Steroids 70 (2005) 515–524

Altered profiles of serum neuroactive steroids in premenopausalwomen treated for alcohol addiction

Martin Hill a,∗, Petr Popovb, Helena Havlıkovaa, Lyudmila Kanchevaa, Jana Vrbıkovaa,Radmila Kanchevaa, Vladimır Pouzarc, IvanCerny c, Luboslav Starkaa

a Institute of Endocrinology, N´arodnı Trıda 8, CZ 11694 Prague 1, Czech Republicb Department of Addiction Treatment, General Faculty Hospital, Charles University, Apolin´arska 4, CZ 12808 Prague 2, Czech Republic

c Institute of Organic Chemistry and Biochemistry, Flemingovo N´am. 2, CZ 16610 Prague 6, Czech Republic

Received 21 October 2004; received in revised form 14 January 2005; accepted 14 February 2005Available online 28 March 2005

Abstract

Long-term alcohol consumption results in menstrual irregularities due to the inhibition of progesterone secretion. Some progesterone metabo-l e (DHEAS)i e formationi ctions andp(p ring thet OVA withs

creased ratioo y restoredt n ovarians ysiologicalr nt of serume chosomatics©

K

ites, including three pregnanolone isomers (PI), abate, while pregnenolone sulfate (PregS) and dehydroepiandrosterone sulfatncrease, alcohol tolerance. The rationale of this study was to evaluate how the neuroactive steroids reflect the impaired progesteronn premenopausal women treated for alcohol addiction, and whether detoxification therapy could restore female reproductive funsychosomatic stability by reinstatement of the steroid biosynthesis. Accordingly, serum allopregnanolone (3�-hydroxy-5�-pregnan-20-oneP3�5�)), pregnanolone (P3�5�), isopregnanolone (P3�5�) and epipregnanolone (P3�5�), progesterone, PregS, pregnenolone, 17�-hydroxy-regnenolone (Preg17), 17�-hydroxy-progesterone (Prog17), DHEA, DHEAS, cortisol and estradiol were measured in 20 women du

herapy (start, 3 days, 14 days, 1 month, 4 months), and in 17 controls, using GC-MS or RIA and evaluated by age-adjusted ANCtatus and phase of the menstrual cycle (PMC) as factors, and status–PMC interaction.The patients exhibited depressed progesterone, Prog17, PI, and estradiol, a decreased progesterone/pregnenolone ratio, a de

f neuroinhibiting P3�5� to neuroactivating PregS, and an elevated PregS and PregS/pregnenolone ratio. The treatment mostlhe indices. The reduction of neuroinhibiting pregnanolone isomers in the patients is primarily associated with the impairment iteroid biosynthesis. Nevertheless, changes in enzyme activities connected with the formation of PI and the influence of altered phequirements on the balance between endogenous neuroinhibiting and neuroactivating steroids are also likely. The reinstatemestradiol, progesterone, and PI during the therapy demonstrates its favorable effect on both reproductive functions and the psytability of the patients.2005 Elsevier Inc. All rights reserved.

eywords:Alcohol detoxification; Pregnanolone isomers; Neuroactive steroids; Pregnenolone sulfate; GC-MS; Premenopausal women

1. Introduction

∗ Corresponding author. Tel.: +420 224 905267; fax: +420 224 905325.In addition to a variety of other deleterious effects,

excessive alcohol ingestion in women is associated withasee

trast,ica-usald by

E-mail addresses:mhill@endo.cz (M. Hill),petr.popov@vfn.cz (P. Popov),hhavlikova@endo.cz (H. Havlıkova),lkancheva@endo.cz (L. Kancheva),jvrbikova@endo.cz (J. Vrbıkova),rkancheva@endo.cz (R. Kancheva),vladimir.pouzar@uochb.cas.cz (V. Pouzar),ivan.cerny@uochb.cas.cz (I.Cerny),lstarka@endo.cz (L. Starka).

menstrual irregularities including anovulation, luteal phdysfunction, recurrent amenorrhea and early menopaus[1].No effect of alcohol intake on gonadotropin synthesis[2–6]has been reported in most of the studies to date. By conprolactin levels are stimulated at the beginning of intoxtion [4]. Concerning serum sex steroids in premenopawomen, estradiol levels have been seen to be elevate

0039-128X/$ – see front matter © 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.steroids.2005.02.013

516 M. Hill et al. / Steroids 70 (2005) 515–524

acute alcohol intake[3–7]. Although some studies havereported no influence of alcohol intake on basal progesterone[2–6], most of these have included low subject numbers,resulting in the low power of the corresponding statisticaltests. The more recent study of Sarkola et al.[7] evaluatinga greater number of subjects has reported a suppressionof progesterone levels and an elevation of estradiol levelsafter acute low-dose alcohol intake, particularly duringthe luteal phase of the menstrual cycle (MC), although theeffect was not dose-dependent. The authors assumed that theelevation of the estradiol/estrone ratio and the suppressionof progesterone levels was connected with alcohol-inducedchanges in the redox state of the liver enzymes providingreversible conversions of estradiol to estrone and proges-terone to 20�-dihydroprogesterone, and that both systemsare catalyzed by 17�-hydroxysteroid dehydrogenase type 2,sharing the same cofactor (NAD) as alcohol dehydrogenase[8]. Even earlier studies reported similar progesteroneeffects during stimulatory conditions among premenopausalwomen not using oral contraceptives[9,10]. Contradictoryresults have been reported in a studies of female adolescents[11] and of groups of light to moderate drinkers[12,13].

