Effects of Neurosteroid and Benz[e]indene Enantiomers on GABAA Receptors in Cultured Hippocampal Neurons and Transfected HEK-293 Cells

Neuropharmacology,Vol. 35, No, 9/10, pp. 1161–1168, 1996 Copyright01996 Elsevier ScienceLtd. All rights reserved Printed in Great Britain 0028-3908/96$15.00+ 0.00

P —

PII:S0028-3908(96)00035-4

Effects of Neurosteroid and Benz[e]indene Enantiomers on GABAAReceptors in Cultured Hippocampal Neurons and Transected HEK-293 Cells C. F. ZORUMSKI,l’2*L. L. WITTMER,l’3K. E. ISENBERG,l YUEFEI HU4 and D. F. COVEY4 ‘Departmentof Psychiatry; 2Departmentof Anatomy and Neurobiology; 3Departmentof Chemistty and 4Departmentof Molecular Biology and Pharmacology, WashingtonUniversi~ Medical School, St Louis, MO 63110, U.S.A. (Accepted 27 March 1996)

Summary-The effects of the enantiomersof the neurosteroid,3ct-hydroxy-5ct-pregnan-20-one (DHP),and the benz[e]indene, BI-1, on y-aminobutyric acid (GABA) responses were studied using whole-cell recording techniquesin culturedrat hippocampalneuronsand humanembryonickidneycells (HEK-293)transected with either al/12y2or ct6/?2y2GABA~receptor subunits.At 10vM, the (+)-enantiomersenhancedcurrentsgated by 2 pM GABA in all cells, whereas the (–)-enantiomers were significantly less effective. The enhancement of 2 pM GABA responsesin HEK-293cells transected with a6~2y2 subunitswas about half that of hippocampal neuronsor HEK-293cells transected with al/?2y2. The lower sensitivityof u6~2y2receptorsfor (+)-DHPand (+)-BI-1 is accounted for by their greater apparent affinity for GABA. When the GABA concentration was decreased to 0.5 pM to take into account the four-fold higher apparent affinity of a6/i’2y2receptors, these receptors exhibited enhancement similar to ulf12y2 receptors. These results indicate that both native and recombinant GABA~ receptors have enantioselective sites at which neurosteroids and benz[e]indenes modulate GABA responses,and that differencesin agonist affinity contribute to apparent differencesin steroid sensitivity among GABAAreceptors. Copyright 0 1996 Elsevier Science Ltd Keywords—Neurosteroids,GABA, enantiomers, transected cells,

actionsby interactingwith neurosteroidsites on GABAA receptorsin hippocampalneurons(Rodgers-Neameet al., 1992; Covey et al., 1993; Hu et al., 1993). The benz[e]indenes may be thought of as tricyclic steroid analogues in which the steroid A-ring has been opened and partially removed (Fig. 1). Recently we found that both neurosteroidsand benz[e]indenesexert their effects on hippocampal GABAA receptors enantioselectively, providing strong support for the existence of specific steroid binding sites on GABAA receptor proteins (Wittmer et al., 1995). Furthermore, steroid and benz[e]indene-inducedanesthesiain tadpolesand mice shows similar enantioselectivity,supportingthe hypothesis that effects on GABAAreceptors contributeto the behavioral effects of these agents (Wittmer et al., 1995). In cultured hippocampal neurons, the effects of neurosteroids and benz[e]indenes show considerable *To whomcorrespondenceshouldbe addressed,at: DepartmentofPsychiatry,WashingtonUniversityMedicalSchool, cell-to-cell variability (Rodgers-Neame et al., 1992). 4940 Children’s Place, St Louis, MO 63110, U.S.A. Tel: Although the mechanisms contributing to this hetero314362 8650; Fax: 3143628649. geneity are uncertain, differences in the molecular

Certain neurosteroids are potent modulators of GABA~ receptors (Lambert et al., 1995). At low concentrations, these agents potentate the actions of GABA by enhancing the flow of chloride ions through GABA~ ion channels. At higher concentrations, neurosteroids directly open chloride channels in the absence of GABA (Majewska et al., 1986). These actions are thought to result from the binding of steroidsto specificrecognition sites on GABAAreceptor proteins that are distinct from the sites at which benzodiazepines,barbiturates,loreclezole and picrotoxin act (Macdonald and Olsen, 1994; Lambert et al., 1995). However, the Iipophilicnature of the steroidsmakes it difficultto exclude that an effect on membrane lipids contributesto the receptor modulation. Certain benz[e]indenesalso appear to augment GABA

