Neurosteroid pregnenolone induces sleep-EEG changes in man compatible with inverse agonistic GABAA-receptor modulation

Neurosteroid pregnenolone induces sleep-EEG changes in man compatible with inverse agonistic GABAA-receptor modulation

Brain Research, 615 (1993) 267-274 267 © 1993 Elsevier Science Publishers B.V. All rights reserved 0006-8993/93/$06.00 BRES 18963 Neurosteroid preg...

768KB Sizes 0 Downloads 65 Views

Brain Research, 615 (1993) 267-274

267

© 1993 Elsevier Science Publishers B.V. All rights reserved 0006-8993/93/$06.00 BRES 18963

Neurosteroid pregnenolone induces sleep-EEG changes in man compatible with inverse agonistic GABAA-receptormodulation Axel Steiger a, Lorenz Trachsel b, Jfirgen Guldner a Ulrich Hemmeter a,, Barbara Rothe Rainer Rupprecht b, Helmut Vedder b and Florian Holsboer a,b

a,

Max Planck Institute of Psychiatry, Clinical Institute, Departments of a Psychiatry and b Neuroendocrinology, Munich (Germany)

(Accepted 26 January 1993)

Key words: Neurosteroid; Pregnenolone; Human sleep; Electroencephalography; Hormone secretion

The steroid pregnenolone (P) and its sulfate (PS) can accumulate in the central nervous system independent of peripheral sources. Pharmacologically, the sulphated form of P interacts with the GABAA receptor complex, and functional assays show that this steroid behaves as an allosteric GABAA receptor antagonist. The present study explored the effect of a single dose of P upon the sleep-EEG and concurrent secretion of growth hormone and cortisol in male volunteers. P increased the amount of time spent in slow wave sleep and depressed EEG sigma power. Sleep-associated nocturnal cortisol and growth hormone secretion remained unchanged, ruling out the possibility that P exerted its effect via altered regulation of these hormones. Furthermore, results from in vitro studies on the potency of P to activate gene transcription via corticosteroid receptors made a genomic action of P via hormone receptor-sensitive DNA sequences unlikely. We conclude that P acts in a non-genomic fashion at or in the vicinity of the benzodiazepine binding site, modulating allosterically the GABAA receptor like a partial inverse agonist.

INTRODUCTION A d r e n o c o r t i c a l steroids are known to interact via their specific cytosolic receptors in various brain areas to p r o d u c e n e u r o e n d o c r i n e effects 2'43. A large n u m b e r of studies showed that corticosteroids are also important regulators of behavior, mood, anxiety, learning and sleep 15'31'46. Fifty years ago Selye 53 r e p o r t e d that certain steroid h o r m o n e s had hypnotic and anesthetic effects. Reevaluation of these early findings indicated that in addition to the classic genomic m e c h a n i s m of action corticosteroids may also rapidly m o d u l a t e neuronal excitability by altering the conductivity of steroid-sensitive ion channels. It is now well established that certain precursors or metabolites of major corticosteroids are able to interact with the g a m m a aminobutyric acid ( G A B A A) r e c e p t o r - c o u p l e d chloride channel 49. A m o n g these hormones, p r e g n e n o l o n e (P) and its sulfate (PS) are of particular interest because they accumulate in the central nervous system, where

they a p p e a r to be synthesized in glial cells 33'36, thus rendering the brain i n d e p e n d e n t of peripheral steroid sources. P can be converted directly into its conjugated form by sulfatase (and vice versa). Interestingly, the P biosynthesis from mevalonolactone in C6-2B glioma cell m i t o c h o n d r i a is regulated by d i a z e p a m binding inhibitor, a polypeptide a b u n d a n t in steroidogenic cells, which suggests a link between neurosteroids and benzodiazepines ( B D Z s ) 27'48. Electrophysiological studies support the view that

p54 and PS 40'42'54 interacts with the G A B A A - g a t e d chloride channel in an antagonistic fashion. In addition, PS has been shown to evoke a decrease in the postsynaptic potential in a slice preparation of adult neocortical neurons 59. F u r t h e r insight into P - G A B A A receptor interactions has been gained from binding studies. A t n a n o m o l a r concentrations PS showed allosteric G A B A A - a g o n i s t i c activity such as increasing the binding of muscimol and B D Z s in brain synaptosomal m e m b r a n e s p r e p a r e d from a d r e n a l e c t o m i z e d

Correspondence: A. Steiger, Max Planck Institute of Psychiatry, Clinical Institute, Department of Psychiatry, KraepelinstraBe 2-10, D-8000

Munich 40, Germany. Fax: (49) (89) 3062 2605. * Present address: University of Basle, Department of Psychiatry, Wilhelm-Klein-StraBe 27, CH-4025 Basle, Switzerland.