In contrast to moderate alcohol drinkers, premenopausalwomen with a history of early alcohol abuse have shownlower estradiol, progesterone, androstenedione and sex hor-mone binding globulin levels and higher levels of free testos-tb ls mab wthf ranu-l

reg-n rsorsp ctives s oft bilityo es-t pi vity[a nt rideu teh ande er[ holt nge thept nea tings

thes ntrols ol ad-

diction, how they reflect impaired progesterone biosynthesis,and to what extent therapy could restore female reproductivefunctions and psychosomatic stability by influencing steroidbiosynthesis. Accordingly, changes were evaluated in theserum levels of some reduced progesterone metabolites,including all such neuroactive PI as allopregnanolone(3�-hydroxy-5�-pregnan-20-one (P3�5�)), pregnanolone(3�-hydroxy-5�-pregnan-20-one (P3�5�)), isopregnano-lone (3�-hydroxy-5�-pregnan-20-one (P3�5�)) and epipre-gnanolone (3�-hydroxy-5�-pregnan-20-one (P3�5�)) in20 premenopausal women during alcohol detoxificationtherapy, and in 17 age-matched controls. In addition, thelevels of such corresponding precursors as progesterone,PregS, pregnenolone, 17�-hydroxy-pregnenolone (Preg17),17�-hydroxy-progesterone (Prog17), dehydroepiandros-terone sulfate (DHEAS), and estradiol were measured inthe 20 women during therapy (at start, 3 days, 14 days, 1month and 4 months) and in the 17 age-matched controls.The effect of MC was considered for all of the steroids.

The authors hypothesized that (particularly in the lutealphase) deficiency in PI reflects impaired progesteronebiosynthesis in women of fertile age treated for severealcohol addiction. There was also speculation as to whethertreatment could remedy the deficit in the formation ofprogesterone and its neuroactive reduced metabolites. Inaddition, the question was addressed as to which steps inn onica tiont

2

2

der-g ays,1 d 17h itteea in-f fromt MCw tientss s, thes t andd rone,e ibu-t ep-r es oft

e 9,1 s, 1m r thel 1, 9,1 ys, 1m ples

erone, suggesting their hyperandrogenic status[14]. It haseen suggested that the decrease in progesterone levee related to the alcohol inhibition of epidermal gro

actor-stimulated progesterone secretion from human gosa cells[10,15].

Reduced progesterone metabolites—including panolone isomers (PI) and one of the progesterone precuregnenolone sulfate (PregS)—are known as neuroateroids (NS), and are primarily effective as modulatorhe neurotransmitter receptors influencing the permeaf ion channels[16–22], while some also operate at prog

erone receptors[23,24]. While the PI with a hydroxy-groun the 3�-position are known to attenuate neuronal acti16,25]via positive allosteric modulation of�-aminobutyriccid receptors, type A (GABAA-r), the PI hydroxylated i

he 3�-position, exert an opposite effect, reducing chloptake stimulation by 3�-PI [26,27]. Pregnenolone sulfaas been reported to be a strong neuroactivatingxcitotoxic substance modulating theN-methyl-d-aspartateceptors (NMDA-r) positively and the GABAA-r negatively17,18,20]. In addition, PregS is known to enhance alcoolerance[28–30]. In terms of the different neuromodulatiffects of individual steroids, it is interesting to traceroportional changes between the 3�- and 3�-PI, between

he 5�- and 5�-PI, between individual PI and progesterond between the neuroinhibiting and neuroactivateroids.

The primary aim of this study was to investigate howerum levels of neuroactive steroids differ between coubjects and premenopausal women treated for alcoh

y

,

eurosteroid biosynthesis might be influenced by chrlcoholism as well as by subsequent alcohol detoxifica

herapy.

. Experimental

.1. Subjects

The patient group consisted of 20 female patients unoing alcohol detoxification therapy at 5 stages (start, 3 d4 days, 1 month and 4 months/termination of therapy) anealthy age-matched volunteers. The local ethical commpproved the protocol for the study. After signing written

ormed consent, the patients underwent blood samplinghe cubital vein. In all of the subjects the effect of theas considered. For practical reasons, the individual patarted therapy at various stages of the MC; neverthelestage of the menstrual cycle was known from the outseuring the treatment, and it was checked by progestestradiol, lutropin and folitropin assays. The wide distr

ion in respect of the MC enabled a relatively uniform resentation of both phases thereof in the individual staghe treatment.

