The Antidepressant-Like Effects of the 3Β-Hydroxysteroid

Neuropharmacology xxx (2011) 1e11

Contents lists available at SciVerse ScienceDirect

Neuropharmacology

journal homepage: www.elsevier.com/locate/neuropharm

The antidepressant-like effects of the 3b-hydroxysteroid dehydrogenase inhibitor trilostane in mice is related to changes in neuroactive steroid and monoamine levels

Julie Espallerguesa,b,c,g, Takayoshi Mamiyad, Monique Valléee,f, Takenao Kosekid, Toshitaka Nabeshimad, Jamal Temsamanig, Claude Laruellee, Tangui Mauricea,b,c,* a Université de Montpellier II, 34095 Montpellier, France b Inserm U. 710, 34095 Montpellier cedex 5, France c E.P.H.E., 75017 Paris, France d Department of Chemical Pharmacology, Graduate School of Pharmaceutical Science, Meijo University, Nagoya 468-8503, Japan e Inserm U. 862, Neurocentre Magendie, Physiopathology of Addiction Group, 33000 Bordeaux, France f Université de Bordeaux, 33000 Bordeaux, France g CLL Pharma, 06200 Nice, France article info abstract

Article history: In the present study, we analyzed the effects of a systemic treatment with the competitive Received 25 February 2011 3b-hydroxysteroid dehydrogenase (3b-HSD) inhibitor trilostane on: (i) neurosteroid and monoamine Received in revised form levels in the brain, and (ii) the antidepressant activity of steroids and antidepressants in the forced 30 August 2011 swimming test (FST). 3b-HSD converts pregnenolone (PREG) into progesterone (PROG) or dehy- Accepted 6 September 2011 droepiandrosterone (DHEA) into androstenedione. These neuroactive steroids are known to regulate neurotransmitters effects in the brain, particularly glutamate, g-aminobutyric acid (GABA) and sero- Keywords: tonin (5-HT), with consequences on mood and depression. We previously reported that trilostane Trilostane 3b-hydroxysteroid dehydrogenase showed antidepressant-like properties in the FST and concomitantly regulated plasma adrenocortico- e e Neurosteroids tropin (ACTH) and corticosterone levels, markers of the stress-induced hypothalamus pituitary Monoamines adrenal (HPA) axis activation. We here observed that adrenalectomy/castration blocked the trilostane Forced swimming test effect, outlining the importance of peripheral steroid levels. Trilostane (25 mg/kg) decreased hippo- Depression campus PROG contents and paradoxically increased circulating PROG levels. It also increased PREG levels in the hippocampus and frontal cortex. In the FST, a co-treatment with trilostane facilitated DHEAS (5e20 mg/kg) antidepressant activity, but showed only an additive, not facilitative, effect with PREGS (10e40 mg/kg), PROG (10e40 mg/kg) or allopregnanolone (ALLO, 1e8 mg/kg). Trilostane (25 mg/kg) treatment significantly increased 5-HT and (-)-norepinephrine (NE) turnovers in the hippocampus, an effect likely related to its antidepressant action. In co-administration studies, trilostane further decreased immobility following fluoxetine (30e60 mg/kg), sertraline (20e40 mg/kg) and imipramine (20e40 mg/kg), but not desipramine (20e40 mg/kg), treatments. A significant additive effect was observed for the selective 5-HT reuptake inhibitors (SSRI) at their highest dose. This study confirmed that a systemic administration of trilostane directly affected peripheral and brain levels in neuroactive steroids and monoamine turnover, resulting in antidepressant activity. The drug could be proposed as a co-treatment with SSRI. This article is part of a Special Issue entitled ‘Anxiety and Depression’. Ó 2011 Elsevier Ltd. All rights reserved.

Abbreviations: ALLO, 5a-pregnan-3a-ol,20-one, allopregnanolone; 3a-HSOR, 3a-hydroxysteroid oxydoreductase; 3a/3b-HSD, 3a/3b-hydroxysteroid dehydrogenase; 5b-DHP, 5b-dihydroprogesterone; 5-HIAA, 5-hydroxyindole-3-acetic acid; 5-HT, serotonin; ACTH, adrenocorticotropin; AdX/CX, adrenalectomized/castrated; BDNF, brain- derived neurotrophic factor; CSF, cerebrospinal ﬂuid; DA, dopamine; DHEA, dehydroepiandrosterone; DHEAS, dehydroepiandrosterone sulfate ester; DOC, deoxycorticosterone; DOPAC, 3,4-dihydroxyphenylacetic acid; Epi, (-)-epinephrine; EPIALLO, 3b-hydroxy-5a-pregnan-ol-20-one, epiallopregnanolone; HVA, homovanillic acid; i.p., intraperitoneally; MHPG, 4-hydroxy-3-methoxyphenylglycol (MHPG); NE, (-)-norepinephrine; PREG, pregnenolone; PREGS, pregnenolone sulfate ester; PROG, progesterone; s.c., subcutaneously; SSRI, selective serotonin reuptake inhibitor; THDOC, 3a,5a-tetrahydrodeoxycorticosterone. * Corresponding author. University of Montpellier 2, Inserm U. 710, cc 105, PL E Bataillon, 34095 Montpellier cedex 5, France. E-mail address: [email protected] (T. Maurice).