268 rats 42. In contrast, at micromolar concentrations, PS inhibits TBPS binding, decreases muscimol binding, blocks GABA agonist-stimulated chloride uptake in synaptoneurosomes 42 and diminishes chloride conductance by reducing channel-opening frequency 4~'45. All these effects are compatible with an antagonistic mode of action. In addition to these effects upon the GABA A receptor, P and PS also interfere with voltage-gated calcium channels 54 and PS potentiates the NMDAgated ion current 67. In the light of these diverse in vitro actions, the supracellular (behavioral) relevance of P and its sulphate is difficult to predict. Only few data are available on this issue. PS and more distinctly P showed memoff-enhancing effects in mice 21. In this paradigm P was effective over a wide dose range (10,000 fold). Corticosteroids exert a number of effects upon sleep regulation 9"~°'25'26, and the G A B A A receptor, at which PS has been shown to act, is involved in mediating some of the sleep-promoting effects of hypnotics such as the BOZs 7'8'16'20"22'34'35'37'61'66.Therefore, we elected to use sleep-EEG recordings as a tool to investigate physiologic effects of P in man. We measured plasma growth hormone (GH) and cortisol concentrations serially during sleep because the hypothalamic neuropeptides controlling these hormones have been shown to be involved in regulation of sleep 19'32'55. In addition, we studied the effects of P and of PS on the expression of corticosteroid-regulated genes to rule out confounding genomic effects. MATERIALS AND METHODS Subjects 12 healthy male volunteers aged 20-30 years (mean _+S.D., 25.9 + 2,2) were enrolled after the purpose of the study had been explained and after written informed consent was obtained. The study was approved by the Ethics Committee for H u m a n Experiments of the Max Planck Institute of Psychiatry. The height and weight of the subjects were within the normal range and all subjects underwent extensive psychiatric, physical and laboratory (hematology, virology, clinical chemistry, endocrinology, E E G and electrocardiographic [ECG]) examinations. Individuals with a personal or family history of psychiatric disorders or a recent stressful life event were excluded, as were shift workers and persons who had made a recent transmeridJan flight. Abuse of drugs, nicotine, alcohol and caffeine and any history of medical treatment during the past 3 months were also ruled out. Sleep and sleep-EEG spectra studies The sleep-endocrine study consisted of two sessions with administration of placebo or an oral dose of 1 mg P at 22.00 h according to a randomized schedule. During each session the subjects spent 2 nights in the laboratory. The first night served for adaptation to the laboratory setting. On the second day the subjects received an electrolyte- and calorie-balanced diet at 08.00 h, 12.00 h and 18.00 h. No snacks were allowed between meals. At 19.30 h an indwelling forearm catheter was inserted and connected to a plastic-tubing extension, placed through a soundproof lock into the adjacent room. It was kept patent with a 0.9% saline drip. In the adjacent room

subjects were observed on a TV screen. Polygraphic recordings (EEG, electrooculogram, electromyogram and ECG) were monitored between 23.00 h and 07.00 h. Electrodes were fixed between 19.30 h and 20.00 h. Sleeping was not allowed before 'lights off' at 23.00 h. If necessary, the subjects were awakened at 07.00 h. In the adjacent room blood was sampled every 30 rain from 20.00 h to 22.00 h and every 20 rain from 22.00 to 07.00 h for later analysis of plasma concentrations of cortisol and G H by radioirnmunoassay as described in detail elsewhere 55. Specimens collected between 20.00 h and 22.00 h served to control for stress effects on hormone secretion due to cannulation, whereas those sampled from 22.00 h to 7.00 h were used to assess study effects. Sleep-EEG recordings were scored manually according to conventional criteria 52. Definitions of sleep-EEG variables were given in detail elsewhere 56. The electrophysiological variables (notch filter at 50 Hz; calibration: 50 IzV, 10 Hz sine wave) were digitized at 100 Hz and stored on tape. Subsequently, the 8 h E E G recordings (C3-A2 derivation) of a random subset of 6 subjects were submitted to a Fast Hartley Transform algorithm ~162. E E G power spectra were computed for consecutive rectangular windows of 256 samples (approx. 2.56 s). This sample size in combination with the 100 Hz-sampling rate yields a frequency resolution of approx. 0.39 Hz between 0 and 50 Hz. Frequency bins above approx. 20 Hz were omitted from further analysis. E E G spectra were aligned with 30-s epochs of the visual scores by averaging the EEG spectra over 12 consecutive windows (approx. 30.72 sec). The routine stepped back 72 samples before analyzing the following 12 windows. E E G power was averaged across three frequency bands (delta, 0.76-4.56 Hz; theta, 4.56-10.26 Hz; sigma, 10.26-14.1 Hz). 30-s epochs were rejected as E E G artefacts whenever the power exceeded individual thresholds in one of the three frequency bands. Visual comparison of the hypnogram and corresponding evolutions of delta, theta and sigma power allowed the determination of consecutive n o n - R E M - R E M sleep cycles (Fig. 1). Mean values of E E G spectral power per frequency bin in each sleep cycle were computed for combined n o n - R E M sleep stages 2, 3 and 4. Statistical differences between placebo-EEG and P-EEG were determined separately for sleep cycles 1-4 (all n = 6). Each frequency bin was individually normalized to its placebo value of the corresponding sleep cycle (100% = placebo level in each cycle; Fig. 2), and a Bonferroni-adjusted t-test (two-sided) was performed on the log-transformed values.