The numbers of patients in the follicular phase wer1, 9, 9 and 7 on the beginning, after 3 days, 14 dayonth and 4 months of the treatment, respectively. Fo

uteal phase, the numbers of patients were and 17, 11, 10, and 10 on the beginning, after 3 days, 14 daonth and 4 months of the treatment, respectively. Sam

M. Hill et al. / Steroids 70 (2005) 515–524 517

were also obtained from the 15 and 17 control subjects infollicular and luteal phase, respectively. The all patients inthe trial were chronic alcoholics. From the total of 20 pa-tients, 18 were treated for the first time and two of themfor the second time. Two subjects undergone one episodeof prior detoxification. Neither patients nor control subjectshave taken any oral contraceptives before and during thetreatment.

2.2. Sample collection

The blood was allowed to clot for serum collection afterclot retraction. The serum was obtained after centrifugationfor 5 min at 2000×gat 0◦C. The serum samples were storedat−20◦C until analyzed.

2.3. Steroids and chemicals

The steroids were from Steraloids (Wilton, NH, USA).The solvents for the extraction and HPLC, and pyri-dine, were of analytical grade, from Merck (Darmstadt,Germany). The derivatization agent Sylon BFT was pur-chased from Supelco (Bellefonte, PA, USA). The inter-nal standard was synthesized using our recently publishedmethod[31].

2.4. Instruments

The GC-MS system was supplied by Shimadzu (Kyoto,Japan). The system consisted of a GC 17A gas chromatographequipped with automatic flow control, AOC-20 autosamplerand for the MS a QP 5050A quadrupole electron-impact de-tector with a fixed electron voltage of 70 eV. The liquid scintil-lation spectrometer was supplied by Beckmann Instruments(Fullerton, CA, USA).

2.5. Analytical methods

The PI were measured using a modified method publishedpreviously[32]. The first modification of the method was theuse of less steep temperature and pressure gradients, as fol-lows: 1 min high pressure injection at 120◦C and 100 kPafollowed by a pressure release to 30 kPa and a rapid linear gra-dient 40◦C and 8.5 kPa up to 220◦C and 51 kPa, then a slowlinear gradient 2.9◦C and 0.5 kPa up to 240◦C and 54.5 kPaand finally a rapid linear gradient 40◦C and 9 kPa up to 310◦Cand 70 kPa with a 2 min delay. The second modification wasthe substitution of 17�-methyl-3�,17�-androstenediol as aninternal standard for trideuterated dehydroepiandrosterone(d3-DHEA), added to the standard solution or to the samplein a 1 ng/�l concentration and recorded at an effective mass

Fu

ig. 1. Mass spectra of pregnanolone isomers. A QP 5050 A quadrupole elsed for the measurement.

ectron-impact detector from Shimadzu with a fixed electron voltage of 70 eV was

518 M. Hill et al. / Steroids 70 (2005) 515–524

of 307. The third change was the addition of pregnenolonemeasurement. The pregnenolone was recorded at the effectivemasses 298 and 398, and the former was used for further pro-cessing. The overall time taken for the analysis was 14.2 min.The retention times were 7.10, 7.25, 7.40, 7.60, 7.80, 8.71,and 8.73 min for d3-DHEA, estradiol, epipregnanolone, al-lopregnanolone, pregnanolone, isopregnanolone, and preg-nenolone, respectively. The final change was the substitutionof micro-extraction in the vials by a more rapid drying of thederivatization agent under a stream of nitrogen.

Pregnenolone sulfate, Preg17 and progesterone were mea-sured using the authors’ specific radioimmunoassays as de-scribed elsewhere[33–35]. 17�-Hydroxy-progesterone and

DHEAS were measured using RIA kits from Immunotech(Marseilles, France).

2.6. Statistical analysis of the data

The results were evaluated using ANCOVA with subjectstatus (control group and 5 stages of the treatment) and thephase of the menstrual cycle (PMC) as main factors, and sub-ject status–PMC interaction and age as a covariate. Given thevarious days of the menstrual cycle at the beginning of thetreatment in the individual subjects, it was impossible to en-sure the same phase of the MC for all of the subjects at theindividual stages of the treatment. Accordingly, the data was

F4or

ig. 2. Comparison between a sample of female serum and standard solung of each of the steroids under study. Panel B shows the response of af the menstrual cycle. E-diol, P3�5�, P3�5�, P3�5� and P3�5� indicate estradespectively.m/z indicates the effective masses of the fragments and the num

tion. Panel A demonstrates the response of a standard solution (4�l) containingsample (4�l) corresponding to 200�l of serum from a patient in the luteal phaseiol, epipregnanolone, allopregnanolone, pregnanolone and isopregnanolone,bers in parentheses denote multiples of the original responses.

M. Hill et al. / Steroids 70 (2005) 515–524 519

treated not using the more powerful repeated measures modelbut with regular multi-directional ANCOVA. The originaldata were transformed by a power transformation to attaina Gaussian distribution and a constant variance of studen-tized residuals. The experimental points after the transfor-mation with absolute values of studentized residuals greaterthan 3 were excluded from calculations; such points neveraccounted for more than 5% of the total. The ANCOVA wasfollowed by Bonferroni (control versus stage of the treatment)multiple comparisons. Statgraphics Plus v.5.1 from Manugis-tics (Rockville, MA, USA) and NCSS 2000 from NumberCruncher Statistical Systems (Kaysville, Utah, USA) wereused for the calculations.