1. Introduction P450 (Wang and Strobel, 1997). Moreover, both estrogens and antidepressants alter the levels of pregnane steroids by affecting Neuroactive steroids are steroid hormones that exert rapid non- enzymes involved in the synthesis of ALLO (Griffin and Mellon, genomic effects on nervous cells. A pool of neuroactive steroid is 1999), in relation with the relief of depressive symptoms during synthesized de novo by neurons and glial cells and contributes to the premenstrual syndrome and the perimenopausal period (Eser the higher concentrations found for several steroids in the brain et al., 2006; Marx et al., 2006; Uzunova et al., 2006). It appears (Corpéchot et al., 1981). These so-called ‘neurosteroids’ (Baulieu, therefore that several studies illustrate how PROG and estrogen 1981) include pregnenolone (PREG), dehydroepiandrosterone levels are responsible for the sensitivity of women to mood disor- (DHEA), progesterone (PROG) and its tetrahydro-reduced metabo- ders in direct relation to changes in steroid levels. lites, such as 5a-pregnan-3a-ol,20-one (allopregnanolone, ALLO). DHEA and its sulfate ester DHEAS may also play a role in The role of neuroactive steroids is to regulate the inhibitory/excit- depressive symptoms in humans, since the disease altered their atory balance in the brain and their impact in several psychiatric serum and urinary levels (Tollefson et al.,1990; Thomas et al.,1994). conditions is well known. For instance, some neuroactive steroids In open-label or double-blind randomized placebo-controlled have been reported to possess anti-stress, anxiolytic, antidepres- clinical trials, oral administration of DHEAS decreased depressive sant, antipsychotic, anticonvulsant, ataxic, and/or anesthetic symptoms in patients with major forms of depression (Wolkowitz properties in animals (Crawley et al., 1986; Landgren et al., 1987; et al., 1995, 1997, 1999a; Bloch et al., 1999). Moreover, in rodents, Belelli et al., 1990; Bitran et al., 1993; Korneyev and Costa, 1996; DHEA counteracted glucocorticoid actions by inhibiting glucocor- Urani et al., 2001). Their role in depressive states has gained ticoid enzyme activity, a plausible mechanism for its antidepres- major attention (see the recent reviews by Van Broekhoven and sant effect (Svec and Lopez, 1989; Browne et al., 1992; Wolkowitz Verkes, 2003; Eser et al., 2006; Girdler and Klatzkin, 2007; Reddy, et al., 1995, 1997, 1999b). DHEAS improves performance of 2010). In patients suffering from major depression, plasma and rodents in the forced swimming test (FST), a procedure that relies cerebrospinal fluid (CSF) levels in tetrahydro-reduced PROG on behavioral despair response in rodents and is routinely used to metabolites, including ALLO and pregnanolone, were found to be measure the antidepressant-like activity of drugs (Reddy et al., significantly lower than those measured in control subjects 1998; Urani et al., 2001). (Uzunova et al., 1998). In contrast, 3a,5a-tetrahydrodeox- The serotonergic system has also long been implicated in the ycorticosterone (THDOC) plasma levels were found to be signifi- neurobiology of mood disorders because, amongst other observa- cantly increased (Rupprecht, 2003; Ströhle et al., 1999). Plasma tions, most antidepressant treatments resulted in an enhanced concentrations of both steroids returned to physiological baseline serotonin (5-HT) neurotransmission through various mechanisms levels following a successful treatment with selective serotonin of action (Blier & de Montigny, 1994; Owens, 1996). Interestingly, reuptake inhibitors (SSRI), after a period of 30 days (Romeo et al., numerous studies indicate that steroids modulate gene expression 1998; Ströhle et al., 2000). Griffin and Mellon (1999) have re- and functional activity of different components of the 5-HT system ported that the SSRI fluoxetine increases the affinity for its (Bethea et al., 2000). Steroids, particularly, modulate the electrical substrates of 3a-hydroxysteroid dehydrogenase (3a-HSD), the activity of dorsal raphe nucleus 5-HT neurons. A 7-day adminis- enzyme that catalyzes the reduction of the dihydroxy-reduced tration of 5b-pregnane-3,20-dione (5b-DHP), ALLO or DHEA metabolites of PROG and deoxycorticosterone (DOC). The levels of increased the firing activity of 5-HT neurons in female rats pregnanolone and ALLO were found to be lower in the CSF and (Robichaud and Debonnel, 2004). Testosterone and 17b-estradiol, plasma of depressed patients as compared within healthy volun- in both male and female rats, also increased the 5-HT neuronal teers. Moreover, successful antidepressant treatment resulted in firing activity (Robichaud and Debonnel, 2005). Together, these increases up to normal levels and proportional to the mood data strongly suggest a reciprocal modulation between the 5-HT improvement (Romeo et al., 1998; Uzunova et al., 1998; Ströhle system and neuroactive steroids, mainly involving the dihydroxy- et al., 1999, 2000). In animal models, injection of fluoxetine or reduced metabolites of PROG and/or DHEA. paroxetine to male rats also resulted in a rapid increase in the brain Trilostane is an effective inhibitor of the conversion of content of ALLO and a concomitant decrease in 5a-DHP, without D5-3b-hydroxysteroids into the corresponding D4-3-ketosteroids. It any change in PREG, PROG or DHEA (Uzunov et al., 1996). Moreover, acts as a competitive inhibitor of 3b-hydroxysteroid dehydrogenase the role of neuroactive steroids on mood is particularly important (3b-HSD) activity with a Ki of 230 nM (Potts et al., 1978). The in females, due to the natural fluctuations in estrogens and PROG concentrations of PREG and PROG are largely and significantly levels, with particular impacts during menstrual cycle or gestation increased by trilostane (Young et al., 1994). In adrenalectomized/ in predisposed individual. Indeed, women suffer twice more from castrated male rats, trilostane significantly increased the brain major depression than men (Weissman and Olfson, 1995). For concentration of PREG, while it significantly decreased the brain instance, women suffering from premenstrual dysphoric disorders concentration of PROG, as expected from a substrate-to-product showed a blunted response to ALLO, responsible for a diminished relationship. This observation confirmed that PROG is a neuro- functional activity of GABAA receptors (Girdler and Klatzkin, 2007). steroid as defined by Baulieu (Jung-Testas et al., 1989; Corpéchot Moreover, estrogens participate in the modulation of depression et al., 1993; Young et al., 1994). associated with the endocrinal changes along the life of women A treatment with trilostane is therefore expected to directly (Stahl, 1998; Robinson, 2001). In depressive women, therapeutic modulate the brain and circulating levels in PROG, its reduced treatments with estrogens decrease depressive symptoms and metabolites, and PREG and its metabolites. Consequently, a change improve the effects of antidepressants including imipramine, ser- in the activity of GABAA and 5-HT systems must be expected, that traline, and fluoxetine (Amsterdam et al.,1999; Halbreich and Kahn, could lead to modulations of the antidepressant response in 2001; Oppenheim, 1983; Robinson, 2001; Schneider et al., 1997). In pathological conditions. We have indeed reported that trilostane rodents submitted to the FST, estrogen co-treatment facilitated that showed antidepressant-like effect in mice submitted to the forced action of antidepressants like venlafaxine, fluoxetine or desipra- swimming test (FST) and anxiolytic effect in the elevated plus-maze mine, by shortening their onset of action (Estrada-Camarena et al., and black-and-white exploration box (Espallergues et al., 2009). 2004, 2008). Chronic treatment with estrogens affected the Trilostane also reduced the increase in plasma corticosterone and metabolism of antidepressants by interacting with monoamine adrenocorticotropin (ACTH) levels provoked by a 15-min duration oxidase (Holschneider et al., 1998; Ma et al., 1995) or cytochrome forced swimming stress in mice, showing direct effect on HPA axis

4a,5-epoxy-17b-hydroxy-3-oxo-5a-androstane-2a-carbonitrile (trilostane) was 2.4. Adrenalectomy/castration (AdX/CX) provided by CLL Pharma. Trilostane was suspended in pure sesame oil and injected subcutaneously (s.c.) in a volume of 100 ml per 20 g of body weight. The dose-range AdX/CX was performed as previously described by us and validated by used for the in vivo injection of trilostane was 6.3e50 mg/kg. The steroids: measuring the steroid levels in AdX/CX mice (Urani et al., 2001). Brieﬂy, animals 3b-hydroxy-pregn-5-en-20-one (pregnenolone, PREG), pregnenolone sulfate were anaesthetized with an intramuscular injection of mixed xylazine 10 mg/kg

Fig. 1. Experimental procedures. (a) In experiments measuring the antidepressant activity of trilostane, animals were submitted to two FST sessions and trilostane was injected before the second session. (b) In biochemical experiments, two groups were considered, non-stressed and stressed animals. The latter went to two 15-min duration FST sessions before sacriﬁce. (c) In co-administration experiments (trilostane þ steroids or antidepressants), the treatments were applied before the second FST session.

(Rompun, Bayer, 2%) and ketamine 80 mg/kg (Imalgene, Merial, 10%). Both adrenal d4 and THDOC-d3 e were used as internal standards to quantify PREG, ALLO, EPI- glands were removed, through incisions in the back of the animal, just below the ALLO and THDOC, respectively. None of the deuterated internal standard has breast ribs (AdX). The skin was sutured. Both testes were ligatured and cut, through a significant amount of their respective steroid-d0 analogs. The procedure was an incision in the scrotum (CX). Animals recovered within a couple of hours from suitable for measuring concentrations of endogenous unconjugated neuroactive surgery. After AdX/CX, drinking tap water was replaced by a saccharose 1%, NaCl steroids in mouse frontal cortex and hippocampus. The limits of quantification 0.9% solution. Animals were used for behavioral experiments at least six days after defined as the lowest amount of steroid measured by GC/MS with a minimum of surgery. error of 20% in triplicate standard samples (Vallée et al., 2000) ranged from 0.3 to 0.6 pg in brain tissue depending on the steroid. 2.5. Forced swimming test 2.9. Statistical analyses Each mouse was placed individually in a glass cylinder (diameter 12 cm, height 24 cm) filled with water at a height of 12 cm. Water temperature was maintained at Data were expressed as mean S.E.M. and analyzed using one-way ANOVA 22e23 C, a value that maintained a high level of immobility without hypothermia- (F value), followed by a Dunnett’s test, or two-way ANOVA, with trilostane and stress induced stress as obtained with lower temperature. The animal was forced to swim as independent factors, as concerns the monoamine dosages. F values less than 1 for 15 min on day 1 (pre-test session) and 6 min on day 2 (test session). The test could not reach significance, whatever the degree of freedom on samples and session was recorded by a CCD camera connected to a computer and the quantity of population could be, and were thus indicated as such. The level of statistical Ò movement was analyzed using the Videotrack software (Viewpoint, Champagne- significance was p < 0.05, but p < 0.10 suggesting notable trends were indicated in au-Mont-d’Or, France). Two levels of pixel changes were analyzed to discriminate the text. between immobility and struggling and between struggling and swimming. Since only immobility is unequivocally regarded as a parameter of behavioral despair, 3. Results other behaviors were not analyzed. When the FST was used as a stressor, the day session duration was 15 min. 3.1. The antidepressant-like effect of trilostane is abolished 2.6. Measure of monoamine levels in AdX/CX animals