In ~,itro studies Cell culture and tran~fection. To investigate hormone-dependent transactivation mediated by h G R and hMR, we employed a cis-trans co-transfection system using the h u m a n neuroblastoma cell line SK-N-MC. The cis-vector (reporter gene) consists of the steroid receptor-inducible mouse m a m m a r y tumor virus (MTV) promoter fused to the luciferase gene. The trans-vector contains the h G R or h M R cDNA inserted in an expression vector. Transfection efficiency is monitored by co-transfection of p C H l l 0 (Pharmacia LKB, Freiburg, Germany), an SV 40 promoter-driven /3-galactosidase expression vector. SK-N-MC cells were grown in Dulbecco's modified Eagle's medium ( D M E M ) supplemented with 10% fetal calf serum (FCS). Transfections were performed using an electroporation system (Biotechnologies & Experimental Research Inc., San Diego, CA) after determination of the optimal electric field strength 13. Usually, 2.5 tzg reporter gene, 5 /~g receptor expression vector and 2.5 p,g carrier D N A (pGEM4) (Promega Corporation, Madison, WI) were co-transfected. Electroporated cells were replated in D M E M supplemented with 10% charcoal-stripped steroid-free FCS 3 and incubated immediately with various concentrations of hormones as indicated. After 24 h, cells were harvested and extracts assayed for luciferase 64 and /3-galactosidase 29, Addition of steroid receptor agonists activates the expressed receptors, which act as hormone-inducible transcription factors at the M T V promoter. The bioactivity of luciferase is directly proportional to the induction of the MTV promoter. The construction of the plasmids M T V - L U C 3°, h G R ( p R S h G R a ) 24 and h M R (pRShMR) 3 has been reported previously. Proopiomelanocortin (POMC) gene expression. Pituitary cells (AtT 20) were cultured under serum-free completely defined conditions as described elsewhere in detail ~3. After incubation with P and PS for

269 TABLE I

Sleep-EEG variables after 1 mg pregnenolone and after placebo in normal controls (n = 12) SPT, sleep period time; SEI, sleep efficiency index; SOL, sleep onset latency; SWS, slow wave sleep; SWS cycle 1, SWS during 1st REM-non-REM cycle (min); WRT, Wilcoxon rank test.

1. Placebo (mean + S.D.)

2. 1 mg pregnenolone (mean +_S.D.)

WRT 1:2

443.2 435.1 95.5 20.7

43.8 47.3 8.5 19.1

447.0 450.8 98.9 23.3

32.9 29.3 1.6 20.1

ns ns P < 0.05 ns

18.6 19.8 253.5 56.1 12.8 31.5 84.3 118.5

36.2 15.2 41.1 17.2 4.2 15.9 31.5 71.3

4.8 19.2 256.2 66.9 15.1 44.0 87.5 128.0

7.5 10.2 37.6 25.5 6.0 22.6 18.1 55.1

ns ns ns P < 0.05 P < 0.05 P < 0.05 ns ns

Sleep continuity SPT (min) TST (min) SEI SOL (min)

Sleep architecture Wake (min) Stage 1 (min) Stage 2 (min) SWS (min) SWS (%) SWS cycle 1 REM (min) REM latency (min)

PLACEBO M~A

i

i I

23

III I

. . . . . . . . . . . . .

"

I

I

I

.

I

24

I~1

1

2

3

4

I

5

6

7

lmg PREGNENOLONE p.o. ..A

.

-,..:'

,

.

:,,

,,,

,

, , ,,

,,

;,

,

i

. v ~.

i

23

24

1

2

3 DAY

4

5

6

7

TIME (hours)

Fig. 1.8-h recordings of the same subject after placebo and after 1 mg P orally administered at 23.00 h. Panel A represents the hypnogram (30-s epochs; visual scores: M = movement time; W = wakefulness; R = REM sleep; 1-4 = non-REM sleep stages). Panels below illustrate the evolution of EEG power of the delta (B: 0.76-4.56 Hz; calibration 275 ~V 2 per tick mark) and sigma (C: 10.26-14.1 Hz; calibration 10 p,V e) bands across the sleep periods (curve connects 30-s mean values). Vertical dotted lines delimit the completion of non-REM-REM sleep cycles. Note that the power plots reflect continuous changes in sleep EEG processes rather than stepwise functions as suggested by the hypnogram. An increase in SWS (stages 3 and 4) and a reduction of EEG sigma power can be observed after P intake.

270

1 mg Pregnenolone p.o.

48 h, cytoplasmatic mRNA was extracted according to Berger 5, separated on 1.2% formaldehyde gels, capillary blotted to Hybond (Amersham, Braunschweig, Germany) membranes and hybridized with nick-translated cDNA probes for POMC 1 and /3-actin 6°. Resuits are based on data from at least three experiments with different concentrations of steroid.

50 25 0

O m

RESULTS

I.H 0 <

Sleep and sleep-EEG spectra Results of visual scoring of sleep-EEG variables, following placebo or 1 mg P are given in Table I. The sleep efficiency index increased significantly on study nights with active medication ( P < 0.05). Other variables describing sleep continuity, including sleep latency, were not influenced by P. The absolute amount and the percentage of SWS for the total night increased significantly ( P < 0.05) after P (Table I). This effect was also detectable when the time in SWS during the first R E M - n o n - R E M cycle was inspected separately. P reduced intermittent wakefulness, but this effect was not statistically significant. Moreover, other variables of sleep architecture, including R E M sleep, failed to show significant differences between placebo and treatment. In the subgroup (n = 6) used for spectral analysis, SWS for the total night increased from 106.0 min (S.D. 28.5) to 129.2 min (30.4; n.s.), and SWS for the first half of the night increased from 95.2 rain (25.5) to 120.7 rain (23.8; P < 0.05), which is an indication that the subsample used for spectral analysis was representative of the entire group. As illustrated in Figs. 1 and 2 during the first sleep cycle after P administration the E E G power between 0 and 8.5 Hz was close to placebo levels, whereas it was reduced by at least 15% from placebo values between 8.5 and 19.5 Hz ( P < 0.05). The strongest decrease was at 11-13 Hz (approx. - 4 0 % from control levels). Dur-