3. Results

3.1. Detection of the steroids

The mass spectra of the PI are shown inFig. 1. The steroidsrecorded on the chosen fragments in SIM were well separatedboth from each other and from the background (Fig. 2). Thesensitivity was sufficient for the quantification of all of the in-vestigated steroids. The detection limits for all the pregnane-neurosteroids were below 0.5 pg and 1 mL of serum was usedf

3.2. Unsaturated steroids and estradiol

As demonstrated inFig. 3A, PregS levels were elevatedin the patients in comparison to the controls. A similar sit-uation appeared in unconjugated pregnenolone but was lesspronounced (Fig. 3B). The changes in the aforesaid steroidsduring the treatment were not significant. No differences be-tween the patients and controls and no changes during thetreatment, were observed in Preg17 (data not shown). Bycontrast, 3-oxo-4-ene steroids, progesterone (Fig. 3C) andProg17 (Fig. 3D) in the luteal phase exhibited significant de-pendence on the subject status, this being more pronouncedin progesterone. Progesterone levels in the controls in theluteal phase were significantly higher than in the patients atthe beginning of treatment, but no such difference was foundwhen comparing the patients after 4 months of treatment.In progesterone, the character of the differences significantlydiverged in individual PMC as indicated by the significant in-teraction between the factors subject status and PMC. In bothPMC, estradiol showed a similar pattern of differences to theprogesterone in the luteal phase but with less pronounceddifference in the follicular phase (Fig. 4). The DHEA andDHEAS levels did not differ significantly between alcoholicsubjects and controls and they did not show any change dur-ing the treatment as well as the levels of cortisol (the data notshown).

F women , with errbscAa

or the analysis from each subject.

ig. 3. Serum levels of unsaturated C21-steroids in controls and in

ars, represent the retransformed mean values with their 95% confidence iignificance andR2 is the squared correlation coefficient expressing the peronsisting of subject status (controls or stage of treatment) and phase of thNCOVA was followed by Bonferroni (control vs. stage of the treatment) msterisk.

during alcohol detoxification treatment. The empty and dotted barsor

ntervals in the follicular and luteal phases respectively;p is the level of statisticalcentage of the total progesterone variability explained by an ANCOVA modele menstrual cycle as factors, inter-factor interaction and age as a covariate. Theultiple comparisons and significant differences (p< 0.05) are labeled with an

520 M. Hill et al. / Steroids 70 (2005) 515–524

Fig. 4. Serum levels of estradiol in controls and in women during the alcoholdetoxification treatment. The drawings and symbols are the same as forFig. 3.

3.3. Pregnanolone isomers

The differences in pregnanolone isomers did not simplyreflect the variation in progesterone but were more prominent(Fig. 5); particularly in the 5�-PI (Fig. 5A and B). The dis-similarity between the controls and the untreated patients wasrelatively less pronounced in the P3�5� (Fig. 5C). Moreover,in contrast to progesterone and similarly as in estradiol, thecharacter of the differences in PI did not diverge between thefollicular and the luteal phase.

3.4. Steroid ratios

To identify in which steps of neurosteroid biosynthesisthe alterations occur in women treated for alcohol addiction,the product/precursor ratios were traced in the pathway frompregnenolone and PregS to PI. As is apparent fromFig. 6,the ratio of PregS to unconjugated pregnenolone showedhigher values in the patients but with a tendency to compen-sation during the treatment. The progesterone/Prog17 andpregnenolone/Preg17 ratios showed similar patterns in theluteal phase, with significantly decreased values at the begin-ning of treatment but with a trend to compensation duringtherapy.

In addition to the product/precursor ratios, the overallratios of the 3�- to 3�-PI and of 5�- to 5�-PI were alsoevaluated, as were the proportions between neuroinhibiting(and most abundant) PI allopregnanolone and neuroactivatingPregS. The ratios of the 3�- to 3�-PI exhibited no significantdifferences between the controls and patients, even duringtreatment. By contrast, the ratios of the 5�- to 5�-PI showeda trend to lower values in the patients and again a trend tocompensation during therapy (Fig. 7A). The ratio of the fore-most neuroinhibiting PI allopregnanolone to neuroactivatingPregS exhibited markedly depressed values in the patientsbut with a trend to incomplete but significant compensation(Fig. 7B).

Fa

ig. 5. Serum levels of pregnanolone isomers in controls and in women durs forFig. 3.

ing the alcohol detoxification treatment. The drawings and symbols are the same

M. Hill et al. / Steroids 70 (2005) 515–524 521

Fig. 6. Serum steroid product/precursor ratios of unsaturated C21-steroids incontrols and in women during alcohol detoxification treatment. The drawingsand symbols are the same as forFig. 3.

