13(12):446-50.

Sieghart, W., 1995, Structure and pharmacology of gamma-aminobutyric acidA receptor subtypes, Pharmacol Rev 47(2):181-234.

Sieghart, W., Fuchs, K., Tretter, V., Ebert, V., Jechlinger, M., Hoger, H., and Adamiker, D., 1999, Structure and subunit composition of GABA(A) receptors, Neurochem Int 34(5):379-85.

Stromstedt, M., Warner, M., Banner, C. D., MacDonald, P. C., and Gustafsson, J. A., 1993, Role of brain cytochrome P450 in regulation of the level of anesthetic steroids in the brain, Mol Pharmacol 44(5):1077-83.

Sundstrom, I., Backstrom, T., Wang, M., Olsson, T., Seippel, L., and Bixo, M., 1999, Premenstrual syndrome, neuroactive steroids and the brain, Gynecol Endocrinol 13(3):206-20.

Tsilchorozidou, T., Honour, J. W., and Conway, G. S., 2003, Altered cortisol metabolism in polycystic ovary syndrome: insulin enhances 5alpha- reduction but not the elevated adrenal steroid production rates, J Clin Endocrinol Metab 88(12):5907-13.

Turkmen, S., Lundgren, P., Birzniece, V., Zingmark, E., Backstrom, T., and Johansson, I. M., 2004, 3beta-20beta-dihydroxy-5alpha-pregnane (UC1011) antagonism of the GABA potentiation and the learning impairment induced in rats by allopregnanolone, Eur J Neurosci 20(6):1604-12.

62 Turner, D. M., Ransom, R. W., Yang, J. S., and Olsen, R. W., 1989, Steroid and naturally occurring analogs modulate the gamma- aminobutyric acid receptor complex at a site distinct from barbiturates, J Pharmacol Exp Ther 248(3):960-6.

Uzunova, V., Sheline, Y., Davis, J. M., Rasmusson, A., Uzunov, D. P., Costa, E., and Guidotti, A., 1998, Increase in the cerebrospinal fluid content of neurosteroids in patients with unipolar major depression who are receiving or fluvoxamine, Proc Natl Acad Sci U S A 95(6):3239-44.

Wafford, K. A., 2005, GABAA receptor subtypes: any clues to the mechanism of benzodiazepine dependence?, Curr Opin Pharmacol 5(1):47-52.

Wang, M., Backstrom, T., Sundstrom, I., Wahlstrom, G., Olsson, T., Zhu, D., Johansson, I. M., Bjorn, I., and Bixo, M., 2001, Neuroactive steroids and central nervous system disorders, Int Rev Neurobiol 46:421-59.

Wang, M., He, Y., Eisenman, L. N., Fields, C., Zeng, C. M., Mathews, J., Benz, A., Fu, T., Zorumski, E., Steinbach, J. H., Covey, D. F., Zorumski, C. F., and Mennerick, S., 2002, 3beta -hydroxypregnane steroids are pregnenolone sulfate-like GABA(A) receptor antagonists, J Neurosci 22(9):3366-75.

Wang, M., Seippel, L., Purdy, R. H., and Backstrom, T., 1996, Relationship between symptom severity and steroid variation in women with premenstrual syndrome: study on serum pregnenolone, pregnenolone sulfate, 5 alpha-pregnane-3,20-dione and 3 alpha-hydroxy-5 alpha- pregnan-20-one, J Clin Endocrinol Metab 81(3):1076-82.

Wang, M. D., Backstrom, T., and Landgren, S., 2000, The inhibitory effects of allopregnanolone and pregnanolone on the population spike, evoked in the rat hippocampal CA1 stratum pyramidale in vitro, can be blocked selectively by epiallopregnanolone, Acta Physiol Scand 169(4):333-41.

Wellman, C. L., 2001, Dendritic reorganization in pyramidal neurons in medial prefrontal cortex after chronic corticosterone administration, J Neurobiol 49(3):245-53.

Wenzel, R. R., Bartel, T., Eggebrecht, H., Philipp, T., and Erbel, R., 2002, Central-nervous side effects of during transesophageal echocardiography, J Am Soc Echocardiogr 15(10 Pt 2):1297-300.

Wheling, M., 1997, Specific, nongenomic actions of steroid hormones, Annual Review of physiology 59:365-93.

63 Whiting, P. J., 2003, GABA-A receptor subtypes in the brain: a paradigm for CNS drug discovery?, Drug Discov Today 8(10):445-50.

Wikinski, S. I., Acosta, G. B., Gravielle, M. C., Bonavita, C. D., Bisagno, V., Fiszer de Plazas, S., and Rubio, M. C., 2001, Diazepam fails to potentiate GABA-induced chloride uptake and to produce anxiolytic-like action in aged rats, Pharmacol Biochem Behav 68(4):721-7.

Wilson, M. A., and Biscardi, R., 1997, Influence of gender and brain region on neurosteroid modulation of GABA responses in rats, Life Sci 60(19):1679-91.

Wisden, W., Laurie, D. J., Monyer, H., and Seeburg, P. H., 1992, The distribution of 13 GABAA receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon, mesencephalon, J Neurosci 12(3):1040-62.

Vorobjev, V., 1991, Vibrodissociation of sliced mammalian nervous, J Neuroscience methods 38:145-150.

Zhu, W. J., and Vicini, S., 1997, Neurosteroid prolongs GABAA channel deactivation by altering kinetics of desensitized states, J Neurosci 17(11):4022-31.

64