Abstract

Thih study examined the effc:c~s of ncuro~teroicl~ on Ihe beha\ ior of mice in the mirn,red chamber ~.tht of ;mxicty. and determined the potential mechanisms hy which ~~urostcron.l~ 11hcr the behavior n animal moJclh 01‘ .m\icty Alloprefrr~nololie ( AP) (0.5 anJ 1 mg/hg) and prcgnenolonc sulfate (PSI (0.5 and 3 mg/kg) significantly reduced the latency to enter the chamber. and increased both number of entries and total time spent in the chamber in u dose-dependent mlmne?. without affecting the spontaneous locomotor activity. In contmht, deRydroepiandrosterone sulf;lte (DHEASJ ( I and 2 mg/kp) increased motor activity and caused an anxiogenic response, i.e., an increase in latency 10 enter the mirrored chamber, and a decrease in the number of entries and time spent in the chamber. Progesterone (PROG) (I-10 mg/kg). ;1 neurosteroid precursor. and 4’-chlordiazepam ($-CD) (025-l mg/kg). a specific lipand for the mitochondrial diazepam binding inhibitor (DBI) receptor (MDR). produced a clear dose-dependent anxiolytic response in the mirrored chamber. The

AP-. PROG- and 3’-CD-elicited anxiolytic behavior was blocked by picrotoxin (I “g/kg), it GABA-A chloride channel antagonist. hut not by tlumaLcnil (7 mg/kg). a selective benzol9iazepine (BZD) antagonist. In contrast, the anxiolytic effect of PS was not blocked by picrotoxin. The 4’-CD-induced anxiolytic effect W;L~ prevented by pretrracment with PKI I I95 (1 mg/kp). a selective partial MDR antagonist. Nifedipine (1 and 5 mg/kg). ;I dihyrlropyridine-type Co’ + channel blocker. produced a tlumazenil-resistant unxiolytic effect. Combined ndniinistnrtion of nifedipine (2 illld S nip/kg) and PS (05 Uncl 2 my/kg) exerted a significant additive cffec~ in the mirrored chamber test. The potcni anxiolytic cffecl of‘ dizocilpinc (0.5 and I mg/kg). an NMDA receptor anlupmist. was hloched by pretreutmen[ with DHEAS (7 mg/hp). Neurobterilidb c\‘c)kerl changes in mirrorcd chamber octiviiiec raembling those elicirccl by triarolam (0.3 anJ 0.5 mg/kp). However. Ihcsc efl’ect~ wcrc WCII at LIOWS that JiJ not marhcdly affect locomotor acti\-ity. thereby su,, ~~~~csting Ihcsc changes

in behavior represent anxiolyrlc action\. Togcthcr. thc~ rcsuha pr~~viiJc cviclencc I’ijr tliffi’rsntiai h&aviorJ action\ of the neurok!eroiJs AP, PS rind DHEAS in lhe mirrored ohnmber test of anxiety. The unxiolytic effect ot PROG may be impu&J lo ita metabolism to

neurosteroid AP. while the 4’-CD-induced anxiolytic response is related to its MDR-stimulated neurosteroidogenesi\ and subsequent modulation of GABA-A receptor. Further. these differential effects reaffirm the contention that neurosteroids could he involved in the homcoslasis of stress response. 8 I997 Elsevier Science B.V. All rights reserved.

1. Introduction

Neurosteroids synthesized de nova in the brnin include

allopregnanolonc CAP), allotctrahydrodeoxyc:orticosterone (THDOC), prcgncnolone. prcgncnolone sulfate (PS) and dehydroepiandrosterone sulfate (DHEAS) [27,33.47]. while the metabolites of’ progesterone and deoxycorticosterone. 3 rw-hydroxy-5 a-pregnane-20-one and Sty-pregnane-3 a- 21 a-dihydroxy-20-one, have been propr.)scd us endogenou5 neuromodulators [33]. Recently. a pathway for neuros-

’ Corresponding nuthor. Fax: +Yl ( 172) %I 132.

teroidogenesis has been shown in brain [47]. and a role for mitochondrial diazepam binding inhibitor IDBI) receptors (MDR) in the regulation of neurosteroid biosynthesis has been proposed [25,42]. Multiple lines of evidences suggest that ncurosteroids mediate their action through y-amino- butyric acid type A (GABA-A) receptors [38]. The GABA- A receptors are known to contain allosteric modulatory sites i‘or therapeutically useful drug:; such as benzodi- azcpines (BZD) and barbiturates. It is now well recognized that certain neurosteroids such as AP. THDOC and progcs- teronc (PROG) can potentiate, whereas PS and DHEAS inhibit the GABA-A chloride channel responses [21,28,3()].

tMM)6-XYY~/Y7/%I7.O0 C?opyrighl 8~3 iYY7 Elsevier Science R.V. All right\ rr*ervod. P/I s0006-sYY~(Y6~01~47-3

Those neuroactive steroids, which are positive illlosteric modulators cf the GABA-A receptor acting via a unique site on the receptor complex, have been termed ‘epalons’ [20]. These epalon neurosteroids. like other GABA-A al- losteric modulators, could have anxiolytic, sedative and anticonvulsant activity [3.9.24], whereas the GABA-ant- agonistic neurosteroids have been implicated in the physio- logicai stress response [33.45,46].