1161

1162

C. F. Zorumski et al,

a cytomegalovirus promoter (u1, a6 and y2s — pRc/ CMVTM;f12 — pBC12) 12-24 hr after plating using H H calciumphosphate–DNAprecipitation(Chen and OkayaHH HH ma, 1987;Wong et al., 1995,1996).Equal amountsof the HO’”@ & Ho P ~ appropriate (ml/32y2s,a6f12y2sor ul~2) subunit con(+)-DHP (+)-BI-I struct DNA (5 pg total) were mixed with 250 mM CaC12. The CaCIJDNA solutionwas added dropwiseto an equal COCH3 COCH3 volume of a HEPES-buffered saline solution (2 x H: tI ‘ stock:46mM HEPES, 280 mM NaCl, 1.4 mM Na2HP04, H ‘H H’H with pH adjustedto 7.06), bubbling air continuouslywith P HO H HO H a Pasteur pipette, to form a precipitate. About 1.5pg of (-)-B1-I total DNA in the buffered CaC12mixture was added to (-)-DHF’ each culture and cells were maintainedin this solutionfor Fig. 1. Structuresof the steroids and benz[e]indenesused in this 6 hr (Wong et al., 1995, 1996); the media was removed, study. the cultures gently rinsed twice with 37°C phosphatebuffered saline (pH =7.4) and fresh media added to plates. Isolated HEK-293 cells were used for electrocompositionof GABAAreceptors may be important.To physiological experiments 24-96 hr after transection. examine potential molecular contributionsto the effects Control cultures were exposed concurrently in the same of neurosteroids and to examine whether different fashion to the buffered CaC12mixture without DNA. GABA~ receptors have enantioselectivesteroid recognition sites, we examined the effects of one neurosteroid Electrophysiology and one benz[e]indene enantiomeric pair on GABA~ Voltage clamp recordings were obtained using the receptors in hippocampal neurons and HEK-293 cells whole-cell patch clamp technique and fire-polished transected with either alf12y2 or Lx6/32y2GABAA pipettes with 3–7 Mf2 tip resistances. Pipettes were receptor subunit combinations. We chose to work with routinely filled with a solution containing (in mM): 140 these GABAA receptor subtypes because prior studies CSC1,4 NaCl, 0.5 CaC12,4 MgC12,5 EGTA, 10 I+EPES, have shown marked differences between these receptors with pH adjusted to 7.3 using CSOH. At the time of in the potentiation of GABA responses by benzodiaze- experiment, the growth medium was exchanged for a pine agonists(Liiddenset al., 1990)and y-butyrolactones recordingsolutioncontaining(in mM): 140 NaCl, 4 KC1, (Mathews et al., 1994). 2 CaC12,2 MgC12,10 D-glucose, 10 HEPES, with pH adjusted to 7.3 using NaOH. Tetrodotoxin (0.5 PM) was added to the extracellular solution to diminish spontaMETHODS neous synaptic currents in hippocampal cultures. In Cell cultures HEK-293 cells, cell capacitances were estimated using Hippocampalcells were prepared from 1 to 2 day old the compensationcircuitry of a List EPC-7 patch clamp post-natalalbino rats usingpreviouslydescribedmethods amplifier.HEK-293 cells selected for study typically had (Rodgers-Neame et al., 1992). Cells were plated at a capacitances in the range of 1540 pF. Steroids and benz[e]indenes were prepared in DMSO densityof W3 x 105cells/mlonto collagencoated tissue culture dishes in which a central well was prepared using as 10 mM stock solutions, and were diluted to final a dish imprinter (Mennerick et al., 1995). Cultures were concentrationat the time of experiment.The finalDMSO maintained in a humidified 95V002–590 C02 environ- concentrationused in these experimentswas <0.1!?70and ment at 37”C, and were used for study between 4 and 7 had no effect on GABA responses. GABA was applied days in vitro. for 0.5 sec via close-range pressure ejection (w20 psi) HEK-293 cells were obtainedfrom the American Type from glass pipettes with 1–2flm openings positioned CultureCollection(ATCC CRL 1573)and were prepared within 5 pm of the recorded cell at 20-30 sec intervals as described previously(Wong et al., 1995, 1996).Early until a stable baselinewas achieved.“Subsequently,drugs passage (2–10) cells in the exponentialphase of growth were co-applied for 0.5 sec with GABA until stable were plated onto 35 mm2tissueculturedishesat a density responses were achieved. This method diminished the of w 1 x 105cells/ml in Eagle’s MEM containing 109o duration that cells were exposed to steroids and fetal calf serum and 2 mM glutamine with 50 U/ml benz[e]indenes. In all cells, the effects of steroids and penicillin and 50 pg/ml streptomycin. HEK-293 cells benz[e]indeneswere at leastpartially reversible,although were maintainedunder the same incubationconditionsas some cells showed a small enhancement (50-100’%0of the hippocampalcells. control) that persisted. The concentrations reported are those in the drug de~iverypipette and are an upper limit Transection procedure for the concentrationsactually bathing the cell. HEK-293 cells were transiently transected with rat During an experiment cells were voltage clamped at GABA~receptor subunitcDNA cloned into vectors with –60 mV. Currents were filtered at 2 kHz and digitized COCH3