n

Ow u_ uS u_ n" m © a_ _z < rr I-o tu a_ co

-25 -50

Sleep Cycle 2 25 0 -25

~-~

Sleep Cycle 3

T

25 0 -25 50

Sleep Cycle 4

25 0 -25 -50 0

5

10

15

Fig. 2. P EEG power spectra (0-19.3 Hz) for sleep cycles 1-4 in man (n = 6). Bars show deviation from placebo level ( = 100%) in each 0.39 Hz bin (error bars: 1 S.E.M.). Solid bars denote significant differences between placebo and P (two-sided Bonferroni t-test, nominal P < 0.05).

ing the second sleep cycle the E E G power was still markedly decreased, by more than 15% between 10.5 and 13 Hz ( P < 0.05). The E E G power of the third

TABLE II

Endocrine cariables under 1 mg pregnenolone and under placebo in normal controls (n = 12) conc., concentration; AUC, area under curve; WRT, Wilcoxon rank test; corr. rise, cortisol rise (for definition see ref. 56).

1. Placebo (mean +_S.D.) Cortisol t,ariables Cortisol conc. ( n g / m l ) 23.00-07.00 h 23.00-03.00 h 03.00-07.00 h Latency 22.00 h - cort. rise (rain) G H cariables GH conc. ( n g / m l ) 23.00-07.00 h 23.00-03.00 h 03.00-07.00 h GH peak AUC ( n g / m l . min) Latency 22.00 h - GH maximum (rain)

20

EEG FREQUENCY (Hz)

2. 1 mg pregnenolone (mean +_S.D.)

W R T 1:2

53.8 19.0 82.5 279.2

17.9 19.1 28.5 63.0

57.6 17.0 97.4 318.3

19.2 8.7 35.1 78.8

ns ns ns ns

4,2 6,5 1,2 1865 158.3

2.6 4.4 1.2 1407 64.1

4.2 5.9 1.6 2164 150.0

3.2 3.9 2.3 1 607 37.7

ns ns ns ns ns

271 fold induction

hGR

hMR

10o0 ~

, 20 I

hGR hMR

15 100-

lO lo-

I

: baselne

~ RU 28362

o ALDO

pregnenolone

Fig. 3. Transactivation properties of h G R and hMR after addition of RU 28362, ALDO and P. Results are expressed in terms of induction of the MTV promoter; the baseline promoter activity without addition of hormone is set as 1. Results represent the mean_+ S.D. from four independent experiments.

sleep cycle was almost back to control levels except for the significant reduction of 15% around 12 Hz (P < 0.05). No significant changes in EEG power spectra can be observed during the fourth sleep cycle. Hormone secretion

The pattern of cortisol and GH secretion under placebo is in agreement with that previously reported for younger normal control subjects 56. P treatment did not change the secretory activity of GH and cortisol compared to placebo (Table II). In vitro studies

As illustrated in Fig. 3, addition of the synthetic hGR agonist RU 28362 strongly activated the hGR, as shown by the approximately 1,000-fold induction of the MTV promoter. Aldosterone also activated the hGR, but to a lesser extent than the pure hGR agonist RU 28362. In contrast, the aldosterone-induced stimulation of the transcription rate by the hMR was only 25-fold due to the low transactivation potential of the hMR, whereas RU 28362 did not stimulate the hMR. Both hGR and hMR agonists were capable of activating gene expression via their respective receptors, but P did not exert any effect via either hGR or hMR. Quantitative analysis of POMC mRNA concentrations after 48 h of treatment of pituitary AtT 20 cells revealed no measurable change as induced by P and PS, ruling out a GR-mediated effect in this assay system. DISCUSSION This study is the first to demonstrate that a low dose of systemically administered P can have numerous rapid effects on human sleep EEG. These effects are un-