Recently, the neurosteroid AP has been shown to be a potent anticonvulsant [2] and anxiolytic in the Vogel test [55], light/dark transition test [54], burying behavior test [43] and elevated plus-maze test [3]. THDOC has been shown to attenuate stress-induced increases in plasma cor- ticosterone [4l]. and produce an anxiolytic effect in the light/dark transition test and Vogel test 191. PROG has hen shown to elicit an anxiolytic etfect in the defensive burying test by metabolizing to allopregnanolone 161. In the elevated plus-maze test, pregnenolone has been shown to cause an anxiogenic response. whereas PS showed a biphasic response [36]. The anxiolytic actions of these neurosteroids have been proposed to occur through its ability to potentiate GABA-A receptors, although the mechanisms underlying these effects are as yet unclear. Furthermore, recent studies have shown that some neuros- teroids can also inhibit glycine-chloride channel [27]. in- hibit voltage-dependent Ca’+ channels 114,521 and en- hance N-methyl+aspartate (NMDA) receptor-mediated excitatory amino acid responses [8.56]. These various ac- tions could lead to differcli:,?s among the neurosteroids in their anxiolytic profile.

The present study was aimed to explore the behavioral effects of neurosteroids in yet L”?:)ltler test of anxiety, the mirrored chamber paradigm. and to compare the effects with the proto:ypicnl short-acting benzodiazepine. triazo- lam. Further attempts were made to determine the potential mechanisms tncdiating the anxiolytic/anxiogcnic eflcct of neurosteroids, particularly the roles of MDR and CABA-A receptors, and Ca? + channels using specific ligands such u PKI 1195, a selective MDR antagonist, picrotoxin, a GABA-regulated chloride channel blocker, flumazenil, a selective BZD antagonist, and nifedipine, a Ca’+ channel blocker, respectively.

2. Materials and methods

2.1. Attitttuls

Male BALB/c mice (Institute of Microbial Technol- ogy. Chandigarh) were used in the study. The animals were housed five per cage and acclimatized to the labora- tory conditions for at least 2 days prior to the initiation of any experiment. Food and water were available ad libitum, except during testing. The experiments were carried out lmwn 9:OO and 17:oO h at the ambient temperature. Mice were only administered drug or vehicle once.

AP. DHEAS. and PS (Sigma Chemical Co.. USA). PROG (Unichem. Bombay, India). flumazenil (F. Hoff- mann-La Roche, Basel. Switzerland). triazolam (Upjohn. Kalazamazo, USA). picro!oxin (BDH, UK). nifedipinc (Bayer, Leverkusen, Germany), dizocilpine (MSD. UK). 4’-chlordiazepam (4’-CD) and PK I I I95 (RBI, Natick. USA) were used in the present study. Drug solutions were made in saline, except progesterone which was dissolved in vegetable oil. DHEAS and PS were dispersed in 0. I c/r Tween and diluted with saline. Triazolam was dissolved in a few drops of diluted HCI and then dissolved in waler (pH 7.5). whereas flumazenil was dissolved in a few drops of dimethyl sulfoxide and then in water. The AP and 4’-CD were prepared as suspension in a vehicle of saline containing 0.5% Tween 80.

All drugs were injected intraperitoneally 30 min prior to the experiment. except PROG which was injected subcuta- neously 60 min prior to the testing. In rhe interaction studies, flumazenil and PK 11195 were administered 30 min before the neurosteroids. Control groups received the respective vehicle treatment and all the control obscrva- tions were pooled since no significant differences were found in control studies with animals which received different vehicles. The selection of doses was based on our previous studies [46].

The behavioral et’fect of ncurostcroids was assessed using the mirrored chamber anxiety test as described by Toubas t’~ al. [SJ], whit h m~asurt!s approach-contlict hc- havior [ I7.261. The chamber consists of a mirrored cube open on one side that is placed inside a square wooden box. The mirrored cube measuring 30 cm on a side was constructed of five pieces of mirrored glass with one mirrored side and an opposite side painted dark brown. The three mirrored side panes, a top pane, and the tloor pane faced the interior of the cube. The container box was 40 X 40 X 30.5 cm. The mirrored cube was placed in the center of the container to form a 5 cm corridor that completely surrounded the milTur chamber. A mirror was also placed on the container wall so that it faced the single open side of the mirrored chamber. Except for this one mirrored portion on the container wall. all portions of the container walls were dark brown.

Group-housed mice were brought into the testing room and allowed to acclimate to the new environment for at least 30 min. Mice were placed in the chamber of mirrors and scored oace. Mice injected with vehicle or drug were placed at a single. fixed starting point in a corner of the chamber. The mouse was allowed to explore the chamber for 5 min. In this paradigm, three parameters are recorded: (a) latency to enter the chamber, i.e.. the time in seconds for the first entry into the chamber of mirrors. (b) number

of entries, an (cl k&d tillle in seconds sptxl in 11162

chamber during the test. The critcriun t;tr entry into ihc

chamber was al1 four feet being placed on the floor panel of the mirrored chamber. After exploration by each animd.

the kor wab cleaned thoroughly to eliminate pi~tentiul cuts such as excrete and urinntion left hy the previous test

animal. Since behavioral changes were reported after the first experience in this paradigm [26], all animals were tested only once.