Animals were killed by decapitation. The brains were quickly removed and the We first examined the antidepressant-like effect of the 3b-HSD frontal cortex and hippocampus were dissected out on an ice-cold glass plate, as in previous reports (Noda et al., 1997; Mamiya et al., 1998). Each section was rapidly inhibitor in AdX/CX mice. Trilostane was administered s.c. in sham- frozen and stored in a deep freezer at 80 C until assayed. The levels of mono- operated mice, twice, 16 h and 2 h before the forced swimming test amines and their metabolites were determined with an HPLC system equipped with session. The compound induced a dose-dependent decrease of an electrochemical detector (HTEC-500, Eicom, Kyoto, Japan). Briefly, each frozen immobility (F(4,28) ¼ 2.88, p < 0.05; Fig. 2a), with a significant brain sample was weighed and homogenized with an ultrasonic processor in 350 ml < e of 0.2 M perchloric acid containing isoproterenol as an internal standard. The difference at 25 mg/kg ( 86 s, p 0.05). The dose response curve homogenates were placed on ice for 30 min and centrifuged at 20,000 g for 15 min at was however U-shaped and the dose of 50 mg/kg was ineffective. In 4 C. The supernatants were mixed with 1 M sodium acetate to adjust the pH to 3 AdX/CX animals, the compound appeared ineffective whatever the and injected into the HPLC system equipped with a reversed-phase ODS column dose (F < 1; Fig. 2b). In particular, significant differences were (Eicompak SC-ODS; 3.0 150 mm; Eicom) and the electrochemical detector. The measured as compared with sham-operated animals, at the doses column temperature was maintained at 25 C, and the detector potential was set at þ750 mV. The mobile phase was 0.1 M citric acid and 0.1 M sodium acetate, pH of 12.5 and 25 mg/kg (Fig. 2b). 3.6, containing 17% methanol, 180 mg/l sodium-L-octanesulfonate, and 5 mg/l EDTA, and the flow rate was set at 0.5 ml/min. 3.2. Effect of trilostane on peripheral and brain levels of steroids

2.7. Measure of plasma and hippocampus PROG levels We investigated the effect of trilostane, administered in the Animals were killed by decapitation and blood samples were collected on 6.3e50 mg/kg dose-range, on the circulating levels of PROG. In 1 mg/ml ethylenediaminetetraacetic acid (Sigma-Aldrich), centrifuged 15 min at non-stressed mice (Fig. 3a), trilostane induced a dose-dependent 4000 g at 4 C. Plasma samples were stored at 20 C, until assayed. They were increase in PROG (F(4,37) ¼ 5.69, p < 0.01), with significant effects assayed in 100 ml unextracted samples. The hippocampus was dissected out, steroid in the 12.5e50 mg/kg dose-range (þ150% at 25 mg/kg). Stress- extracted using ethyl acetate/isooctane 1:1 and sonication, followed by a centrifu- fi gation 5 min at 3000 g at 4 C, drying of the supernatant and resuspension in induced a highly signi cant increase in circulating levels of PROG methanol 100%. Hippocampus PROG was assayed in 200 ml methanol resuspended (þ340%, t(15) ¼ 4.20, p < 0.001 for vehicle solution-treated groups, extracts. Samples were then assayed using the radioimmunoassay kit according to Fig. 3b). However, the trilostane treatment failed to affect PROG ’ the manufacturer s instructions (DSL active RID, DSL-France, Cergy-Pontoise, France). levels in stressed mice, whatever the dose (F < 1, Fig. 3b). The lowest detectable PROG concentration was 2 pg/ml and the intra- and inter- assay coefficients of variation were 6% and 11%, respectively. The effect of trilostane on central level of PROG was then eval- uated using a simplified organic extraction/RIA dosage method in 2.8. Measure of steroid levels by gas chromatography/mass spectroscopy (GC/MS) the hippocampus, in control and stressed animals (Fig. 3c). Signif- icant effects were measured for stress (F(1,45) ¼ 26.1, p < 0.0001), Animals were killed by decapitation. Brains were quickly removed and the ¼ < trilostane (F(1,45) 42.2, p 0.0001) and the interaction frontal cortex and hippocampus were dissected out and frozen at 80 C until ¼ < fi assayed. Brain levels of PREG, ALLO, EPIALLO and THDOC were determined by (F(1,45) 23.6, p 0.0001). Indeed, stress highly signi cantly isotope dilution combined with GC/MS according to the protocol previously increased PROG level whereas trilostane decreased by 30% PROG described (Vallée et al., 2000; George et al., 2010). The method was validated in levels in non-stressed mice and blocked the stress-induced terms of sensitivity, accuracy and precision for these steroids. Briefly, steroids were increase (Fig. 3c). These observations were coherent with its extracted from brain tissue by a simple solid-phase extraction (SPE) method using 3b-HSD inhibition activity. The effects of trilostane on the levels of reverse-phase C18 columns (Vallée et al., 2000). The evaluation of recovery, repro- ducibility, and generation of calibration curves were performed for each analysis. other steroids were investigated using a GC/MS quantification The response was linear in the range of 0e32 pg for all the steroids analyzed technique in two brain structures involved in the response to stress (R2 ¼ 0.999) and intra- and inter-assay coefficients of variation ranged from 4 to 14% and antidepressants, namely the frontal cortex and hippocampus, and 12 to 20%, respectively, depending on the steroid analyzed. The method uses according to previously published method (Vallée et al., 2000; negative chemical ionization with GC/MS and involves the formation of penta- fi fluorobenzyloxime/trimethylsilyl ether derivatives of the steroid fraction from brain George et al., 2010). The method did not permit the identi cation extracts to enhance the mass spectrometric analysis. Mass spectra were acquired of PROG, but the precursor PREG and the tetrahydroxy-reduced with a GCMS-QP2010 mass spectrometer (Shimadzu, Kyoto, Japan). The mass metabolites of PROG and 11-deoxycorticosterone, ALLO, EPIALLO spectrometer was operated in a selective ion-monitoring mode, and the concen- and THDOC, were identified. In both structures, trilostane tration of each steroid was calculated by linear regression of the peak area corre- (25 mg/kg) induced significant increases in PREG levels. Trilostane sponding to the diagnostic ion (m/z) with the highest intensity. fi The isotope dilution method was used to achieve accurate quantification failed to signi cantly affect metabolites levels, although a trend to (Alomary et al., 2001). Deuterated analogs of steroids e PREG-d4, ALLO-d4, EPIALLO- diminution was noted particularly for EPIALLO (Table 1).

Fig. 2. Doseeresponse effect of the 3b-HSD inhibitor trilostane in the forced swim test Fig. 3. Effects of stress and trilostane treatment on progesterone (PROG) levels in the in AdX/CX Swiss mice: (a) sham-operated control animals and (b) AdX/CX mice. Mice plasma (a, b) and hippocampus (c). Trilostane (25 mg/kg) or the vehicle solution were submitted to a 15 min pre-test forced swimming session and 24 h after to a 6 min (sesame oil, V) was administered s.c. 16 and 2 h before sacriﬁce. Stressed animals were test session. Trilostane or the vehicle solution (sesame oil, V) was administered s.c. 16 submitted to a 15-min duration forced swimming immediately before sacriﬁce. The and 2 h before the test session. The number of animals per group was 5e6 in (a), 9 in number of animals per group was 7e8 in (a), 8e10 in (b), 12e13 in (c). Group (b). Group comparisons: *p < 0.05 vs. V-treated group; #p < 0.05, ##p < 0.01 vs. sham- comparisons: *p < 0.05, **p < 0.01, ***p < 0.001 vs. the V-treated group, ###p < 0.001 operated group; Dunnett’s test. vs. the stressed group, Dunnett’s test.