likely to be mediated via neuroendocrine sleep regulators or genomic actions at intracellular corticosteroid receptors. According to conventional visual scoring P improved the quality of sleep as documented by an increase in SWS, a higher sleep efficiency and a trend to decreased intermittent wakefulness. Spectral analysis showed profound power reduction in the fast frequency range above the delta and theta bands. This effect was most pronounced during the first sleep cycle and gradually disappeared in subsequent cycles. EEG delta power, however, did not change after P. This was unexpected in view of the increased amount of visually scored SWS. Such discrepancies between visual scores and EEG spectral data have been repeatedly demonstrated in both physiological and pharmacological settings 6'12'17'61, supporting the view that conventional sleep-EEG scoring should be complemented with quantitative signal analysis of the sleep EEG. In the light of electrophysiological data suggesting an action of P at the GABA-BDZ receptor, we compared the present data with those reported for ligands of the BDZ binding site. BDZs such as flunitrazepam decrease visually scored S W S 22'37 a s well as EEG amplitude and power around 2 Hz 7'8"16'20'34'35'61"66 and prompt increased activity at frequencies around 12 Hz 7'8'16'35'61. Furthermore, non-BDZ hypnotics such as zopiclone and zolpidem, which bind specifically to the GABA-BDZ receptor complex, depress EEG delta power and increase EEG sigma power 12'61"66. P, on the other hand, decreases EEG sigma power, but increases SWS. Based on these findings we posit that P exerts inverse agonistic properties at the GABA receptor, probably at or in the vicinity of the BDZ site. The hypothesis has been advanced that EEG changes in the 12-14 Hz-frequency range represent a common 'spectral BDZ signature 'j2'61. The present data support this hypothesis and may reflect a relatively close linkage between receptor action and specific EEG changes in human sleep: BDZ agonists increase EEG sigma power, whereas BDZ inverse agonists decrease it. Further differences between P and BDZ agonists can be observed outside the 12 Hz range. Under the influence of BDZ receptor agonists EEG power is reduced in almost all frequencies below the sigma band and is very close to control levels above 15 Hz 8'12'16'61. In contrast, P does not significantly affect low EEG frequencies and depresses power above the sigma band during the first sleep cycle. Although this could be a specific inverse agonistic effect on the cortical EEG, unspecific effects cannot be ruled out. For example, the depression of SWS and the reduction of slow wave activity by BDZ agonists is observed much beyond their

272 elimination half-life 722'35 and is therefore considered to reflect unspecific B D Z effects. It should be kept in mind that in vitro micromolar concentrations of PS were necessary to prompt inverse agonistic effects 3~ and that it is unlikely that this concentration was reached in our experiment. On the other hand it is suggested that a wide dose range of P is capable to exert behavioral effects from an animal study 21. To date no pharmacokinetic data and particularly no elimination half-life are available for P. However, the effect of P on the sleep-EEG 12-Hz activity is limited to the first three sleep cycles, allowing the assumption that elevated blood levels of P occur after oral intake of l mg of the substance and that the elimination half-life of P could be in the range of a few hours. It has been shown that augmentation of E E G power in the 12-14Hz range or of sleep spindle activity may be closely related to the elimination half-life of BDZ-agonistic hypnotics 12'35'6t. Total sleep deprivation attenuates spectral power of the spindle frequency range during recovery sleep in man w. It is possible that the downregulation of E E G sigma power after sleep loss and the reduction after P could be mediated by similar mechanisms, e.g. the G A B A - B D Z receptor complex. Changes in E E G sleep after systemic administration of corticosteroids that bind to G R and M R are well documented. For example, R E M sleep is suppressed by cortisol 4'm, prednisone 25, dexamethasone 9'm and fluocortolone 1°. Increased amount of time spent in SWS is induced by cortisol 4'm and dexamethasone 9,m, whereas the combined glucocorticoid and progesterone receptor antagonist R U 486 (mifepristone) abolishes the physiological sleep-EEG structure completely 65. Some studies 9"26, but not all 57'5s suggest that MRs are directly involved in a steroid-mediated increase in SWS. In the current study we investigated whether P-induced effects are mediated by corticosteroid receptors. Employing co-transfection assays, we showed that P is devoid of any measurable genomic effects mediated via G R or MR. In addition, the failure of P and of PS to alter the level of cytoplasmatic P O M C m R N A makes a corticosteroid receptor-driven activity of P unlikely ~s. Also the possibility that P exerts its effect by interacting with sleep-promoting neuropeptides must be considered, particularly because P is a precursor of cortisol via the biosynthetic route and may well have interacted with hormones involved in the secretion and release of G H and ACTH. To date, animal and clinical studies consistently show that growth hormone-releasing hormone increases SWS, whereas corticotropin-releasing hormone has an opposite effect 19'32'47'55. Any action of P upon these two neuropeptides seems to be unlikely because serial monitoring of plasma hormone

concentrations throughout the night yielded no changes in plasma cortisol or G H concentrations. However, the possibility that P or its sulfate are transformed into compounds which act through a yet unidentified mechanism, e.g. after conversion into one or several other steroids can not be ruled out. The new concept of neurosteroid function in the brain was recently put forward after several 3a-hydroxy A-ring-reduced pregnane steroids were found to bind with high affinity to recognition sites associated with the G A B A receptor complex, potentiating G A B A chloride currents at low n a n o m o l a r concentrations 23"2s'4°'49. The potentiating action of one of these steroids, 3a,21-dihydroxy-5apregnan-200ne (allotetrahydroDOC) was confirmed by experiments using recombinantly expressed G A B A receptors, and its efficacy was independent of the presence of the 72-subunit, a structural requirement necessary for mediating effects of BDZs 5°. The functional significance of allotetrahydroDOC is suggested by its anxiolytic TM and hypnotic 44 effects in rats, where the levels of this steroid (derived from the adrenal cortex) are elevated in the brain when the animals are exposed to acute stress 51. Taken together, the effects reported here are best understood as partial inverse agonistic effects of P at the G A B A receptor complex, modulating chloride conductance at or in the vicinity of the B D Z binding site. Such a pharmacology may provide a potential lead for development of specific sleep regulators. Interestingly, partial inverse agonists at the G A B A - B D Z receptor derived from /~-carbolines are now being clinically tested for their potential to ameliorate cognitive deficits in early dementia. P, which according to current data may also be regarded as an inverse G A B A - B D Z receptor agonist, was recently reported to have memory-enhancing effects in mice 21. This study was supported by a grant from the Deutsche Forschungsgemeinschaft(Ste 486/1-1). We thank Drs. M. Lancel and D.J. Dijk for their invaluable comments on parts of the manuscript.