2.3. Locor?llmr- rirtil~itY tr.st

The effects of neurosteroids on motor activity were evaluated with an Opto-Varimax automated animal activity meter (Columbns h~strumcnts, Ohio, USA). The two vani- ables recorded are ambulatory and total activities. After the mirrored chamber test the animal was individually placed in a Plexiglas cage (9-Y X 21 X 32 cm) and the ambulatory and total activities registered for ;I 5 min period [Ih]. An

array of IS itlltared emitter/detector pairx (spaced at 2.65 cm intervals; beam wavelength 875 nm; distance between the sensor 50 cm) measured the animal activity along LL single axis of motion. the dipital data being d~spluycd on the front panel meters as ambulatory and total clctivity. After each test the cage was carefully cleaned. The motor activity was expressed in terms of total photo bean1

counts/S min per minial.

2.5. stiati.sfic~.s

Ah results a-~: cxprcsscd 3s mc’an + S.E.M. Dais were sihjrctcd 11.) one-way aiinlysis ol’ varimce (ANOVA) lill- lowed hy Duiloail’a new illllltipl~ range lcsl, the criterion for significance being P < 0.05. A correlational analysis of the three measures of the mirrored chamber activity was also carried out.

3. Results

The behavioral effects of neurosteroids in mice assessed by the mirrored chamber test are shown in Fig. 1. The average latency to enter the mirrored chamber by the vehicle-treated mice was I87 + Y.8 s. The vehicle-trentcd mice entered the mirrored chamber less frequently (1.75 + 0.37) and spent ;1 mean total time of 13 + 3.2 s in the chamber during the S-nun test period. Triaznlnm (0.35-0.5 mp/kg) treatment significantly reduced the latency to enter the mirrored chamber and increased both the number of cntries ;1s welt as the time spent in the mirrored chamber, AP (0.5 und I mg/kg) produced ;t marked and

Latency For Entry Iseconds) boo /

200 T

150 {

100

50

Number of Entries 14 3

Total Time in Chamber (secor?ds)

a0

60 j

40

20

n I- :.; ii. .: , ”

VENICLE 0 25 0.5 0.5 1 0.6 2 TRIAZOLAM AP PS

dose-dependent decrease in the latency to enter, and in- creased the number of entries and time spent in the mir- rored chamber us compared to vehicle-treated controls. PS

(0.5 and 2 mg/kg) exhibited a shortened latency to enter the mirrored chamber, and also increased both the number of entries and the time spent in the chamber in a dosage- dependent fashion. In contrast, DHEAS (1 and 2 mg/kg) signiticantly increawd the latency to enter the chamber. and also dec.reased both the number of entries as well as the time spent in the mirrored chamber. a behavioral profile opposite to that observed with triozolam and AP.

The anxiolyl:ic effect of AP was blocked by the con- comitant treatment with picrotoxin (1 mg/kg), but not by pretreatment with flumazenil (2 mg/kg) (Fig. 2). In con- trast. the anxiolytic effect of PS in the mirrored chamber test was. resistznt to picroloxin (Fig. 2). The dose of picrotoxin used in these experiments was previously shown to reverse the &stress effects of neurosteroids without producing convulsant activity [46]. The selective benzodi- azepine antagonist, tlumazenil (2 mg/kg), did not signifi- cantly affect the mirrored chamber activity when adminis- tered alone (data not shown).

As shown in l?g. 3, administration of PROG (I-10 mg/kgI, a neurouctive slcroid i\nd precursor fw ncuroh- tcroids. markedly shorlcnud the time lakcn IO cntcr the mirrored chamber in a dosage-dcpcndcnl manner. PROG trealment si@ficnntly increiiscd Ihe number of cntrics und

time spent in rhc mirrored chamber compared to vehicle conUds. 4’-CD (0.?5- 1 mp/kg), iI MDR receptor agonist and neurosteroid inducer, produced a dose-dependent de- crWsc in fhe latency to enter lhc mirrored chamber and 81~0 augmented both the number of entries and the time spent in the chamber as compared to the vehicle-treated group.

The interaction between the neurosteroidal agents PROG and @-CD. and tlumazcnil. picrotoxin or PKI 11% was in:rs?;tigated, As is evident from the data presented in Fig. 4. IfIe unxiolytic effect of PROG in Ihe mirrored chamber test was blocked by picrotoxin (1 mg/kg). but no1 by pretreatment with either flumlzenil(2 me/kg) or PK~ 1195 (2 q/kg). However. the anxiolytic effect of #-CD was S~llnifiC~nt~Y reversed by both picrotoxin and PKI I 105, but Was r&runt lo inhibition by fiumazenil. The selective

Latency for Entry (seconds)

b

J.:. :..r :

VEHICLE AP f’s

Number of Entries 12,

a

101 T

I r .’ /

a -r

a

I

6

b

T . . . . / ,jj/:: : ..::.:;: :.:::: '.. :::::: : /:j ..I:: ::: :: "' ::. . . . . . . . a.. VEHICLE AP I’S

Total Time in Chamber (suconds) I a

1ool T

80 -

60

40

20

01 __ VEHICLE

a T a

; a - -1

b :-x::. :.:::::::. :,:.:::::’ /ij( !:j’j “’ :“. .: ::::. :::. .’ .:, .:, ,.j;,,j:.; :“““’ . I

MDR partial antagonist. PKI 1105 (2 mg/kg). did not

affect the mirrored chamber behavior of mice when dmin- istered alone (data not shown).