OF

COCH3

1163

Steroid enantiomers and GABA~ receptors

using pCLAMP (V 5.51). Data were analyzed using pCLAMP and routines written in Axobasic. Dose– response data were tit using a least squares minimization routine with an equation of the form: Response= Response~JIGABA]N/([GABA]N + ECU)), where Response~,X is the maximum response, [GABA] is the GABA concentration,EC50is the half-maximaleffective concentration and n is the Hill coefficient. Results are presented as mean t s.e.m. Differences in means were evaluated using two-tailed Students t-tests. Materials

Most chemicals were obtained from Sigma Chemical Co. (St Louis, MO). The GABA~ a6 and y2s subunit clones were provided by C. Fraser, j?2was a gift from P. Malherbeand A. Tobin providedthe al-subunit. (+)-BI-1 (Fig. 1) was synthesizedas previously described (Covey et al., 1993). The (–)-enantiomers of DHP and BI-1 were prepared from (lR-cis)-l-(l,l -dimethylethoxyl,2,3,6,7,7a-hexahydro-7a-methyl-5H-inden-5-one by total synthesis (Grieco et al., 1976; Rychnovsky and Mickus, 1992) and had NMR and IR spectra identical to those of (+)-DHP and (+)-BI-1. Full syntheticdetailswill be reported elsewhere. Enantiomeric purity was determined by optical rotation measurements of chloroform solutions of the compounds. (+)-DHP had [u]~4 = +101.5°, (–)-DHP had [u]~4 = –98.4°, (+)-BI-1 had [u]:’ = +96.2°, and (–)-BI-1 had [a]:’ = –92.4°. The (+)-enantiomers have the absolute configuration of naturally occurring steroids.

0.1

1

10 100 GABAConcentration(PM)

1000

RESULTS Fifty-sixper cent of isolatedHEK-293 cells transected with either al~2y2 or u6f12y2GABAA subunits expressedfunctionalreceptorsas evidencedby the presence of a macroscopicchloridecurrent in responseto 2–10PM GABA (n= 290). No systematic differences were observed in transection efficacy between cells expressing al f12y2or a6fi2y2 (ulf12y2:5670,n = 121; u6~2y2:55Y0, n = 169). Sham transected cells exposed to calcium phosphate without GABAA subunit cDNAs failed to respond to GABA at concentrations ranging from IOO#Mto 1 mM (n= 12). Consistent with prior studies (Ducic et al., 1995), HEK-293cells transected with a6f12y2had a lower EC50 for GABA than cells expressing ulf12y2 (Fig. 2). The EC50for alf12y2cells was 7.7 t 0.8PM (Hill coefficient = 1.2 ~ 0.1) while the EC50 for a6f12y2 cells was 1.9 ~ 0.3PM (Hill coefficient= 1.1 i- 0.2). HEK-293 cells exhibited cell-to-cellvariability in maximal GABA responses with peak currents of 883 t 228 pA (n = 19) and 1556 t 347 pA (n = 22) in response to 100VM GABA in cells transected with alf12y2 and a6~2y2, respectively. For comparison, the GABA ECSO in cultured hippocampal neurons is 9.8 + 0.1 KM (Hill coefficient= 1.4 f 0.1), with a peak response to 100PM GABA of 2122 + 158 pA (n = 30) (Table 1). Prior studies have also demonstrated that HEK-293 cells transected with alf12y2differ markedly from cells expressing E6/12y2in their response to benzodiazepine agonists (Liiddens et al., 1990). Although we did not verify the actual subunits expressed with antibody staining, we found that 10flM diazepam increased peak currentsevokedby 2 pM GABA to 199 f 349t0of control (n= 7) in cells expressing al~2y2 and had no effect on GABA currents in cells expressing cY6f12y2 (99 ~ 2% of control, n = 6). Because benzodiazepine agonists are significantly less effective on GABA responses at saturating GABA concentrations, we examined the effects of 10PM diazepam on responses to 0.5PM GABA in u6~2y2 transected cells and again found no effect (100 t 370 of control, n = 8). Diazepam also failed to enhance GABA responses in cells transected with al~2 (99 ~ 2Y0 of control, n = 3). In cultured hippocampal neurons, 10PM diazepam increases peak currentsevokedby 2PM GABA to 188 ~ 18Y0of control (n= 5).