3.3. Effect of co-treatments between trilostane and neuroactive steroids in the FST in a dose-dependent manner, with significant effects at 2e8 mg/kg (Fig. 4d). Co-administration of trilostane (25 mg/kg) resulted in PREGS, administered in the 5e40 mg/kg s.c. dose-range, failed a significant decrease in immobility in vehicle solution-treated to decrease the immobility in the forced swimming test (Fig. 4a). animals, but failed to affect the doseeresponse curve of allopreg- ¼ < < < Co-administration of trilostane (25 mg/kg) resulted in a significant nanolone (F(4,42) 3.70, p 0.05 for ALLO, F 1 for trilostane, F 1 decrease in immobility but did not change the doseeresponse for the interaction; Fig. 4d). curve of PREGS (F < 1 for PREGS and trilostane, F(3,44) ¼ 1.63, p > 0.05 for the interaction; Fig. 4a). DHEAS, administered in the 5e20 mg/kg s.c. dose-range, decreased significantly the immobility Table 1 in a dose-dependent but U-shaped manner, the most active dose Effects of trilostane on frontal cortex and hippocampus levels of pregnenolone (PREG) and progesterone and 11-deoxycorticosterone metabolites, 3a,5a-tetrahy- being 10 mg/kg (Fig. 4b). Co-administration of trilostane (25 mg/kg) droprogesterone (ALLO), 3b,5a-tetrahydroprogesterone (EPIALLO) and 3a,5a-tetra- shifted the DHEAS doseeresponse curve to the left, the most active hydrodeoxycorticosterone (THDOC). dose becoming 5 mg/kg. This was indicated by a highly significant ¼ < Steroid Frontal cortex Hippocampus ANOVA effect on the interaction parameter (F(3,54) 4.28, p 0.01 (pg/mg tissue) Veh Trilostane Veh Trilostane for DHEAS, F < 1 for trilostane, F(3,54) ¼ 7.09, p < 0.001 for the (25 mg/kg sc) (25 mg/kg sc) interaction; Fig. 4b). PROG, administered in the 10e40 mg/kg s.c. dose-range, failed to decrease the immobility (Fig. 4c). PREG 1.15 0.27 3.21 0.94** 0.69 0.17 2.85 0.87* fi ALLO 0.81 0.18 0.68 0.23 0.88 0.19 0.74 0.15 Co-administration of trilostane (25 mg/kg) resulted in a signi cant EPIALLO 1.04 0.24 0.59 0.26 0.77 0.20 0.38 0.14 decrease in immobility but did not change the doseeresponse THDOC 0.80 0.26 0.56 0.15 0.30 0.04 0.39 0.05 curve of PROG (F < 1 for PROG, F ¼ 14.41, p < 0.001 for tri- (3,61) Trilostane (25 mg/kg s.c.) or sesame oil (Veh) was administered 16 h and 2 h before < lostane, F 1 for the interaction; Fig. 4c). ALLO, administered in the sacrifice. The number of animals per group was n ¼ 6. *p < 0.05, **p < 0.01 vs. Veh- 1e8 mg/kg s.c. dose-range, decreased significantly the immobility treated mice, ManneWhitney’s test.

Fig. 4. Effects of co-treatments between trilostane and neuroactive steroids in the forced swimming test: (a) pregnenolone sulfate (PREGS); (b) dehydroepiandrosterone sulfate (DHEAS); (c) progesterone; (d) allopregnanolone. Trilostane (25 mg/kg) or the vehicle solution (sesame oil, Veh) was administered s.c. 16 and 2 h before the test session. The steroid or the vehicle solution (sesame oil, V) was injected s.c. 30 min before the test session. The number of animals per group was 8e12 in (a), 7e9 in (b), 8e10 in (c), 4e7 in (d). Group comparisons: *p < 0.05, **p < 0.01 vs. the (Veh þ V)-treated group, Dunnett’s test.

3.4. Effect of trilostane on brain levels of monoamines p < 0.01; Fig. 5c). The increases remained modest, þ19% for 5-HT and metabolites and þ32% for DA. No effect was measured for the trilostane treatment or the stress trilostane interaction on monoamine The effects of the trilostane treatment and swimming stress on levels or turnover ratios. monoamine levels were examined in the frontal cortex (Table 2a) In the hippocampus (Table 2b), the ANOVA revealed a signifi- and hippocampus (Table 2b) of non-stressed mice and stressed cant effect stress of for 5-HT, NE, MHPG, DOPAC and HVA. Post-hoc animals, i.e., exposed to forced swimming during 15 min on days 1 comparisons showed that 5-HT levels were significantly decreased and 2. Trilostane (25 mg/kg) or vehicle was administered 16 h and and MHPG levels significantly increased (Table 2b). These varia- 2 h before sacrifice. A measure of the immobility duration showed tions resulted in different effects on monoamine turnover ratios 222 8 s for vehicle solution-treated animals and 114 11 s for (Fig. 5def). The ANOVA showed a significant stress effect for the trilostane-treated mice (n ¼ 8, p < 0.001). 5-HT turnover (F(1,28) ¼ 5.53, p < 0.05; Fig. 5d). Significant tri- In the frontal cortex (Table 2a), the stress-induced significant lostane effects were measured for both the 5-HT turnover changes in monoamine levels: the contents in neurotransmitter (F(1,28) ¼ 8.91, p < 0.01; Fig. 5d) and NE turnover (F(1,28) ¼ 8.01, serotonin (5-HT) and norepinephrine (NE) were highly signifi- p < 0.01; Fig. 5e). The stress trilostane interaction failed to show cantly decreased. The ANOVA revealed a significant effect of stress a significant effect on 5-HT turnover (F(1,28) ¼ 3.09, p ¼ 0.09; for the dopamine (DA) metabolite 3,4-dihydroxyphenylacetic Fig. 5d) but appeared significant for the NE turnover acid (DOPAC) and for the NE metabolite 4-hydroxy-3- (F(1,28) ¼ 12.08, p < 0.01; Fig. 5e). The group comparisons revealed methoxyphenylglycol (MHPG), although post-hoc comparisons that only the trilostane-treated stressed mice group showed did not reach significance (Table 2a). These variations resulted in significant increase in 5-HT and NE turnovers as compared with significant effects on monoamine turnover ratios (Fig. 5aec). The Veh-treated non-stressed controls (Fig. 5d,e). These increases were ANOVA showed a significant stress effect for the 5-HT turnover modest, þ32% and þ13% respectively. No effect was measured on (F(1,23) ¼ 9.78, p < 0.01; Fig. 5a) and DA turnover (F(1,23) ¼ 8.53, DA turnover (Fig. 5f).

Table 2 treatment data showed a significant increase of the anti-immobility Effects of trilostane on the frontal cortex and hippocampus levels of monoamines effect (p < 0.05; Fig. 6d). and metabolites in mice exposed, or not, to forced swimming stress. Monoamine Non-stressed mice Stressed mice 4. Discussion (pg/mg tissue) Veh Trilostane Veh Trilostane (25 mg/kg sc) (25 mg/kg sc) In the present study, we extended the analysis of the a) Frontal cortex antidepressant-like effect induced by the 3b-HSD inhibitor trilos- 5-HT 17.07 0.50 16.68 0.47 14.16 0.60** 13.76 0.25** tane, which was previously reported in mice submitted to the FST 5-HIAA 5.23 0.18 5.58 0.21 5.17 0.26 4.89 0.15 (Espallergues et al., 2009). We analyzed the impact of a trilostane NE 23.07 0.42 22.54 0.80 19.27 0.73** 19.14 0.57** Epi 0.86 0.06 1.03 0.06 1.14 0.08 0.94 0.07 treatment on steroid and monoamine levels and whether it could MHPG 3.29 0.09 3.28 0.12 2.75 0.38 2.62 0.28 affect the pharmacological efficacy of exogenously administered DA 3.22 0.21 2.66 0.43 3.03 0.43 2.90 0.13 steroids and antidepressants, acting as monoamine transporters DOPAC 1.07 0.10 1.03 0.04 1.26 0.09 1.15 0.04* inhibitors. Our purpose was to determine if the compound could HVA 2.01 0.16 1.91 0.09 2.25 0.17 2.00 0.14 putatively induce a direct antidepressant-like effect through b) Hippocampus 5-HT 22.22 0.63 22.11 0.52 19.49 0.59** 19.16 0.92** modulation of neuroactive steroid levels and modulation of 5-HIAA 11.00 0.72 11.83 0.91 10.12 0.73 13.01 0.97 monoaminergic systems. The results confirmed the antidepressant NE 18.95 0.89 19.70 0.45 16.70 0.99 15.89 0.63* potential of the 3b-HSD inhibitor but outlined the limited impact Epi 0.66 0.03 0.81 0.12 0.78 0.03 0.82 0.04 on neurosteroid levels and mainly modulatory effect in combina- MHPG 1.94 0.06 1.97 0.06 1.54 0.17* 1.87 0.12 DA 1.26 0.12 1.34 0.11 1.36 0.06 1.17 0.05 tion studies with known antidepressants. DOPAC 0.51 0.05 0.49 0.04 0.63 0.04 0.60 0.04 HVA 1.03 0.10 1.05 0.09 0.90 0.04 0.86 0.05 4.1. The trilostane treatment affected neuroactive steroid levels in Trilostane (25 mg/kg s.c.) or sesame oil (Veh) was administered 16 h and 2 h before brain and periphery sacrifice. Animals were exposed to forced swimming stress 15 min on day 1 and on day 2, immediately before sacrifice. The number of animals per group was n ¼ 7. We first confirmed that, although trilostane paradoxically fi ¼ Two-way ANOVA: in (a), signi cant effects were measured for stress (F(1,23) 39.6, increased circulating PROG levels, it significantly decreased the p < 0.0001 for 5-HT; F ¼ 30.2, p < 0.0001 for NE; F ¼ 5.22, p < 0.05 for (1,23) (1,23) brain levels in PROG, in the hippocampus and frontal cortex. In MHPG; F(1,23) ¼ 5.06, p < 0.05 for DOPAC), but not for the trilostane treatment or the stress trilostane interaction. In (b), significant effects were measured for stress parallel, it increased PREG levels, in accordance with its inhibitory