Latency for Entry k+econds)

40

1 5 10 0.25 0.5 1 mg/kg

Number of Entries 121

I

8

6.

2

1 6 10 0.26 06 1 mg/kg

,ot,Total Time in Chamber (seconds)

00

20

0 ~--‘- VEHICLE 1 6 10

PROG 0.25 0.5 1 mglkg

4’-CD

Latency for Entry (seconds) 300

(]I. VEHICLE

a

PIIOG

Total Time in Chamber (seconds) 100

a 00

60

40

20

O! VEHICLE

a

b T .!. ::.

, PROG

receptor nntagonist. elicited ;I dose-dependent anxiolytic activity in the mirrored chamber test. However. the anxi- olytic effect of dizocilpine (0.5 mg/kg) W;IS significantly

Latency for Entry (seconds)

Latency for Entry (seconds)

200 I T b

200

160

120

0 : _.. VEHICLE

60

40

6 2 mufku

a

Number of Entries 12

I

6- 10

6

6

4

2

O! Vr3llCLE

b

1

i 1

2 6 2 0.6 1 0.6 m6lk6

BOTotal Time In Chamber (seconds1

ooTotal Time in Chamber (seconds)

I 407

a b

30

20 i)

10

0. -.l._.~ ---- VEHICLE ’ 2 5 2 0.6 1 0.6 Wku NIFEDIPINE DIZOCILPINE

Fig. 5. Effect of nlfcdipine and dizofilpine (MK SOI) on the mirrored chimlhcr activity in mice. Flum;lzcnil (2 me/kg) and DHEAS (2 mg/kg) Wcrc injcctcd intrapritoncillly 30 min prior to nif&Iipinr: (2 q/kg) ild

disucilpinc t0.S mg/kpl trcdmcnth. rc\pcrtivcly. Each h;u wprcrcnrt the I~ICUII f S.E.M. of S-$4 unimal~ per gmup. ’ P < (WI compercd IO vchi- clc-IrWWl group: P < 0.W when ccanparcd 10 dirtrilpine (05 mg/kg&utcd mice.

2 Wku

dihydmpyridine (L-type) calcium channel blocker, pro- duced ;P significant anxiolytic activity in the mirrored chamber test. which was flumozenil (2 mg/kg) resistant. Dizocilpine (0.5 and I mg/kg). a non-competitive NMDA

600

450

300

150

0 VEHICLE 0.5 1.5 0.5 2 1 2 5 10 2 5 1 hg/kg)

AP PS ocims PRQG NlFE DIZOC

i

blocked by pwtrc;itment with the ncurr~steroid DHEAS (2 mg/kg).

Further. nn interaction hc-twccn the non-nc’iirostcroid nit’cdipinc and the nc’urostcroid PS was invc‘htigat&. Ah is cvidcnt I’ron~ Lllc data pIX3~tccl in Fig. 0. cornhind ildlllill-

islriilion of low closes 01’ I’S (0.5 iy!/lq) ;uid nifcdipinc (1. mg/kg9 IIXIIIC~ in ;I ~ignil’icant &litivc cl’l’ect \V~ICII ci~mprcil lo llrc el'lixlh 01' cithcr ol’ thcsc ~rgciils adrninia- tercd done. Howcvcr, administration of higher doses ol’ PS (2 mg/kg9 and nifedipinc (2 or 5 ~ng/kg) t~g~‘tiwr did not produce it further additive 4li2ct as coinlxtrcd lo PS (2 mg/kg9 nltme, but it was stutis&cnlly significant compared to the effect ol’ nit’cdipinc (3 and S mg/kg) alone.

One of the major concerns IXpildillg mcastircnient of potential anxiolytic actions of drugs is whcthcr Clliul~CS in activity rcprcsent anxiolytic actions or cffccts on loconio- tor activity. Therefore. the ncurosteroid effects on locomo- tor rtclivity wcrc nicasured with lhc Opto-Varimax oppitrii- tus. Thehc results arc depicted in Fig. 7. AP itt the high dose (I.5 mg/kg9 suppressed both the ambulatory ml total activity, whilst the lower dose (0.5 111g/kg) hiid 11~1

cl’fcct on the motor activity wlw co~npar~d to vchiclc controls. PS had no cl’tixt on the locomotor activity in the dose rangtz tested. DHEAS increased the motor activity above the level of the vehicle controls. Propestcrone did not significantly alter the motor behuvior of animals, ex-

cept at the higher dose (IO mg/kg9. A high dose ot nit’edipine (S mg/kg9 suppressed both ambulatory imd total activity. whereas a lower dose had no effect. In contraht. ;L significant illClI?ilSC in locomotcn sctivity Wilb hcen li,llowing the dizocilpinc ( I mgJkg9 administration.