Acknowledgements.

REFERENCES 1 Affolter, U. and Reisine, T., Corticotropin releasing factor increases proopiomelanocortin messenger RNA in mouse anterior pituitary tumor cells, J. Biol. Chem., 260 (1985) 15477-15548. 2 Arriza, J.L., Simerly, B.. Swanson, L.W. and Evans, R.M., The neuronal mineralocorticoid receptor as a mediator of glucocorticoid response, Neuron, 1 (1988)887-900. 3 Arriza, J.L., Weinberger, C., Cerelli, G., Glaser, T.M., Handelin, B.L., Housman, D.E. and Evans, R.M., Cloning of human mineralocorticoid receptor complementary DNA: structural and functional kinship with the glucocorticoid receptor, Science, 237 (1987) 268-275.

273 4 von Bardeleben, U., Lauer, C., Wiedemann, K. and Holsboer, F., Nocturnal sleep-endocrine effects of cortisol infusion in normal controls, Neuroendocrinol. Lett., 10 (1989)227. 5 Berger, S.W., Isolation of cytoplasmatic RNA: ribonucleosidevanadyl complexes, Methods Enzymol., 152 (1987) 227-234. 6 Borb61y, A.A., Baumann, F., Brandeis, D., Strauch, I. and Lehmann, D., Sleep deprivation: effect on sleep stages and EEG power density in man, Electroencephalogr. Clin. NeurophysioL, 51 (1981) 483-493. 7 Borb61y, A.A., Mattmann, P., Loepfe, M., Fellmann, I., Gerne, M., Strauch, I. and Lehmann, D., A single dose of benzodiazepine hypnotics alters the sleep EEG in the subsequent drugfree night, Eun J. Pharmacol., 89 (1983) 157-161. 8 Borb61y, A.A., Mattmann, P., Loepfe, M., Strauch, I. and Lehmann, D., Effect of benzodiazepine hypnotics on all-night sleep EEG spectra, Hum. Neurobiol., 4 (1985) 189-194. 9 Born, J., DeKloet, E.R., Wenz, H., Kern, W., and Fehm, H.L., Gluco- and antimineralocorticoid effects on human sleep: a role of central corticosteroid receptors, Am. J. Physiol., 260 (1991) E183-E188. 10 Born, J., Zwick, A., Roth, G., Fehm-Wolfsdorf, G. and Fehm, H.L., Differential effects of hydrocortisone, fluocortolone, and aldosterone on nocturnal sleep in humans, Acta Endocrinol. (Copenh.), 116 (1987) 129-137. 11 Bracewell, R.N., The Fast Hartley Transform, Oxford University Press, New York, 1986. 12 Brunner, D.P., Dijk, D.J., Munch, M. and Borb61y, A.A., Effect of zolpidem on sleep and sleep EEG spectra in healthy young men, Psychopharmacology, 104 (1991) 1-5. 13 Chu, G., Hayakawa, H. and Berg, P., Electroporation for the efficient transfection of mammlian cells with DNA, Nucleic Acids Res., 15 (1987) 1311-1326. 14 Crawley, J.N., Glowa, J.R., Majewska, M.D. and Paul, S.M., Anxiolytic activity of an endogenous adrenal steroid, Brain Res., 398 (1986) 382-385. 15 De Kloet E.R., Brain corticosteroid receptor balance and homeostatic control, Front. Neuroendocrinol., 12 (1991) 95-165. 16 Dijk, D.J., Beersma, D.G.M., Daan, S. and van den Hoofdakker, R.H., Effects of seganserin, a 5-HT 2 antagonist, and temazepam on human sleep stages and EEG power spectra, Eur. J. Pharmacol., 171 (1989) 207-218. 17 Dijk, D.J., Brunner, D.P. and Borb61y, A.A., Time course of EEG power density during long sleep in humans, Am. J. Physiol., 258 (1990) R650-R661. 18 Drouin, J., Charron, J. and Gagner, J.P., Proopiomelanocortin gene: a model for negative regulation of transcription by glucocorticoids, J. Cell. Biochem., 35 (1987) 293-304. 19 Ehlers, C.L., Reed, T.K. and Henriksen, S.J., Effects of corticotropin-releasing factor and growth hormone-releasing factor on sleep and activity in rats, Neuroendocrinology, 42 (1986) 467-474. 20 Feinberg I., Fein G., Walker J.M., Price L.J., Floyd T.C. and March J.D., Flurazepam effects on slow wave sleep: stage 4 suppressed but number of delta waves constant, Science, 198 (1977) 847-848. 21 Flood, J.F., Morley, J.E. and Roberts, E., Memory-enhancing effects in male mice of pregnenolone and steroids metabolically derived from it, Proc. Natl. Acad. Sci. USA, 89 (1992) 1567-1571. 22 Gaillard, J.M., Schulz, P. and Tissot, R., Effects of three benzodiazepines (nitrazepam, flunitrazepam and bromazepam) on sleep of normal subjects, studied with an automatic sleep scoring system, Pharmacopsychiatry, 6 (1973) 207-217. 23 Gee, K.W., Steroid modulation of the GABA/benzodiazepine receptor-linked chloride ionophore, Mol. Neurobiol., 2 (1988) 291-317. 24 Gigubre, V., Hollenberg, S.M., Rosenfeld, M.G. and Evans, R.M., Functional domains of the human glucocorticoid receptor, Cell, 46 (1986), 645-652. 25 Gillin, J.C., Jacobs, L.S., Fram, D.H. and Snyder, F., Acute effect of a glucocorticoid on normal human sleep, Nature, 237 (1972) 398-399. 26 Gillin, J.C., Jacobs, L.S., Snyder, F. and Henkin, R.I., Effects of decreased adrenal corticosteroids: changes in sleep in normal