4. Discussion

The rationale of the study was to evaluate how the neu-roactive steroids reflect impaired progesterone formation inpremenopausal women treated for alcohol addiction. A fur-ther question was as to whether the therapy could restore thefemale reproductive functions and psychosomatic stabilityby reinstatement of the steroid biosynthesis (Fig. 8). Serumpregnanolone isomers, progesterone, pregnenolone and itssulfate, Preg17, Prog17, DHEAS, DHEA, cortisol and estra-diol were therefore evaluated in 20 women during therapy (atstart, 3 days, 14 days, 1 month and 4 months) and in 17 age-matched controls. As shown inFig. 2, a highly sensitive andselective GC-MS method was used for the analysis of the PI,

unconjugated pregnenolone, and estradiol. Pregnenolone sul-fate, progesterone, Preg17, Prog17, and DHEAS were mea-sured by specific radioimmunoassays. Given the possible agedependencies as well as the effect of the menstrual cycle, two-way ANCOVA with subject status and phase of menstrualcycle (PMC) as factors and status–PMC interaction and ageas a covariate was used for evaluation.

In the case of progesterone (Fig. 3C) and its reducedmetabolites (Fig. 5), the resulting data were in accordancewith those in the study of Sarkola et al.[7], which reportedthe suppression of progesterone levels after acute low-dosealcohol intake in the luteal phase of the menstrual cycle, andwith those in the study of Pettersson et al.[14], which re-ported lower progesterone levels in premenopausal womenwith a history of early alcohol abuse. In addition, the newresults showed the compensating effect of the therapy in pro-gesterone (Fig. 3C), Prog17 (Fig. 3D), and particularly the PI(Fig. 5), of which the P3�5� exhibited the least pronounceddifferences—possibly due to its rapid metabolism[36–38].The return of progesterone and PI to their physiological levelsduring alcohol detoxification therapy could be explained bythe regeneration of progesterone formation impaired by alco-hol abuse, or by the recovery of enzyme activities changed byalcohol intake as reported by Sarkola et al.[7]. Given a similarmechanism in the effects of the alcohol and steroid activa-tors of GABA -r, a third possibility could be the adaptation

Fig. 7. I5�/� characterizing the overall conjugated/free isomer ratio of serumpregnanolone isomers (defined as the square root of the ratio of product of the5�- to 5�-isomers) and the allopregnanolone/pregnenolone sulfate ratio incontrols and in women during alcohol detoxification treatment. The drawingsand symbols are the same as forFig. 3.

A

522 M. Hill et al. / Steroids 70 (2005) 515–524

Fig. 8. Simplified scheme of the biosynthesis of pregnanolone isomers. The symbols + and− denote the increase and decrease of serum steroids in patients,respectively, when compared to controls. The symbol “?” indicates missing information.

of the organism to a rising demand for GABAA-r activatingsubstances due to the cessation of alcohol intake.

The opposite, albeit less pronounced pattern to estro-gens, progesterone and its reduced metabolites during ther-apy was indicated in PregS (Fig. 3A) and pregnenolone lev-els (Fig. 3B). In addition, the ratio between PregS and preg-nenolone (Fig. 6A) was higher in the patients, and at thesame time the ratio of the foremost neuroinhibiting serumsteroid allopregnanolone to PregS (Fig. 7B) was markedlydepressed in patients, showing an incomplete reinstatementafter the therapy. As noted above (PregS) is efficient neuroac-tivator acting via negative modulation of GABAA-r and pos-itive modulation of NMDA-r. It is also very likely that serumsteroids, or at least their precursors and metabolites, more orless penetrate the blood–brain barrier and influence the over-

all concentration of brain neuroactive steroids, as do the brainneurosteroids synthesized in situ[39–42]. Given these find-ings, one may speculate as to the influence of altered phys-iological requirements on the proportion of neuroinhibiting(i.e. alcohol tolerance reducing) and neuroactivating (alco-hol tolerance enhancing) endogenous steroids. The renewalof the ratio of 5�-PI to 5�-PI in the patients during the therapy(Fig. 7A) also supports such speculation.

Given the renewal of the ratio of 5�-PI to 5�-PI (Fig. 7A),the more pronounced changes in 5�-PI during treatmentin comparison to progesterone (Fig. 3C), and the signifi-cant differences in both PMC in contrast to progesteroneshowing differences in the luteal phase only, a restorationof the farther progesterone metabolism to PI is also likely.A similar pattern as in the progesterone and its metabolites

M. Hill et al. / Steroids 70 (2005) 515–524 523

was found in estradiol (Fig. 4), even though the restora-tion was incomplete. The latter finding suggests that therestoration of ovarian activity during therapy is not lim-ited to progesterone and its metabolites; it involves estro-gens as well. Estrogens are also neuroprotective[43], andtheir lack may be connected to signs of metabolic syndrome[44].

Concerning the adrenal activity, we have completed thecortisol measurement. The latter steroid showed no changeduring the treatment and no significant differences betweenpatients and controls were detected as well. The results in-dicate no significant alterations in adrenal activity of thepatients. Formerly elevated cortisol levels were reported inchronic alcoholics[45] but the results were in controversywith other reports and some authors did not recommend touse the cortisol as an indirect marker of chronic alcoholism[46,47].