4. Discussion

SCVUill lines Of rcccnt ill\oStigiltiO~~ hiI\-C Shown that

ncurc~~lc‘roids ;rlic&rcrically nl~~Juli\tc (iABA through ;I unicluc’ binding \itc OII IIK GABA-A rcccpcor (‘1 h1nncl

[ I8,lc).28.30,33]. This binding site has been few cl to ho distinctly diffcrcnt from that ol’ the benzcldinxfipirt. or the barbiturate site. rcspcctively [I X-2 I ,2X-301. The 3cuhy- droxy ring A-reduced metabolites 01’ PROG an 1 DOC. such as AP and THDOC. produced a potent bnrbitu ‘ Itc-like enhancement of GABA-A receptor respca~.~s I vitro [X33] and induced behaviornl sedation wnen a&~unistcred in vivo [2.3]. w;iile PS and DHEAS hnve’becn SIIOWII to inhibit GABA-A Cl conductanccs [ IO.3 I]. PS also antiI$- oni the gl)TillC-ilCtiVilld currents [57]. tAlilllCl3 tllC NMDA-p&J current [Se] and the NMDA receptor-mcdia- ted excitatory responses [K] in cultured neurons. The IICU- rostcroids prcgnenolonc. t~tr~hydmcorticostcrol,e and PS were rcportcd to inhibit the voltage-gated Ca” channels i!l is<jlittcd neurons [ I4,!52]. AP IWS bee11 SIW\~II to bC a pnteiit i\nti~ollvulsilnl [2], anxiolytic [3,54.55] alld neuro- protective ilgcnt [ 161. More recent studies have dcmon- struted thut ThlDOC has anxiolytic [‘,I and an:iaggressive [??I properties, and thnt PS reduces pcntobarbital-induced sleep time [Xl. DHEA has been shown to abolish aggrcs- si\~encss in cnstr:d male rats [~g]. and impr:sve memory

in nt~r& Inice as well us mice With age-induced memory deficits [37]. Furthermore. a role of neurosteroids has been inlplicated in pathological states such as catamenial epilepsy [IS], stress [4I ,4X46] and dementia 1401; they may also have potential therapeutic roles as anXiOlytiC Or

anticonvulsant agents [3.9,i2.48]. Triuolam produced an antianxiety effect in the mir-

rored chamber paradigm, a result similar to that showed by dimepam [53]. The anxiolytic behavior of neurosteroid AP in the mirrored chamber paradigm was in line with other behavioral studies with AP [3,43,54,55], and supports the increased explomtory activity observed in the elevated plus-maze after intracerebroventricular administration of AP [3]. Moreover, differences in the anxiolyiic effects of the two endogenous neurosteroid isomers AP and epipreg- nanolone have been demrnstrated in various models of anxiety [S4]. However, the decrease in motor behavior has been argued against a non-specific effect of AP at higher doses [24]. In the present study- direct support for the coc:ention that the anxiolytic effect of AP is mediated via potentiation of GABA-A receptor was found in the obser- vations that co-administration of picrotoxin prevented the effects of AP in the mirrored chamber test. These data are comparable to those previously reported for the picrotoxin antagonism of the AP-induced anxiolytic effect in rodents [3]. Since, pretreatment with flumazenil had no effect on the anxiolytic actions of AP in the mirrored chamber test, it is unlikely that AP’s psychotropic effect is due to the occupation of the BZD site on the GABA-A receptor complex. Neurosteroids, like BZDs. produce their action through the GABA-A receptor mnplex but they act on a binding site which is distinct from the BZD binding site [l8,29]. Although a putative ‘epalon’ site, at which the ncurosteroids act as positive allosteric modulators. on the

GABA-A receptor complex has been proposed 1191. the cum?nt understanding of this neurosteroid ‘epalon’ binding site is yet unclear.

PS produced an anxiolytic effect in the mirrored cham- ber test without affecting motor behavior. It has been demonstrated that PS also exhibited anxiolytic activity in the elevated plus-maze test [36]. Picrotoxin, a GABA-A receptor Cl- channel blocker, failed to block the effect of EL PS has been shown to possess a mixed aeonist/antagonist action at the GABA-A receptor [33]. In high concentration it blocks the GABA-A receptor-media- ted currents 1381. and at low doses PS shows an agonistic Profile 1311. In another study, higher doses of PS produced no effect on pentylenetetraidol-induced convulsions [34] and did not ulter ethanol-induced sleep [7]. In the present study. a relatively high dose of PS (2 me/kg) was em- ployed and it exhibited an anxiolytic response. It appears that this response may not involve the GAB&A receptor- coupled Cl - channel [7,34.36].

PS has been shown previously to inhibit the voltage- gated Ca’ + current in hippocampal cells [14.X!], and there is evidence that calcium channel inhibitors have an anxi-

olytic effect [44]. particularly without influencing the ilnxi-

olytic action of ethanol in the elevated plus-maze test [23]. In addition. low doses of nifedipine (2 mg/kg), a dihydro- pyridine Ca’+ channel blocker, that do not alter the motor activity. elicited a tlumazenil-insensitive anxiolytic effect in the mirrored ~ilii~iib~r tcsi. Further, combine:! ;r&!i+.- tration of lower doses of PS and nifedipine exerted a marked additive effect in the mirrored chamber test. When a maximal anxiolytic dose of PS was given, nifedipine produced no further effect following combined exposure. This suggested that the anxiolytic profile of the neuros- teroid PS involved the same mechanism as that of nifedip- ine, i.e.. through modulation of voltage-gated calcium channels in its anxiolytic response in the mirrored chambel paradigm.