(F(1,28) ¼ 17.29, p < 0.001 for 5-HT; F(1,28) ¼ 15.67, p < 0.001 for NE; F(1,28) ¼ 5.03, activity on 3b-HSD. These measures confirmed the data presented < ¼ < ¼ < p 0.05 for MHPG; F(1,28) 8.23, p 0.01 for DOPAC; F(1,28) 4.75, p 0.05 for in the pioneer studies by Young et al. (1994, 1996). As trilostane is HVA), but not for the trilostane treatment or the stress trilostane interaction. a competitive inhibitor, an increase in substrate concentrations will Group comparisons: *p < 0.05, **p < 0.01 vs. non-stressed mice, Dunnett’s test. lead to a lifting of the enzymatic inhibition. Trilostane triggers an adrenal hyperstimulation similar in mice to the one observed in 3.5. Effect of co-treatments between trilostane and antidepressants rats and its blockade of the synthesis of CORT directly impacts its in the forced swimming test negative feedback on ACTH secretion. ACTH in turn stimulates the P450scc, which catalyzes the synthesis of PREG from cholesterol In co-treatment studies, trilostane (12.5, 25 mg/kg) was (Young et al., 1994). The resulting increases in the production of co-administered with selected antidepressants (Fig. 6). Sertraline PREG (Potts et al., 1978; this study) generate a lifting of the 3b-HSD and fluoxetine were tested as reference SSRI. Desipramine and inhibition. Therefore, our results confirm the first observations by imipramine were tested as reference tricyclic antidepressants Young et al. (1994), that trilostane increases, not decreases, the acting through the serotonin and norepinephrine transporters. PROG level in plasma. The same authors reported that brain PREGS Sertraline decreased the immobility duration in the 20e40 mg/kg concentrations were not significantly increased by trilostane, but dose-range (Fig. 6a). This effect was increased in trilostane-treated ALLO, barely detectable in control brains, became measurable in the groups, since the ANOVA showed a significant effect for sertraline brain of trilostane-treated mice, at a nanomolar level (Young et al., (F(2,118) ¼ 19.2, p < 0.0001) and trilostane (F(2,118) ¼ 4.48, p < 0.05), but 1996). We could not here confirmed such results but confirmed that not for the interaction (F < 1). In particular, the trilostane (25 mg/kg) trilostane induced a significant increase in brain PREG levels, sug- treatment highly significantly increased the anti-immobility effect gesting that DHEA levels, not measurable in the present study, induced by the highest dose of sertraline (40 mg/kg; Fig. 6a). Similar could also be affected (Young et al., 1994). We observed no signif- data were obtained with fluoxetine. The drug decreased the immo- icant changes in the tetrahydroxy-reduced metabolites of PROG bility duration in the 30e60 mg/kg dose-range (Fig. 6b). This effect and DOC, and a tendency to decreased brain EPIALLO levels. These was increased by trilostane (25 mg/kg), since a significant effect was observations are coherent with the inhibitory effect of trilostane on measured for fluoxetine (F(2,113) ¼ 25.1, p < 0.0001) and trilostane 3b-HSD enzymatic activity, but they may minimize its (F(2,118) ¼ 3.26, p < 0.05), but not for the interaction (F < 1). The tri- antidepressant-like potency. Indeed, the trilostane treatment lostane (25 mg/kg) treatment significantly increased the anti- affected not only the excitatory neurosteroids PREGS and DHEAS, immobility effect induced by the highest dose of fluoxetine (Fig. 6b). but also the steroids acting as GABAergic positive modulators, Desipramine decreased the immobility duration in the 0e40 including ALLO, a steroid known to play an important role in mg/kg dose-range (Fig. 6c), however trilostane only tended to antidepressant response. increase this effect. The ANOVA showed a significant effect for desipramine (F(2,169) ¼ 31.8, p < 0.0001), but only a tendency for 4.2. PREG, PROG and ALLO differentially impact depressive states trilostane (F(2,169) ¼ 2.54, p ¼ 0.08) and no effect for the interaction (F < 1). Imipramine also decreased the immobility duration in the Brain PREG levels have been linked to depressive symptoms in 0e40 mg/kg dose-range (Fig. 6d). A significant effect was measured humans, since CSF levels of PREG have been found to be decreased for imipramine (F(2,119) ¼ 52.7, p < 0.0001), but not for trilostane in depressed patients (George et al., 1994). In rodents, PREGS (F(2,119) ¼ 2.17, p ¼ 0.12) or the interaction (F < 1). Noteworthy, a two- reduced the immobility time in the FST (Reddy et al., 1998; Urani column comparison between the (imipramine 40 mg/kg þ vehicle et al., 2001). Moreover, fluoxetine or olanzapine have been shown solution) and (imipramine 40 mg/kg þ trilostane 25 mg/kg) to increase PREG concentration in the hippocampus of rats at doses

Fig. 5. Effects of trilostane treatment and forced swimming stress on monoamine turnover in the mouse frontal cortex (aec) and hippocampus (def): 5-HT (a, d), NE (b, e) and DA (c, f). Data were calculated from Table 2a,b. Group comparisons: *p < 0.05, **p < 0.01 vs. non-stressed mice; ##p < 0.01 vs. vehicle-treated mice; NewmaneKeuls’ test.

mediating their antidepressant-like effects (Marx et al., 2006). administration in some studies (Dennerstein et al., 1980; Baker Interestingly, the drugs also increased hippocampus levels of ALLO et al., 1995). However, other studies led a lack of effect as and PROG, and plasma levels in DOC and CORT (Marx et al., 2006), compared with placebo treatment (Freeman et al., 1995; Vanselow suggesting that their effect could be related to a general boost in et al., 1996). So far, no definitive conclusion can be drawn regarding neurosteroid syntheses, putatively through modulation of the the therapeutic potential of PROG for the prevention of these oxydoreductase activity of 3a-HSD (Griffin and Mellon, 1999). subtypes of depression. Moreover, due to its s1 receptor antago- Contradictory effects for PROG have been reported in depres- nistic action, PROG antagonized the antidepressant effects of sion. In ovarectomized mice submitted to the tail suspension test DHEAS and PREGS in mice submitted to the FST (Reddy et al., 1998; (Bernardi et al., 1989) and rats submitted to the FST (Martínez-Mota Urani et al., 2001). The antidepressant efficacy of synthetic s1 et al., 1999), PROG decreased the duration of behavioral despair. In receptor agonists was indeed inversely correlated to the endoge- this study, we confirmed previously published data (Urani et al., nous PROG levels in the mouse hippocampus (Urani et al., 2001). 2001) showing that exogenous administration of PROG failed to These observations suggested that PROG may induce a pharmaco- decrease the immobility duration in mice submitted to the FST. In logical blockade on specific endogenous mood regulatory systems humans, exogenously administered PROG has been tested in post- and such adverse effects may indeed limit the beneficial impact of partum depression and premenstrual dysphoric disorder, a mood PROG-based therapies in depression. disorder that shares some clinical features with major depression. On the contrary, several studies have suggested an important An improvement of mood could be observed after PROG pathophysiological role for tetrahydroxy-reduced PROG metabolites