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46 47

48

subjects and patients with adrenal cortical insufficiency, Electroencephalogr. Clin. NeurophysioL , 36 (1974) 283-289. Guarneri, P., Papadopoulos, V., Pan, B. and Costa, E., Regulation of pregnenolone synthesis in C6-2B glioma cells by 4'-chlorodiazepam, Proc. Natl. Acad. Sci. USA, 89 (1992) 5118-5122. Harrison, N.L., Majewska, M.D., Harrington, J.W. and Barker, J.L., Structure-activity relationship for steroid interaction with the 3,-aminobutyric acid A receptor complex, J. Pharmacol. Exp. Then, 241 (1987) 346-353. Herbomel, P., Bourachot, B. and Yaniv, M., Two distinct enhancers with different cell specificities coexist in the regulatory region of polyoma, Cell, 39 (1984), 653-662. Hollenberg, S.M. and Evans, R.M., Multiple and cooperative transactivation domains of the human glucocorticoid receptor, Cell, 55 (1988) 899-906. Holsboer, F., Psychiatric implications of altered limbic-hypothalamic-pituitary-adrenocortical activity, Eur. Arch. Psychiatry Neurol. Sci., 238 (1989) 302-322. Holsboer, F., von Bardeleben, U. and Steiger, A., Effects of intravenous corticotropin-releasing hormone upon sleep-related growth hormone surge and sleep EEG in man, Neuroendocrinology, 48 (1988) 32-38. Hu, Z.Y., Bourreau, E., Jung-Testas, I., Robel, P. and Baulieu, E.-E., Neurosteroids: oligodendrocyte mitochondria convert cholesterol to pregnenolone, Proc. NatL Acad. Sci. USA, 84 (1987) 8215-8219. Johnson, L.C., Seales, D.M., Naitoh, P., Church, M.W. and Sinclair, M., The effects of flurazepam hydrochloride on brain electrical activity during sleep, Electroencephalogr. Clin. Neurophysiol., 47 (1979) 309-321. Johnson, L.C., Spinweber, C.L., Seidel, W.F. and Dement, W.C., Sleep spindle and delta changes during chronic use of a shortacting and long-acting benzodiazepine hypnotic, Electroencephalogr. Clin. Neurophysiol. , 55 (1983) 662-667. Jung-Testas, I., Hu, Z.Y., Baulieu, E.E. and Robel, P., Neurosteroids: biosynthesis of pregnenolone and progesterone in primary cultures of rat glial cells, Endocrinology, 125 (1989) 20832091. Kales, A., Soldatos, C.R., Bixler, E.O., Kales, J.D. and Vela Bueno, A., Diazepam: effects on sleep and withdrawal phenomena, J. Clin. Psychopharmacol., 8 (1988) 340-345. Majewska, M.D., Neurosteroids: endogenous bimodal modulators of the GABA A receptor. Mechanism of action and physiological significance, Prog. Neurobiol., 38 (1992) 379-395. Majewska, M.D., Demirg6ren, S. and London, E.D., Binding of pregnenolone sulfate to rat brain membranes suggests multiple sites of steroid action at the GABA A receptor, Eur. J. Pharmacol., 189 (1990) 307-315. Majewska, M.D., Harrison, N.L., Schwartz, R.D., Barker, J.L. and Paul, S.M., Steroid hormone metabolites are barbiturate-like modulators of t he GABA receptor, Science, 232 (1986) 1004-1007. Majewska, M.D., Mienville, J.M. and Vicini, S., Neurosteroid pregnenolone sulfate antagonizes electrophysiological responses to GABA in neurons, Neurosci. Lett., 90 (1988) 279-284. Majewska, M.D. and Schwartz, R.D., Pregnenolone-sulfate: an endogenous antagonist of the 3,-aminobutyric acid receptor complex in brain? Brain Res., 404 (1987) 355-360. McEwen, B.S., Steroids affect neural activity by acting on the membrane and the genome, Trends PharmacoL Sci., 12 (1991) 141-147. Mendelson, W.B., Martin, J.V., Perils, M., Wagner, R., Majewska, M.D. and Paul, S.M., Sleep induction by an adrenal steroid in the rat, Psychopharmacology, 93 (1987) 226-229. Mienville, J.-M. and Vicini, S., Pregnenolone sulfate antagonizes GABA A receptor mediated currents via reduction of channel opening frequency, Brain Res., 489 (1989) 190-194. Murphy, B.E.P., Steroids and depression, J. Steroid Biochem. Mol. Biol., 38 (1991) 537-559. Obfil Jr., F., Alf61di, P., Cady, A.B., Johannsen, L., Sary, G. and Krueger, J.M., Growth hormone-releasing factor enhances sleep in rats and rabbits, Am. J. Physiol., 255 (1988) R310-R316. Papadopoulos, V., Guarneri, P., Krueger, K.E., Guidotti, A. and