In contrast to the anxiolytic effects of AP and PS. the increased latency to enter tbe chamber. and the reduction in the number of entries and the total time spent in the mirrored chamber produced by DHEAS is consistent with the auxiogenic-like effect of this neurosteroid in an animal model of anxiety. The GABA-A receptoy antagonistic properties of DHEAS [27,33] could explain these anxio- genie-like effects. Recently, Melchior and Ritzmann [35] have reported that DHEAS attenuates the anxiolytic effect of ethanol that would support the present results with DHEAS. Moreover, DHEAS has recently been reported to be a positive allosteric modulator of the NMDA receptor complex at the sigma site [39], which may possibly con- tribute to the anxiogenic responses observed with DHEAS. This action may be partly explained by the observation that dizocilpinc. a non-competitive NMDA receptor antag- onist, exhibited a potent anxialytic effect in the mirrored chamber test. Pretreatment with the neurosteroid DHEAS blocked the anxiolytic cffict of dizocilpine. which further strcngtltens the hypothesis that apart from GABA-A recep- tors, NMDA receptors may be at least partly involved in the action of DHEAS, and thus in the neuronal pathway of neurosteroidal anxiogenesis.

The anxiolytic effect of PKOG was in parallel and comparable to that of AP in the mirrored chniilber test. In an earlier study, it has been shown that PROG exhibited antistress actions via its possible in vivo conversion to AP, which was not related to its :i~tracellular progestin receptor actions [46]. Several lines of recent evidence have shown that the anticonvulsant and anxiolytic effects of proges- terone are mediated by its metubolite AP [5.6.13,48]. and it is therapeutically used in certain endocrine states such as the menopause [SO] and premenstrual tension [I I]. Al- though it is known that PROG potentiates the GABA responses both in vitro [57] and in vivo [5l], it appears to be a poor modulator of GABA A $:.z, eptor function [4]. In

addition, Bitran et al. [4.6] have observed that blockade of the bioconversion of PKOG to AP prevents the anxiolytic effect and potentiation of cortical GABA-A receptor func- tion in rats. On these bases, in the mirrored chamber test, it is likely that PROG induces the anxiolytic effect by indi-

rect action on GA A-A RXCptW B‘UllCtion. nmdiattd by ils

reduced metabolitc. A In the present study. -I’- rxbibited m snxiolytic

effect in the mirrored chan ’ test. ~~ctreatl~le~lt with PKI 1 I95 blocked the anxiolytic effects of -$‘-CD in ~hc mirr1ired chamber icst. Thcsr re?xllls p!“ovidL- i?;dkx! c\ i- dence for a possible cot~tribu~i~~~ of an increase in neuros- teroid biosynthesis to the action MDR @and 4’-CD. Recent evidence suggests that R receptors reguiatc ncurosteroidogenesis [35,42]. Act on of MDR receptors by selective ligands such as 4’-CD and KIN- l-21 potently stimulates the neurosteroid biosynthesis in brain [I ,351, Thus, the anxiolytic effect of #-CD may be attributed to its agonistic action on MDR-medi d steroidogenesis, and the subsequent modulation of 6 ergic transmission at the GABA-A receptor possibly via stimulation 01’ mite- chondrial synthesis and release of neurosteroid AP. This contention is further supported by the ability of AP to mediate a similar response in the mirrored chamber paradigm, in a manner resistant to llumazcnil and picrcr- toxin inhibition, because of the direct action of AP on ihe neurosteroid site of the GABA-A receptor complex.

The cllanpes in mirrored chamber activities could sirn- ply retlec; ChllgCS in locomotor activity. Therefore, we were concerned that drug actions on locomotion could influence set!eral of the mirrored chamber measures which may confound interpretation of drug effects on the mir- rored chamber test of anxiety. To prevent this ambiguity, we have also examined the effects of nemrosteroid and non-neurostcroid drugs on general locomotor activity. Howewr. in the ~IVWI~I study, comparing the effects 01’ dil fereut aecnts WI locomotor perfonnancc and in the mirrored chanlbcl~ test iudicatcs tllilt the anl;iolytic/;unxlr,- penic cffccts of lhc drugs wcrc not colllimnded hy CI’IiXI5

on locomotion. Moreover, the cviclcnce a\ ailablc from the literature suggests that general motor activity WI be delin- eated from the anxiety response in animals using drug CIICS. Bitran et al. [3] II;IW demonstrated that. although picrotoxin per se had no effect on the plus-maze test ol anxiety, it significantly decreased moior activity; and fol- lowing the combined exposure to picrotoxin and preeg- nanolone, a pregnane neurosteroid. the effect of picrotoxin on locomotion was blocked, and also the effect of preg- nanolone on plus-maze performance. However, it may be possible that when administered in doses higher than the doses tested in the present study, these neurostcroids pro- duce motor impairment [24]. Overall, the present findings indicate that the neurosteruid-induced changes in mirrored chamber activity in mice do not arise because of changes in locomotor activity.

In conclusion, the present study demonstrated the dif- ferential behavioral actions of the neurosteroids AP. PS and DHEAS in the mirrored chamber test of anxiety. The AP-elicited anxiolytic effect was picrotoxin-sensitive and resistant to flumazenil inhibition, and most likely mediated by potentiation of GABA-A receptor function via the

The Senior Research Fellowship (to D.S.R.) of the Council of Scientific and Industrial Research CCSIR). New Delhi, and tlw samples of PK I I 1% and If-CD provided by

I, Natick as part of the Chemical Synthesis Progain 01‘ the U.S. National Institute of SIental Health. Contract NQI MH300(13. arc gratef~llly achnowledgcd.