Fig. 6. Effects of co-treatments between trilostane and antidepressants: sertraline (a), fluoxetine (b), desipramine (c), and imipramine (d). Trilostane (25 mg/kg) or the vehicle solution (sesame oil, Veh) was administered s.c. 16 and 2 h before the test session. The antidepressant or the vehicle solution (water, V) was injected i.p. 30 min before the test session. The number of animals per group was 9e18 in (a), 11e19 in (b), 12e26 in (c), 12e15 in (d). Group comparisons: *p < 0.05, **p < 0.01 vs. V-treated group; #p < 0.05, ##p < 0.01 vs. Veh-treated group; Dunnett’s test. in depressive disorders. The normalization of their endogenous probably as a consequence of HPA axis overdrive (Ströhle et al., levels may contribute to the therapeutic effects of various antide- 2000). Treatment with fluoxetine counteracted these modifica- pressants. ALLO showed an antidepressant-like potential in mice tions in tetrahydroxy-reduced steroid metabolites levels in plasma submitted to the FST (Khisti et al., 2000). Alterations of ALLO levels and CSF (Romeo et al., 1998; Uzunova et al., 1998). Fluoxetine also have been detected in different paradigms of depression-related decreased THDOC levels back to physiological plasma concentra- behaviors in rodents. For instance, protracted social isolation in tions (Ströhle et al., 2000). As the 3a-HSD also catalyzes the mice is accompanied by decreases in ALLO levels and its precursor reduction of the precursor 5a-dihydroDOC toTHDOC, these findings 5a-DHP (Matsumoto et al., 1999; Dong et al., 2001) in the frontal have suggested that antidepressants do not generally shift the cortex. In rats subjected to social isolation immediately after activity of the 3a-HSD toward the reductive direction (Ströhle et al., weaning, decreased levels of ALLO and THDOC have been measured 1999). We confirmed here that exogenous administration of ALLO in the cortex, hippocampus and plasma (Serra et al., 2000). After significantly decrease the duration of behavioral despair of mice in olfactory bulbectomy in rats, decreased ALLO levels have been the FST. measured in the amygdala and frontal cortex (Uzunova et al., 2003). Moreover, restoration of physiological levels in ALLO has been 4.3. The trilostane treatment differentially affected the shown in humans and animal models of depression to accompany antidepressant activity of neurosteroids the therapeutic efficacy of antidepressants. In humans, plasma and CSF levels of ALLO and 3a,5b-THP are found to be decreased in The trilostane treatment failed to significantly affect the patients suffering from major depression (Romeo et al., 1998; concentrations of the reduced PROG and DOC metabolites, since Uzunova et al., 1998). There was, in parallel, an increase of EPI- only a downward trend is observed. ALLO or THDOC may therefore ALLO (Romeo et al., 1998) that may act as a functional antagonist for not contribute to its antidepressant-like effect, suggesting that the GABA-agonistic steroids like ALLO (Eser et al., 2006). Moreover, trilostane did not interact with 3a-HSD activity. In turn, we an increase of THDOC has been observed in depressed patients, observed that a co-treatment between trilostane and ALLO showed

Please cite this article in press as: Espallergues, J., et al., The antidepressant-like effects of the 3b-hydroxysteroid dehydrogenase inhibitor trilostane in mice is related to changes in neuroactive steroid and monoamine levels, Neuropharmacology (2011), doi:10.1016/ j.neuropharm.2011.09.005 10 J. Espallergues et al. / Neuropharmacology xxx (2011) 1e11 a mutually exclusive effect: trilostane decreased immobility in the such effect, and its specific estrogenic effect (Espallergues et al., absence of ALLO but, when the steroid was injected at its active 2011), trilostane moderately but significantly facilitated 5-HT doses, only its own effect was measurable. The trilostane neurotransmission and the action of SSRI. Trilostane appears antidepressant-like effect appeared thus to involve modulation of therefore as an interesting alternative to antiglucocorticoid drugs in neurosteroid contents through a mechanism unrelated to that of depression. Indeed, although modulators of steroid sulfotransfer- fluoxetine and limited to a potentiation of the PREG/S and DHEA/S ase, sulfatase or aromatase failed to show any potential in depres- effects. Indeed, although the trilostane co-treatment failed to sion, modulators of steroid hydroxylases or cholesterol side-chain markedly impact the doseeresponse effect of PREGS, a clear cleavage like ketoconazole, metyrapone and aminoglutethimide potentiation of DHEAS effect was measured with a left-shift in the have been shown to allow significant improvement in major active doses. Young et al. (1994, 1996) have previously reported that depression in humans (Murphy, 1997; Rupprecht et al., 1998; DHEA levels are rapidly and efficiently increased after trilostane Wolkowitz and Reus, 1999; Wolkowitz et al., 1999b). Similarly, treatment in mice, confirming that the main effect of the inhibitor trilostane may thus offer a new opportunity to increase the ther- on neurosteroids is a potentiation of DHEA/S effects. It is clear apeutic efficacy of SSRI, by directly modulating neurosteroid levels. however, from this pharmacological study, that the lack of impact of trilostane on reduced PROG metabolites may contribute to limit its behavioral efficacy. In terms of a direct impact on neurosteroid References levels, the compound acts as a modulator of steroids acting as Alomary, A.A., Fitzgerald, R.L., Purdy, R.H., 2001. Neurosteroid analysis. Int. Rev. negative GABAA receptor modulators (PREGS, DHEAS) and not Neurobiol. 46, 97e115. through the highly effective positive GABAA receptor modulators Amsterdam, J., García-España, F., Fawett, J., Quitkin, F., Reimherr, F., Rosenbaum, J., (PROG, ALLO). Beasley, C., 1999. Fluoxetine efficacy in menopausal women with and without estrogen replacement. J. Affect Disord. 55, 11e17. Baker, E.R., Best, R.G., Manfredi, R.L., Demers, L.M., Wolf, G.C., 1995. Efficacy of 4.4. The trilostane treatment impacts monoamine turnover progesterone vaginal suppositories in alleviation of nervous symptoms in and facilitates SSRI activity patients with premenstrual syndrome. J. Assist. Reprod. Genet. 12, 205e209. Baulieu, E.E., 1981. Steroid hormones in the brain: several mechanisms? In: Fuxe, K., The modification in GABA receptor function induced by Gustafson, J.A., Wettenberg, L. (Eds.), Steroid Hormone Regulation of the Brain. A Pergamon Press, Oxford, pp. 3e14. increases in brain PREGS and DHEAS levels triggered by trilostane is Belelli, D., Lan, N.C., Gee, K.W., 1990. Anticonvulsant steroids and the GABA/ expected to impact monoaminergic, and particularly 5-HT, neuro- benzodiazepine receptor-chloride ionophore complex. Neurosci. Biobehav. Rev. e transmission systems. We observed here that the forced swimming 14, 315 322. Bermack, J.E., Debonnel, G., 2001. Modulation of serotonergic neurotransmission by stress decreased 5-HT contents in the hippocampus and cortex, an short- and long-term treatments with sigma ligands. Br. J. Pharmacol. 134, effect related to increased release of the neurotransmitter. The 5-HT 691e699. turnover, an index of neurotransmitter metabolism (Schwarting Bernardi, M., Vergoni, A.V., Sandrini, M., Tagliavini, S., Bertolini, A., 1989. Influence of ovariectomy, estradiol and progesterone on the behavior of mice in an and Huston, 1992), was however increased only in the cortex. The experimental model of depression. Physiol. Behav. 45, 1067e1068. trilostane treatment resulted in modest increases in 5-HT and NE Bethea, C.L., Mirkes, S.J., Shively, C.A., Adams, M.R., 2000. Steroid regulation of turnovers in the hippocampus that could appear coherent with its tryptophan hydroxylase protein in the dorsal raphe of macaques. Biol. Psychi- atry 47, 562e576. antidepressant-like activity, but remained limited. Indeed, in the Bitran, D., Purdy, R.H., Kellogg, C.K., 1993. Anxiolytic effect of progesterone is cortex, the trilostane treatment failed to further increase the 5-HT associated with increases in cortical allopregnanolone and GABAA receptor turnover or to affect the NE turnover. Co-administration studies function. Pharmacol. Biochem. Behav. 45, 423e428. Blier, P., de Montigny, C., 1994. Current advances and trends in the treatment of with SSRI or tricyclic antidepressants, however, clearly showed that depression. Trends Pharmacol. Sci. 15, 220e226. trilostane allowed additive effect with the drugs. The impacts on Bloch, M., Schmidt, P.S., Danaceau, M.A., Adams, L.F., Rubinow, D.R., 1999. Dehy- SSRI efficacy were significant. Both trilostane-induced modulation droepiandrosterone treatment of midlife dysthymia. Biol. Psychiatry 45,1533e1541. of brain DHEA/S levels or its estrogenic action could be involved in Browne, E.S., Wright, B.E., Porter, J.R., Svec, F., 1992. Dehydroepiandrosterone: antiglucocorticoid action in mice. Am. J. Med. Sci. 303, 366e371. a facilitation of 5-HT systems. First, through its GABAergic or s1 Corpéchot, C., Robel, P., Axelson, M., Sjovall, J., Baulieu, E.E., 1981. Characterization effect, DHEAS could increase 5-HT neurotransmission. For instance, and measurement of dehydroepiandrosterone sulfate in rat brain. Proc. Natl. Acad. Sci. U.S.A. 78, 4704e4707. Bermack and Debonnel (2001) have reported that the s1 ligands þ Corpéchot, C., Young, J., Calvel, M., Wehrey, C., Veltz, J.N., Touyer, G., Mouren, M., 4-IBP, ( )-pentazocine, and DTG after either 2 or 21 days of treat- Prasad, V.V., Banner, C., Sjövall, J., Baulieu, E.E., Robel, P., 1993. Neurosteroids: ment induce a significant effect on the firing activity of 5-HT 3a-hydroxy-5a-pregnan-20-one and its precursors in the brain, plasma, and neurons of the dorsal raphe nucleus, which is increased by more steroidogenic glands of male and female rats. Endocrinology 133, 1003e1009. Crawley, J.N., Glowa, J.R., Majewska, M.D., Paul, S.M., 1986. Anxiolytic activity of an than 50%. Second, the same group has shown that 17b-estradiol, endogenous adrenal steroid. Brain Res. 398, 382e385. and testosterone, increases the firing activity of 5-HT neurons in Dennerstein, L., Spencer-Gardner, C., Gotts, G., Brown, J.B., Smith, M.A., both males and females (Robichaud and Debonnel, 2005). A Burrows, G.D., 1980. Progesterone and the premenstrual syndrome: a double blind crossover trial. Br. Med. J. 290, 1617e1621. possible mechanism could be that 17b-estradiol decreases the Dong, E., Matsumoto, K., Uzunova, V., Sugaya, I., Takahata, H., Nomura, H., 2001. GABAergic tonic inhibition and increases the glutamatergic input Brain 5a-dihydroprogesterone and allopregnanolone synthesis in a mouse on 5-HT neurons. The effects of estrogens on the 5-HT system have model of protracted social isolation. Proc. Natl. Acad. Sci. U. S. A. 98, 2849e2854. Eser, D., Schüle, C., Baghai, T.C., Romeo, E., Rupprecht, R., 2006. Neuroactive steroids indeed been studied extensively and it has particularly been shown in depression and anxiety disorders: clinical studies. Neuroendocrinology 84, that 17b-estradiol altered directly the gene expression and binding 244e254. sites for 5-HT transporters in the dorsal raphe nucleus and other Espallergues, J., Givalois, L., Temsamani, J., Laruelle, C., Maurice, T., 2009. The 3b-hydroxysteroid dehydrogenase inhibitor trilostane shows antidepressant brain regions of ovarectomized rats (Landry and Di Paolo, 2003). properties in mice. Psychoneuroendocrinology 34, 644e659. Espallergues, J., Temsamani, J., Laruelle, C., Urani, A., Maurice, T., 2011. The 4.5. Conclusion antidepressant-like effect of the 3b-hydroxysteroid dehydrogenase inhibitor trilostane involves a regulation of b-type estrogen receptors. Psychopharma- cology 214, 455e463. The present data suggested that a treatment with trilostane Estrada-Camarena, E., Fernández-Guasti, A., López-Rubalcava, C., 2004. Interaction results in a significant modulation of excitatory neurosteroid levels, between estrogens and antidepressants in the FST in rats. Psychopharmacology namely PREG/S and DHEA/S (this study and Young et al., 1994). This 173, 139e145. Estrada-Camarena, E., Rivera, N.M., Berlanga, C., Fernández-Guasti, A., 2008. effect appeared however limited to these steroids and failed to Reduction in the latency of action of antidepressants by 17 beta-estradiol in the affect the tetrahydroxy-reduced metabolites of PROG. Through forced swimming test. Psychopharmacology 201, 351e360.