274

49 50

51

52

53

54

55

56

57

58

Costa, E., Pregnenolone biosynthesis in C6-2B glioma cell mitochondria: regulation by a mitochondrial diazepam binding inhibitor receptor, Proc. Natl. Acad. Sci. USA, 89 (1992) 5113-5117. Paul, M.S. and Purdy R.H., Neuroactive steroids, FASEB J., 6 (1992) 2311-2322 Puia, G., Santi, M.R., Vicini, S., Pritchett, D.B., Purdy, R.H., Paul, S.M., Seeburg, P.H. and Costa, E., Neurosteroids act on recombinant human GABA A receptors, Neuron, 4 (1990) 759765. Purdy, R.H., Morrow, A.L., Moore Jr., P.H. and Paul, S.M., Stress-induced elevations of 7-aminobutyric acid type A receptor-active steroids in the rat brain, Proc. Natl. Acad. Sci. USA, 88 (1991) 4553-4557. Rechtschaffen, A., Kales, A. (Eds.), A Manual of Standardized Terminology Techniques and Scoring System for Sleep Stages of Human Subjects, US Department of Health, Education and Welfare, Neurological Information Network, Bethesda, 1968. Selye, H., Correlations between the chemical structure and the pharmacological actions of the steroids, Endocrinology, 30 (1942) 437-452. Spence, K.T.. Plata-Salaman, C.R. and ffrench-Muller, J.M.H., The neurosteroids pregnenolone and pregnenolone-sulfate but not progesterone, block Ca 2+ currents in acutely isolated hippocampal CA1 neurons, Life Sci., 49 (1991) PL235-PL239. Steiger, A., Guldner, J., Hemmeter, U., Rothe, B., Wiedemann, K., and Holsboer, F., Effects of growth hormone-releasing hormone and somatostatin on sleep EEG and nocturnal hormone secretion in male controls, Neuroendocrinology, 56 (1992) 566-573 Steiger, A., Herth, T. and Holsboer, F., Sleep-electroencephalography and the secretion of cortisol and growth hormone in normal controls, Acta Endocrinol. (Copenh.), 116 (1987) 36-42. Steiger, A., Pollm~icher, T., Lauer, C. and Holsboer, F., Modulation of nocturnal sleep EEG in humans by neuropeptides and adrenal steroids, Biol. Psychiatry, 29 (9A) (1991) 52A. Steiger, A., Rupprecht, R., Spengler, D., Guldner, J., Hemmeter, U., Rothe, B., Damm, K. and Holsboer, F., Functional properties

59

60

61

62

63

64

65

66

67

of deoxycorticosterone and spironolactone: molecular characterization and effects on sleep-endocrine activity, J. Psychiat. Res., in press. Teschemacher, A., Zeise, M.L. and Zieglg~insberger, W., The neuroactive steroids tetrahydrodeoxycorticosterone and pregnenolone sulfate modulate inhibitory postsynaptic potentials in neocortical neurons of the rat in vitro, 1992, ENA Congress, Munich, abstract. Tokunaga, K., Tanigucchi, H. and Yoda, K., Nucleotide sequence of a full-length cDNA for mouse cytoskeletal /3-actin mRNA, Nucleic Acids Res., 14 (1986) 235. Trachsel, L., Dijk, D.J., Brunner, D., Kleene, C. and Borb~ly, A.A., Effect of zopiclone and midazolam on sleep and EEG spectra in a phase-advanced sleep schedule, Neuropsychopharmacology, 3 (1990) 11-18. Trachsel, L., Edgar, D.M., Seidel, W.F., Heller, H.C. and Dement, W.C., Sleep homeostasis in suprachiasmatic nuclei-lesioned rats: effects of sleep deprivation and triazolam administration, Brain Res., 589 (1992) 253-261. Vedder, H., Serum-free culture of AtT 20 pituitary cells: a system for neuroendocrine studies under defined conditions, In Vitro Cell. Det'. BioL, 26 (1990) 1068-1072. de Wet, J.R., Wood, K.V., DeLuca, M., Helinski, D.R. and Subramani, S., Firefly luciferase gene: structure and expression in mammalian cells, MoL Cell. Biol., 7 (1987) 725-735. Wiedemann, K., Lauer, C., Loycke, A., Pollm~icher, T., Durst, P., Mach~r, J.-P. and Holsboer, F., Antiglucocorticoid treatment disrupts endocrine cycle and nocturnal sleep pattern, Eur. Arch. Psychiatry Clin. Neurosci., 241 (1992) 372-375. Wright, N.A., Belyavin, A., Borland, R.G. and Nicholson, A.N., Modulation of delta activity by hypnotics in middle-aged subjects: studies with a benzodiazepine (flurazepam) and a cyclopyrrolone (zopiclone), Sleep, 9 (1986) 348-352. Wu, F.-S., Gibbs, T.T. and Farb, D.H., Pregnenolone sulfate: a positive allosteric modulator at the N-methyl-o-aspartate receptor, Mol. Pharmacol., 40 (1991) 333-336.