Given the stress-reducing and neuroprotective effectsof allopregnanolone[48–50], its increasing profile duringtherapy (Fig. 5A) could contribute to an improvement inthe psychosomatic stability of the patients. On the otherhand, given that the P3�5� (Fig. 3B) and P3�5� (Fig. 3D)levels exhibited a concurrent increase with the levels ofP3�5� (Fig. 3A), their opposite effect on GABAA-r [26,27]should be considered. Nevertheless, the GABAA-r inhibit-ing action of unconjugated 3�-PI is probably less intenset ,i[ ef-f osae

it-i ausalw oci-a ver-t eronem al re-q s andb ctiva-t nismso . Ther lone,p m leve ablee oso-m ohola s notc

A

In-t hee -f

References

[1] Hugues JN, Coste T, Perret G, Jayle MF, Sebaoun J, ModiglianiE. Hypothalamo-pituitary ovarian function in thirty-one women withchronic alcoholism. Clin Endocrinol (Oxf) 1980;12:543–51.

[2] McNamee B, Grant J, Ratcliffe J, Ratcliffe W, Oliver J. Lack ofeffect of alcohol on pituitary-gonadal hormones in women. Br JAddict Alcohol Other Drugs 1979;74:316–7.

[3] Valimaki M, Harkonen M, Ylikahri R. Acute effects of alcohol onfemale sex hormones. Alcohol Clin Exp Res 1983;7:289–93.

[4] Mendelson JH, Mello NK, Cristofaro P, Ellingboe J, Skupny A,Palmieri SL, et al. Alcohol effects on naloxone-stimulated luteiniz-ing hormone, prolactin and estradiol in women. J Stud Alcohol1987;48:287–94.

[5] Teoh SK, Mendelson JH, Mello NK, Skupny A. Alcohol effectson naltrexone-induced stimulation of pituitary, adrenal, and gonadalhormones during the early follicular phase of the menstrual cycle. JClin Endocrinol Metab 1988;66:1181–6.

[6] Mendelson JH, Mello NK, Teoh SK, Ellingboe J. Alcohol effectson luteinizing hormone releasing hormone-stimulated anterior pi-tuitary and gonadal hormones in women. J Pharmacol Exp Ther1989;250:902–9.

[7] Sarkola T, Makisalo H, Fukunaga T, Eriksson CJ. Acute effect ofalcohol on estradiol, estrone, progesterone, prolactin, cortisol, andluteinizing hormone in premenopausal women. Alcohol Clin ExpRes 1999;23:976–82.

[8] Andersson S, Moghrabi N. Physiology and molecular genetics of 17beta-hydroxysteroid dehydrogenases. Steroids 1997;62:143–7.

[9] Teoh SK, Mendelson JH, Mello NK, Skupny A, Ellingboe J. Alcoholeffects on hCG-stimulated gonadal hormones in women. J PharmacolExp Ther 1990;254:407–11.

[ ect in-ecific

[ reg-arma-

[ nAm J

[ n JF.uctivepre-

[ tionMed

[ inhi-an

[ ig-ymp

[ loneteroid07–

[ al-ar-

[ DB,e

[ tor-amo-75–

han the activating effect of P3�5� [26,27]. Moreovern avian CNS, P3�5� acted in a similar way to P3�5�51–54]. In addition, an alcohol tolerance-inhibitingect of P3�5� has been reported in a study by Barbt al.[30].

In conclusion, it is likely that the reduction of neuroinhibng pregnanolone isomers in the serum of premenopomen treated for alcohol addiction is primarily assted with ovarian steroid biosynthesis impairment. Ne

heless, changes in the formation of neuroactive progestetabolites and the influence of decreased physiologicuirements for endogenous neuroinhibiting substancey contrast the increased requirement for steroid neuroa

ors should also be considered. All the same, the mechaf the corresponding feedback loops remain in questioneinstatement of estradiol, progesterone, allopregnanoregnanolone and perhaps also epipregnanolone seruls during detoxification therapy demonstrates its favorffect on both the reproductive functions and the psychatic stability of premenopausal women treated for alcddiction, even though in some indices the restoration iomplete.

cknowledgements

This study was supported by grant NB/6891-3 of theernal Grant Agency of the Czech Ministry of Health. Txcellent technical assistance of Mrs. Ivona Kralova is grateully acknowledged.

-

10] Saxena S, Meehan D, Coney P, Wimalasena J. Ethanol has dirhibitory effects on steroidogenesis in human granulosa cells: spinhibition of LH action. Alcohol Clin Exp Res 1990;14:522–7.

11] Torres JM, Ortega E. Alcohol intoxication increases allopnanolone levels in female adolescent humans. Neuropsychophcology 2003;28:1207–9.

12] Martin CA, Mainous III AG, Curry T, Martin D. Alcohol use iadolescent females: correlates with estradiol and testosterone.Addict 1999;8:9–14.

13] Garcia-Closas M, Herbstman J, Schiffman M, Glass A, DorgaRelationship between serum hormone concentrations, reprodhistory, alcohol consumption and genetic polymorphisms inmenopausal women. Int J Cancer 2002;102:172–8.