Freeman, E.W., Rickels, K., Sondheimer, S.J., Polansky, M., 1995. Ineffectiveness of metyrapone administration in depressed patients and healthy volunteers. Biol. progesterone suppository treatment for premenstrual syndrome. J. Am. Med. Psychiatry 44, 912e914. Assoc. 274, 51e57. Schneider, M., Small, G., Hamilton, S., Bystrisky, A., Nemeroff, C., Meyers, B., 1997. George, M.S., Guidotti, A., Rubinow, D., Pan, B., Mikalauskas, K., Post, R.M., 1994. CSF Estrogen replacement and response to fluoxetine in surgically menopausal neuroactive steroids in affective disorders: pregnenolone, progesterone, and women. Am. J. Geriatr. Psychiatry 5, 97e106. DBI. Biol. Psychiatry 35, 775e780. Schwarting, R.K., Huston, J.P., 1992. Behavioural concomitants of regional changes in George, O., Vallée, M., Vitiello, S., Le Moal, M., Piazza, P.V., Mayo, W., 2010. Low brain the brain’s biogenic amines after apomorphine and amphetamine. Pharmacol. allopregnanolone levels mediate flattened circadian activity associated with Biochem. Behav. 41, 675e682. memory impairments in aged rats. Biol. Psychiatry 68, 956e963. Serra, M., Pisu, M.G., Littera, M., Papi, G., Sanna, E., Tuveri, F., 2000. Social isolation- Girdler, S.S., Klatzkin, R., 2007. Neurosteroids in the context of stress: implications induced decrease in both the abundance of neuroactive steroids and GABAA for depressive disorders. Pharmacol. Ther. 116, 125e139. receptor function in rat brain. J. Neurochem. 75, 732e740. Griffin, L.D., Mellon, S.H., 1999. Selective serotonin reuptake inhibitors directly alter Stahl, S., 1998. Augmentation of antidepressants by estrogen. Psychopharmacol. activity of neurosteroidogenic enzymes. Proc. Natl. Acad. Sci. U. S. A. 96, Bull. 34, 319e321. 13512e13517. Ströhle, A., Pasini, A., Romeo, E., Hermann, B., Spalleta, G., di Michele, F., Holsboer, F., Halbreich, U., Kahn, L.S., 2001. Role of estrogen in the aetiology and treatment of Rupprecht, R., 2000. Fluoxetine decreases concentration of 3a,5a, tetra- mood disorders. CNS Drugs 15, 797e817. hydrodeoxycorticosterone (THDOC) in major depression. J. Psychiatr. Res. 34, Hirani, K., Khisti, R.T., Chopde, C.T., 2002. Behavioural action of ethanol in Porsolt’s 183e186. forced swim test: modulation by 3a-hydroxy-5a-pregnan-20-one. Neurophar- Ströhle, A., Romeo, E., Hermann, B., Pasini, A., Spalletta, G., di Michele, F., macology 43, 1339e1350. Holsboer, F., Rupprecht, R., 1999. Concentrations of 3a reduced neuroactive Holschneider, D., Kumazawa, T., Chen, K., Shih, J., 1998. Tissue specific effects of steroids and their precursors in plasma of patients with major depression and estrogen on monoamino oxidase A and B in the rat. Life Sci. 63, 155e160. after clinical recovery. Biol. Psychiatry 45, 274e277. Jung-Testas, I., Hu, Z.Y., Baulieu, E.E., Robel, P., 1989. Neurosteroids: biosynthesis of Svec, F., Lopez, A., 1989. Antiglucocorticoid actions of dehydroepiandrosterone and pregnenolone and progesterone in primary cultures of rat glial cells. Endocri- low concentrations in Alzheimer’s disease. Lancet 2 (8675), 1335e1336. nology 125, 2083e2091. Thomas, G., Frenoy, N., Legrain, S., Sebag-Lanoe, R., Baulieu, E.E., Debuire, B., 1994. Khisti, R.T., Chopde, C.T., Jain, S.P., 2000. Antidepressant-like effect of the neuro- Serum dehydroepiandrosterone sulfate levels as an individual marker. J. Clin. steroid 3a-hydroxy-5a-pregnan-20-one in mice forced swim test. Pharmacol. Endocrinol. Metab. 79, 1273e1276. Biochem. Behav. 67, 137e143. Tollefson, G.D., Haus, E., Garvey, M.J., Tuason, V.B., 1990. 24 hour urinary dehy- Korneyev, A., Costa, E., 1996. Allopregnanolone (THP) mediates anesthetic effects of droepiandrosterone sulfate in unipolar depression treated with cognitive and/ progesterone in rat brain. Horm. Behav. 30, 37e43. or pharmacotherapy. Ann. Clin. Psychiatry 2, 39e45. Landgren, S., Aasly, J., Backstrom, T., Dubrovsky, B., Danielsson, E., 1987. The effect of Urani, A., Roman, F.J., Phan, V.L., Su, T.P., Maurice, T., 2001. The antidepressant-like progesterone and its metabolites on the interictal epileptiform discharge in the effect induced by s1-receptor agonists and neuroactive steroids in mice cat’s cerebral cortex. Acta Physiol. Scand. 131, 33e42. submitted to the forced swimming test. J. Pharmacol. Exp. Ther. 298, 1269e1279. Landry, M., Di Paolo, T., 2003. Effect of chronic estradiol, tamoxifen or raloxifene Uzunov, D.P., Cooper, T.B., Costa, E., Guidotti, A., 1996. Fluoxetine-elicited changes in treatment on serotonin 5-HT1A receptor. Mol. Brain Res. 11, 82e89. brain neurosteroid content measured by negative ion mass fragmentography. Ma, Z.Q., Violani, E., Villa, F., Picotti, G., Maggi, A., 1995. Estrogenic control of Proc. Natl. Acad. Sci. U. S. A. 93, 12599e12604. monoamine oxidase A activity in human neuroblastoma cells expressing Uzunova, V., Ceci, M., Kohler, C., Uzunov, D.P., Wrynn, A.S., 2003. Region-specific physiological concentrations of estrogen receptor. Eur. J. Pharmacol. 284, dysregulation of allopregnanolone brain content in the olfactory bulbectomized 171e176. rat model of depression. Brain Res. 976, 1e8. Mamiya, T., Noda, Y., Nishi, M., Takeshima, H., Nabeshima, T., 1998. Enhancement of Uzunova, V., Sampson, L., Uzunov, D.P., 2006. Relevance of endogenous 3alpha- spatial attention in nociceptin/orphanin FQ receptor-knockout mice. Brain Res. reduced neurosteroids to depression and antidepressant action. Psychophar- 783, 236e240. macology 186, 351e361. Martínez-Mota, L., Contreras, C.M., Saavedra, M., 1999. Progesterone reduces Uzunova, V., Sheline, Y., Davis, J.M., Rasmusson, A., Uzunov, D.P., Costa, E., immobility in rats forced to swim. Arch. Med. Res. 30, 286e289. Guidotti, A., 1998. Increase in the cerebrospinal fluid content of neurosteroids in Marx, C.E., Shampine, L.J., Khisti, R.T., Trost, W.T., Bradford, D.W., Grobin, A.C., patients with unipolar major depression who are receiving fluoxetine or flu- Massing, M.W., Madison, R.D., Butterfield, M.I., Lieberman, J.A., Morrow, A.L., voxamine. Proc. Natl. Acad. Sci. U. S. A. 95, 3239e3244. 