[ ls] Finn, D.A. and Gee, K.W.. The inllurncc of estrus ~ele on neuro\-

teroid potency 111 Ihe gunma-aminoburyric acid A receptor complex.

J. P/mwrad. F-p. T/w.. 265 ( 1993) I374- 1379. [ 161 Frye, C.A., The neurosteroid 3u.Sn-THP ha antisazure and possi-

ble neurnprotectivc cffcas in an animal model of epilepsy. Bruin

Rrs., 6Y6 (190s) II&170. [ 171 Gullup Jr.. G.G.. Mirror-inqc s~imt~l;!~ion. Pv\rhrrl. h/f.. 70 ( 1468)

782-793. [la] Gee. K.W.. Bolger. M.B.. Brinton. RX.. Coirini. H. and M&ten.

B.S., Steroid modulation of the chloride ionophore in rut brain: strucmre activity requirement\. regional drpendencc and mechanism

of ncrion. J. f/~ctmcrcrr/. hp. T/w.. X6 ( IYXX) X03-8 II!. [IY] Gee, K.W.. McCauley, L.D. and Lan. NC.. A putative receptor for

neurostcroidk on the GABA-A rccrptor complcu: the pharmacolc$- cul properties and therapeutic potential of epalons, Cn’f. /&I.. Nc~rro- bid, 9 t199sj X7-227.

[?(I] Goodnough, D.B. and Hnwkinson, J.E.. Ncuronctivc steroid moduln- lion of [‘H]muscimol binding to tlx GABA-A receptor eomplcx in

rat cortex, Ecrr. /. P/rarftnrt~m/.. 28X (IYYS) lS7- 163. [XI] Hawkinson. J.E.. Kimbrough. C.L., MeCaulcy. L.D.. Bolgcr. M.B..

Lan, N.C. md Gee. K.W.. The neurosctive steroid 3u-hydroxy-SF- pregnun-Xl-one is G two-component modulntor of lignnd binding to

the GABA-A receptor, Effr. J. P/fcfrff~mf/.. 269 ( iYY4) I S7- 163. (221 Kavalierr. M., Inhibitory inlluencca of rhr: udremd steroid 3tr .Scr.te-

~ruhydmxyeoFticostcronr. on aggrchsion rrnd defeat-induced nnalge-

4iu in mice, I’s~c~/~c~/~/rtr~rfrtrc’c~/r~.~~. OS ( 1%) JXK--lY!. 1231 Kiruc. R. and Eroplu. I... Bay K-X&I4 pocnti:itc$ the ;ulxiolyhL

effect of etbi~nol. F’/~mw~trtwl. Awhr. &4ctr ., 30 ( I’M I ) 3W 34.

[?J] KOL~IIC. T.G., Svcn*\on. B.E. and Roguwski. M.A., Anticonvulrunt icrivily of ncurohtcroids: corrrl;ition with ~illlllllil~~Illill~lblI1>ri<

il<irl=cvok~tl chloride current polcnfi;ltioll. 1. /‘/~ctrrwcd. Et/>. 7’h.r..

219 (I!#&) t”t,i”% Cv4I “” * [?S] KorneyQv, A,, Pun, US.. Polo. A.. Romeo, E.. Guidolti. A. and

Coslu. E.. Srimulation of bruin pregncnrku~c synthcsib by mitochon- dksl CiiiiqXun binding inhibitor recephrr liyunds in \iuo, J. Nru- rfwbtvr,.. bl ( 19Y.I) IS IS- 1524.

(261 Kt:lkomi. S.K. und Rcddy. U.S.. Animal behnvioml nu~dels for testing antianxiety agency ~\kt/&t Fiful, Ey. CXfl. P/f~frffl~~ff/. , IX

ii9Y6) 119-130.

1291 Lnn. N.. Chen. J.S.. Bulclli. D.. Pritchett. D.. Sechurg. P.H. & Gee.

K.W.. A steroid recognition \ite is functionully coupled 10 iul cxprcsscd GABA-A/‘~nsrdi;lr.cpillc receptor. EIW. J. ~/,trntrcrt o/ , lxx (1990) 40%.406.

(331 Mnjewsku, M.D.. Neurosteroids: cndogenous bimodnl modulntors of the GABA-A receptor. Mecb;mism of action and phyhiologic;il

hignitirnncc. Prog. Nfwrobifd. 38 ( 1992) 37Y-305. [3-t] Mclchior. C.L. imd Allen. P.M.. Interaction of prrnenolonc sullale

wit11 ethanol and pemobarbit;~l. Phrfffcrcw/. Bioc hon. /I~/f~w.. 11

(IYYZ) 605-61 I. 1351 Mclchior. C.L. and Ritzmann. R.F.. DehgJrclepiandnr~t~rl~ne i\ an

anxiolytic in mice on the plus-rwzc. Pknmnco/. Bioc&fr. Hrhw..

47 (lYY4) 437441.