14] Pettersson P, Ellsinger BM, Sjoberg C, Bjorntorp P. Fat distribuand steroid hormones in women with alcohol abuse. J Intern1990;228:311–6.

15] Wimalasena J, Meehan D, Dostal R, de Silva M. Selectivebition of luteinizing hormone action by ethanol in cultured humgranulosa cells. Alcohol Clin Exp Res 1993;17:340–4.

16] Majewska MD. Steroid regulation of the GABAA receptor: land binding, chloride transport and behaviour. Ciba Found S1990;153:83–97.

17] Majewska MD, Demirgoren S, London ED. Binding of pregnenosulfate to rat brain membranes suggests multiple sites of saction at the GABAA receptor. Eur J Pharmacol 1990;189:315.

18] Wu FS, Gibbs TT, Farb DH. Pregnenolone sulfate: a positivelosteric modulator at theN-methyl-d-aspartate receptor. Mol Phmacol 1991;40:333–6.

19] Hawkinson JE, Acosta-Burruel M, Kimbrough CL, GoodnoughWood PL. Steroid inhibition of [3H]SR 95531 binding to thGABAA recognition site. Eur J Pharmacol 1996;304:141–6.

20] Poisbeau P, Feltz P, Schlichter R. Modulation of GABAA recepmediated IPSCs by neuroactive steroids in a rat hypothalhypophyseal coculture model. J Physiol (Lond) 1997;500:485.

524 M. Hill et al. / Steroids 70 (2005) 515–524

[21] Irwin RP, Maragakis NJ, Rogawski MA, Purdy RH, Farb DH, PaulSM. Pregnenolone sulfate augments NMDA receptor mediated in-creases in intracellular Ca2+ in cultured rat hippocampal neurons.Neurosci Lett 1992;141:30–4.

[22] Park-Chung M, Wu FS, Purdy RH, Malayev AA, Gibbs TT, FarbDH. Distinct sites for inverse modulation ofN-methyl-d-aspartatereceptors by sulfated steroids. Mol Pharmacol 1997;52:1113–23.

[23] Putnam CD, Brann DW, Kolbeck RC, Mahesh VB. Inhibition ofuterine contractility by progesterone and progesterone metabolites:mediation by progesterone and gamma amino butyric acid A receptorsystems. Biol Reprod 1991;45:266–72.

[24] Rupprecht R, Berning B, Hauser CA, Holsboer F, Reul JM. Steroidreceptor-mediated effects of neuroactive steroids: characterization ofstructure–activity relationship. Eur J Pharmacol 1996;303:227–34.

[25] Gerak LR, Stevenson MW, Winsauer PJ, Moerschbaecher JM. Ef-fects of pregnanolone alone and in combination with other positiveGABAA modulators on complex behavior in rats. Psychopharmacol-ogy (Berl) 2004;173:195–202.

[26] Lundgren P, Stromberg J, Backstrom T, Wang M. Allopregnanolone-stimulated GABA-mediated chloride ion flux is inhibited by 3beta-hydroxy-5alpha-pregnan-20-one (isoallopregnanolone). Brain Res2003;982:45–53.

[27] Prince RJ, Simmonds MA. 5Beta-pregnan-3 beta-ol-20-one, a spe-cific antagonist at the neurosteroid site of the GABAA receptor-complex. Neurosci Lett 1992;135:273–5.

[28] Barbosa AD, Morato GS. Pregnenolone sulfate, dehydroepiandros-terone sulfate and allotetrahydrodeoxycorticosterone affect rapidtolerance to the hypothermic effect of ethanol. Brain Res Bull2002;58:99–105.

[29] Barbosa AD, Morato GS. Influence of neurosteroids on the devel-opment of rapid tolerance to ethanol in mice. Eur J Pharmacol

[ reg-chem

[ R.preg-

[ l.uring

den-mers.

[ i-ctantroids

[ ettions

62.[ im-

l Lett

[ s ofn. J

[ gernewlogy

[38] Carl P, Hogskilde S, Lang-Jensen T, Bach V, Jacobsen J, SorensenMB, et al. Pharmacokinetics and pharmacodynamics of eltanolone(pregnanolone), a new steroid intravenous anaesthetic, in humans.Acta Anaesthesiol Scand 1994;38:734–41.

[39] Bixo M, Andersson A, Winblad B, Purdy RH, Backstrom T. Pro-gesterone, 5alpha-pregnane-3,20-dione and 3alpha-hydroxy-5alpha-pregnane-20-one in specific regions of the human female brain indifferent endocrine states. Brain Res 1997;764:173–8.

[40] Bixo M, Backstrom T, Winblad B, Andersson A. Estradiol andtestosterone in specific regions of the human female brain in differentendocrine states. J Steroid Biochem Mol Biol 1995;55:297–303.

[41] Zhu D, Wang MD, Backstrom T, Wahlstrom G. Evaluation and com-parison of the pharmacokinetic and pharmacodynamic properties ofallopregnanolone and pregnanolone at induction of anaesthesia in themale rat. Br J Anaesth 2001;86:403–12.