2006. Olanzapine and fluoxetine administration and coadministration increase Vallée, M., Rivera, J.D., Koob, G.F., Purdy, R.H., Fitzgerald, R.L., 2000. Quantification of rat hippocampal pregnenolone, allopregnanolone and peripheral deoxy- neurosteroids in rat plasma and brain following swim stress and allopreg- corticosterone: implications for therapeutic actions. Pharmacol. Biochem. nanolone administration using negative chemical ionization gas chromatog- Behav. 84, 609e617. raphy/mass spectrometry. Anal. Biochem. 287, 153e166. Matsumoto, K., Uzunova, V., Pinna, G., Taki, K., Uzunov, D.P., Watanabe, H., 1999. Van Broekhoven, F., Verkes, R.J., 2003. Neurosteroids in depression: a review. Permissive role of brain allopregnanolone content in the regulation of Psychopharmacology 165, 97e110. pentobarbital-induced righting reflex loss. Neuropharmacology 38, 955e963. Vanselow, W., Dennerstein, L., Greenwood, K.M., de Lignières, B., 1996. Effect of Murphy, B.E., 1997. Antiglucocorticoid therapies in major depression: a review. progesterone and its 5a and 5b metabolites on symptoms of premenstrual Psychoneuroendocrinology 22 (Suppl. 1), S125eS132. syndrome according to route of administration. J. Psychosom. Obstet. Gynaecol. Noda, Y., Mamiya, T., Furukawa, H., Nabeshima, T., 1997. Effects of antidepressants 17, 29e38. on phencyclidine-induced enhancement of immobility in a forced swimming Villard, V., Meunier, J., Chevallier, N., Maurice, T., 2011. Pharmacological interaction test in mice. Eur. J. Pharmacol. 324, 135e140. with the sigma1 (s1) receptor in the acute behavioural effects of antidepres- Oppenheim, G., 1983. Estrogen in the treatment of depression: neuropharmaco- sants. J. Pharmacol. Sci. 115, 279e292. logical mechanisms. Biol. Psychiatry 18, 721e725. Wang, H., Strobel, H.W., 1997. Regulation of CYP3A9 gene expression by estrogen Owens, M.J., 1996. Molecular and cellular mechanisms of antidepressant drugs. and catalytic studies using P450 3A9 expressed in Escherichia coli. Arch. Bio- Depress. Anxiety 4, 153e159. chem. Biophys. 344, 365e372. Potts, G.O., Creange, J.E., Harding, H.R., Schane, H.P., 1978. Trilostane, an orally active Weissman, M.M., Olfson, M., 1995. Depression in women: implications for health inhibitor of steroid biosynthesis. Steroids 32, 257e267. care research. Science 269, 799e801. Reddy, D.S., 2010. Neurosteroids: endogenous role in the human brain and thera- Wolkowitz, O.M., Reus, M., Keebler, A., Nelson, N., Friedland, M., Brizendine, L., peutic potentials. Prog. Brain Res. 186, 113e137. Roberts, E., 1999a. Double blind treatment of major depression with dihy- Reddy, D.S., Kaur, G., Kulkarni, S.K., 1998. Sigma (s1) receptor mediated anti- droepiandrosterone. Am. J. Psychiatry 156, 646e649. depressant-like effects of neurosteroids in the Porsolt forced swim test. Neu- Wolkowitz, O.M., Reus, V.I., 1999. Treatment of depression with antiglucocorticoid roreport 9, 3069e3073. drugs. Psychosom. Med. 61, 698e711. Robichaud, M., Debonnel, G., 2004. Modulation of the firing activity of female dorsal Wolkowitz, O.M., Reus, V.I., Chan, T., Manfredi, F., Raum, W., Johnson, R., Canick, J., raphe nucleus serotonergic neurons by neuroactive steroids. J. Endocrinol. 182, 1999b. Antiglucocorticoid treatment of depression: double-blind ketoconazole. 11e21. Biol. Psychiatry 45, 1070e1074. Robichaud, M., Debonnel, G., 2005. Oestrogen and testosterone modulate the firing Wolkowitz, O.M., Reus, V.I., Roberts, E., Manfredi, F., Chan, T., Ormiston, S., activity of dorsal raphe nucleus serotonergic neurones in both male and female Johnson, R., Canick, J., Brizendine, L., Weingartner, H., 1995. Antidepressant and rats. J. Neuroendocrinol. 17, 179e185. cognition-enhancing effects of DHEA in major depression. Ann. N. Y. Acad. Sci. Robinson, G., 2001. Psychotic and mood disorders associated with the peri- 774, 337e339. menopausal period. Epidemiology, aetiology and management. CNS Drugs 15, Wolkowitz, O.M., Reus, V.I., Roberts, E., Manfredi, F., Chan, T., Raum, W., Ormiston, S., 175e184. Johnson, R., Canick, J., Brizendine, L., Weingartner, H., 1997. Dehydroepian- Romeo, E., Ströhle, A., Spaelleta, G., di Michele, F., Herman, B., Holsboer, F., Pasini, A., drosterone (DHEA) treatment of depression. Biol. Psychiatry 41, 305e310. Rupprecht, R., 1998. Effects of antidepressant treatment on neuroactive steroids Young, J., Corpéchot, C., Perché, F., Eychenne, B., Haug, M., Baulieu, E.E., Robel, P., in major depression. Am. J. Psychiatry 155, 910e913. 1996. Neurosteroids in the mouse brain: behavioural and pharmacological Rupprecht, R., 2003. Neuroactive steroids: mechanisms of action and neuro- effects of a 3b-hydroxysteroid dehydrogenase inhibitor. Steroids 61, 144e149. psychopharmacological properties. Psychoneuroendocrinology 28, 139e168. Young, J., Corpéchot, C., Perché, F., Haug, M., Baulieu, E.E., Robel, P., 1994. Neuro- Rupprecht, R., Ströhle, A., Hermann, B., di Michele, F., Spalletta, G., Pasini, A., steroids: pharmacological effects of a 3b-hydroxy-steroid dehydrogenase Holsboer, F., Romeo, E., 1998. Neuroactive steroid concentrations following inhibitor. Endocrine 2, 505e509.