(361 Mclchior. C.L. and Ritzmunn, R.F.. Pregnenolone and prcgnenolone sulfate, alone and with ethanol, in mice on the plus-maze. Phrffm-

Id. BirJc/fcffl. fMfff~.. 18 ( IY9-U 893-897. [37] Melchior. C.L.. Glasky. A. and Ritzmann. R.F., Dchydroepiandros-

teronc (DHEA) improves memory in normul mice and mice with age-induced memory deticith. Strc,. Nc~rrrrasr+. ABstr.. IX ( IYYZ) 1660.

[38] Micnville. J.M. and Vicini. S.. Pregncnolonc sulfate antagonizes GABA-A receptor-mediated currents via ;I reduction of chnnnel

opening frcqucnvy. Rnrifl RCJS.. 3X9 ( IWO) IYO- 194. 1.~~1 Monnet. F.P.. M&c. V., Rohcl. I’. and Bnulicu. E.E.. Ncuroatcroids.

vi0 sigma receptors. tt10dUlilt~ the [ ’ ~l]iior~pin~phrIIic rrlcahc cvokcd hy N-methyl-l~~spart:ite in the rat hippocampu\. Froc. I%/. r\cat/.

ki. UJA. ‘12 (IYYS) 3774-3774.

[-IO] Na~m;m. 9.. Olswn. T.. Bockstrom. T.. Erikwm. S.. Grankvist. K.. Viitancn. M. aid Bucht. G.. Srrum drliydrocpiandrt~stcro~ic wlfatc

in Alzhcimcr’s disease und in multi-intarcl demenun, Hiol. /-‘.g~~/rrcc-

try. 30 (1991) 6X-L00. [-II ] Owens. M.J.. Ritchie. J.C. and Nemerof. C.B.. Scu-Prepnane-3u.21-

dial-%-one (THDOC) nttenuates mild xtress-induecd increases in

plasrw corticosteronc via u non-glucocortietid mechanism: compnri-

bon with nlprazolam. fhiff ROT.. 373 (1002) 353-355. [42] P;rpadopoulob. V.. Guameri. P.. Krueger. K.E.. Guidotri. A. and

Cosk~. E.. Prcgnenolone biosynthesis in C6-29 glionra cc/l mito-

chondrie: regulation by mitochondri;d dialep:un hinclmg inhibitor rcccpw. Prw. Mrd Acw/. Sci. USA. W ( lcJc)?_) 45 I 13-45 I 19.

[~3] Picazo, 0. and Fcrnandez-Guuhti. A.. Anti-nnxicty effects of proge+

teronc and MW of its reduced mctaholitcs: a11 ewhmtion wing the lwv~t~g behavior tea. BrcrLr /h.. OX0 t 10%) IX?- 141.

[J-8) Puciltrwhki. (, . P~)I. Ilopliuriiiacol~)~ienl propcrlirh of calcium CIUIII-

WI inhibilor~. /‘\~(,/to/~/f1fr~fII1,.o/o,~~ I Ilwlirr). IO0 ( I W.?) I?-- IO.

[45] Purtly. R.H.. Morrow. A.I... Moore Jr.. PH. ;III~ Pinil. S.M.. SIrah-

induce11 eIeviuion\ ol’ F;II”‘Ilil-illllillohlltyri~ ilcld type- 4 ilcliyc

!&roidh in rut bruin, I”,rrc~. Nc&. &rfl. Sci. USA. 8X (1991) 4SS3- X5.57.

[XI] Rcddy. D.S. and Kulkarni, S.K.. Role of CAl3A-A rind mitochon- drial di;rzepam bindiag inhibitor receptors in the nnlistress activity of neurosteroids in mice, ~.s~c~/lff/f/fcfrfffff(~ff/~~~~~~ I% ( 1996) (in press).

(471 Robel. R. and Bnulieu. E.E., Ncurosteroids biosynthesis and time- tion, Trcfrdv fk/ffc~rifro/. hfcW/f.. 5 ( llY4) I-X.

(4x1 Rodriqurz-Sicrm, J.F.. Howard. J.L.. Pollard, G.T. and Hendricks, SE.. Effect of ovarian hormones on contlict behnvior. P,s,v&f~t~rr- rrrc,fff/ffr.riftfi/ffs~, Y ( 19X4) 293-300.

[SO) Shcrwm. B.B. ;uitl Gclhnd. M.M.. A prospective one-ycur xtudy of e~!rngs~~ and progchtin in postn~unopnusni wonien: effects on chnlciri \)rnptoni\ ;Intl lipoprotein lipids. (I/U% (&rc:c.o/.. 73 ( IOXYD 75% 766.

15 I ] Smidi. S.S.. Wn~erhou~c. B.D.. Chnpin. J.K. and Woodward. D.J..

Propsrcrone alters GABA rind gluUnK\tc msponsiveness: ;I possible tncchanisnr for itz, snxiolytic action. Hrtliff /h.s.. 400 (1987) 759-766.

:sy SpwiL i;.‘i’.. PI&o-Salanla~~. C.R. and Ffrench-Mullen. J.M.H.. The

neurortcri~ids prcgncnolone mid pregnenrlloile-iulfilte but not progcb- Icrone. block calcium currents il l acutely isolated hippocnmpnl CA I

neuron\. Lite Sci.. 49 t !YYl 1 X35-231~.