[42] Wang MD, Wahlstrom G, Backstrom T. The regional brain dis-tribution of the neurosteroids pregnenolone and pregnenolone sul-fate following intravenous infusion. J Steroid Biochem Mol Biol1997;62:299–306.

[43] Sortino MA, Chisari M, Merlo S, Vancheri C, Caruso M, Nicoletti F,et al. Glia mediates the neuroprotective action of estradiol on beta-amyloid-induced neuronal death. Endocrinology 2004;145:5080–6.

[44] Spencer CP, Godsland IF, Stevenson JC. Is there a menopausalmetabolic syndrome? Gynecol Endocrinol 1997;11:341–55.

[45] Gianoulakis C, Dai X, Brown T. Effect of chronic alcohol consump-tion on the activity of the hypothalamic–pituitary–adrenal axis andpituitary beta-endorphin as a function of alcohol intake, age, andgender. Alcohol Clin Exp Res 2003;27:410–23.

[46] Hasselbalch H, Selmer J, Sestoft L, Kehlet H. Hypothalamic–pituitary–adrenocortical function in chronic alcoholism. Clin En-docrinol (Oxf) 1982;16:73–6.

[ Gal-JM.Ann

[ ionducedRes

[ atesneu-

[ bani-e in

[ ofte, andrain

[ atedneu-

[ by 3dingchem

[ partialA(A)–5.

2001;431:179–88.30] Barbosa AD, Morato GS. Effect of epipregnanolone and p

nenolone sulfate on chronic tolerance to ethanol. Pharmacol BioBehav 2000;67:459–64.

31] Cerny I, Pouzar V, Budesinsky M, Bicikova M, Hill M, HamplSynthesis of [19-2H3]-analogs of dehydroepiandrosterone andnenolone and their sulfates. Steroids 2004;69:161–71.

32] Hill M, Parizek A, Bicikova M, Havlikova H, Klak J, Fait T, et aNeuroactive steroids, their precursors, and polar conjugates dparturition and postpartum in maternal and umbilical blood: 1. Itification and simultaneous determination of pregnanolone isoJ Steroid Biochem Mol Biol 2000;75:237–44.

33] Hill M, Hampl R, Lukac D, Lapcik O, Pouzar V, Sulcova J. Elimnation of cross-reactivity by addition of an excess of cross-reafor radioimmunoassay of 17alpha-hydroxypregnenolone. Ste1999;64:341–55.

34] Hill M, Havlikova H, Klak J, Bicikova M, Pouzar V, Hampl R,al. Rapid immunoassay for pregnenolone sulfate and its applicain endocrinology. Collect Czech Chem Commun 2002;67:140–

35] Langer L, Veleminsky J, Hampl R, Starka L, Holan J. Radiomunoassay of plasma progesterone. Radiochem Radioana1978;34:267–72.

36] Parivar K, Wessen A, Widman M, Nilsson A. Pharmacokineticeltanolone following bolus injection and constant rate infusioPharmacokinet Biopharm 1996;24:535–49.

37] Hering WJ, Ihmsen H, Langer H, Uhrlau C, Dinkel M, GeisslinG, et al. Pharmacokinetic-pharmacodynamic modeling of thesteroid hypnotic eltanolone in healthy volunteers. Anesthesio1996;85:1290–9.

47] Vallina Alvarez E, Villaverde Amieva P, Rodriguez Sanchez J,lego Villalobos M, Gonzalez Fernandez E, Arribas CastrilloHypophysio-adrenal axis in patients with chronic alcoholism.Med Interna 2002;19:626–8.

48] Dazzi L, Sanna A, Cagetti E, Concas A, Biggio G. Inhibitby the neurosteroid allopregnanolone of basal and stress-inacetylcholine release in the brain of freely moving rats. Brain1996;710:275–80.

49] Kajta M, Budziszewska B, Lason W. Allopregnanolone attenukainate-induced toxicity in primary cortical neurons and PC12ronal cells. Pol J Pharmacol 1999;51:531–4.

50] Lockhart EM, Warner DS, Pearlstein RD, Penning DH, MehraS, Boustany RM. Allopregnanolone attenuatesN-methyl-d-aspartateinduced excitotoxicity and apoptosis in the human NT2 cell linculture. Neurosci Lett 2002;328:33–6.

51] Matsunaga M, Okuhara K, Ukena K, Tsutsui K. Identification3beta,5beta-tetrahydroprogesterone, a progesterone metaboliits stimulatory action on preoptic neurons in the avian brain. BRes 2004;1007:160–6.

52] Gravielle MC, de Novara AM, Fiszer de Plazas S. GABA-stimulchloride uptake during avian CNS development: modulation byrosteroids. Int J Dev Neurosci 1998;16:469–75.

53] Viapiano MS, Fiszer de Plazas S. Comparative modulationalpha,5 alpha and 3 beta,5 beta neurosteroids of GABA binsites during avian central nervous system development. NeuroRes 1998;23:155–61.

54] Pignataro L, Fiszer de Plazas S. Epipregnanolone acts as aagonist on a common neurosteroid modulatory site of the GABreceptor complex in avian CNS. Neurochem Res 1997;22:221