Table 1. Steroids and GABA EC50

wl/32y2 u6/32y2 Fig. 2. Dose–response relationship for activation of chloride Control 1.9 f 0.3 7.7 ~ 0.8 currents by GABA in HEK-293 cells transected with alf12y2 (+)-DHP 0.77 + 0.07 0.34 f 0.06 0.95 * 0.1 7.3 k 0.6 2.6 ~ 0.5 10.1 * 0.1 (0) or a6f12y2(A). Responseswerenormalizedwithrespectto (–)-DHP the maximumresponse(100%).The solid lines represent the GABAEC50S aregiveninPM.Valuesarederivedfromthebestfitto best fit of the dose–responseequation with EC50 = 7.7 pM for the logisticequationgiven in Methods.For these experiments, ulf12y2and 1.9 MM for a6f12y2.The Hill coefficientswere 1.2 (+)-DHP and ( –)-DHP were administered in the bathing solution at 1PM. for alf12y2and 1.1 for a6/72y2. Hippocampalneurons 9.8 t 0.1

1164

C. F. Zorumski et al.

A-

A

HippocampalNeurons

v

500PA

!

+ (+)-DHP

+ (-)-DHP

HippocampalNeurons

l!3A

B-

B

H

GABA

GABA

+ (+)-DHP

L

&.ID2y2

GABA

L

200PA

200PA + (-)-DHP cs6&?y2

—

—

T

+ (+)-BI.I

200PA L

c-

a6D2y2

r —

GABA +(-) -BI-l

~PA

GABA

GABA CEPA

~ +(-)-DHP

IY’@’ 2s

r

+(+,.B,-,

@?A

I

200PA

+(-) -61-1 2s

Fig. 3. Traces of currents from hippocampal neurons (A) and HEK-293 cells transected with ul/?2y2 (B) and a6ji’2y2(C) voltage clamped at —60mV and exposed to 2 pM GABA in the absence and presence of the DHP enantiomers. The enantiomers were applied at 10pM.

Fig. 4. Response of hippocampal neurons (A) and HEK-293 cells transected with &l/?2y2(B) and a6/?2y2(C) to the BI-1 enantiomers. Cells were voltage clamped at –60 mV and exposed to 2 pkl GABA with and without 10,uMof the BI-1 enantiomers.

To examine the sensitivity to steroids and benz[e]indenes, we used 10~M of each compound,based on prior studies indicating that this concentration gives near maximal responses to (+)-DHP and (+)-BI-1 (RodgersNeame et al., 1992). At 10PM, the naturally occurring neurosteroid, (+)-DHP, increased currents evoked by 2 #M GABA in hippocampalneuronsto 1111 f 178%of control with a range of 470-1680% (n =6) (Fig. 3). Similarly, 10KM(+)-BI-1 increased GABA responsesin hippocampal neurons to 539 t 40% of control (n= 8) with a range of 409–783% (Fig. 4). In contrast, the (–)-enantiomers of these agents were significantlyless effective (Figs 3–5), supporting the hypothesis that neurosteroidsand benz[e]indenesact on GABAAreceptors at enantioselectivesites in hippocampalneurons. In HEK-293 cells transected with either the al/?2y2 or u6/?2y2subunitcombination,both DHP and BI-1 showed enantioselectivity similar to that seen in hippocampal neurons (Figs 3–5). The response of HEK-293 cells transected with al~2y2 subunits was similar to the response of the hippocampal neurons, showing greater than eight-fold enhancement by the (+)-steroid (mean:

829 t 99% of control, n = 16; range: 352–1600%) and more than four-fold enhancement by the (+)-benz[e]indene (mean: 433 ~ 58% of control, n = 12; range: 181– 767%) (Fig. 5). However, the enhancement of 2PM GABA currents in cells expressing m6~2y2was significantly less to both (+)-enantiomers and on average was about half that of either hippocampal neurons or HEK-293 cells with ctlf12y2 [(+)-DHP: 369 A 57%, n = 20; range: 120-994%; (+)-BI-1: 285 f 36%, n = 15; range: 121–531%;Fig. 5]. Possible explanations for the differences between rxlf12y2and u6f12y2receptors include differences in the steroid binding site on these receptors and/or differences in the potency of GABA as an agonist on the different receptors.Previousstudies have shown that (+)-DHP and (+)-BI-1have markedlygreater effects on peak responses at low concentrationsof GABA and have little effect on peak responses at saturating GABA concentrations (Rodgers-Neameet al., 1992).We investigatedthis issue by examining the effects of IO#M (+)-DHP and 10PM (+)-BI-1on currentsgatedby saturatingconcentrationsof GABA. In hippocampalneurons and HEK-293 cells, the

1165

Steroid enantiomers and GABA~ receptors 1200

A

1400 1200

a6p2y2

GZl

T

(+)-DH

1000

1000 z>

s $

Soo

2 z ~

c s

600

s

a 0 s

400 200

800

200

.J1

(+)-DHP

— EZZZl(+)-BI-I O (-)-BI-I

500 < 400 02 5 > 300 * $ 200

**

100 0 HCN

2PM GABA

400

0 600

0.51JMGABA

~

600

0

B

m

(+)-BI-1

Fig. 6. Effects of 10pM (+)-DHP and 10pM (+)-BI-1 on currents gated by 0.5 pM and 2 ~M GABA in HEK-293 cells transected with ct6/12y2.**p c 0.001; *p c 0.02 compared to effects on 0.5 KMGABA.

** * D_llL

alp2y2

ct6132y2

Fig. 5. Summaries of the effects of the DHP (A) and BI-1 (B) enantiomers on GABA currents. The steroids”and benz[e]~ indenes were administered at 10MM.Results are normalized with respect to the response to 2 pM GABA. **p c 0.001; *p <0.01.

to a similar degree, whereas 1 VM (–)-DHP had no effect on the GABA EC50(Table 1). In addition to potentiating GABA currents, neurosteroids directly gate bicuculline-sensitivechloride currents at certain concentrations(Barker et al., 1987; Callachan et al., 1987). (+)-DHP activates these currents at concentrations >1 ~M while (+)-BI-1 requires concentrations >10 wM. Because (+)-BI-1 is relatively inefficient in gating chloride currents (Rodgers-Neame et al., 1992),we focused on the effects of DHP to determine whether steroid enantioselectivity for direct channel gating and GABA potentiation are similar. In HEK-293 cells transected with cYl~2y2and a6~2y2, 10AM (+)DHP activated chloride currents of 56 ~ 21 pA (n = 18) and 209 i 79 pA (n = 7), respectively. In hippocampal neurons and transected cells, the currents gated by 10,uM(+)-DHP were always less than the currents gated by 2PM GABA (typically c 50% of the GABA current) (Rodgers-Neame et al., 1992). However, IO~M (–)DHP activatedcurrentsin both receptortypes that were at or below the limit of detection (alfi2y2: 3 t 2 pA (n= 12); ct6/32y2:11 ~ 7 pA (n= 6), p c 0.05 for each compared to (+)-DHP). Similarly, only the (+)-enantiomer of DHP consistently gated chloride currents in cultured hippocampal neurons at a concentration of IOpM ((+)-DHP: 147 ~ 37pA (n = 7); (–)-DHP: 5 ~ 2 pA (n = 6), P < 0.01).

steroid and benz[e]indenefailed to augmentresponsesto 100pM GABA [hippocampal neurons: (+)-DHP: 98 f 4% control (n = 6); (+)-BI-1: 106 f 6$Z0control (n= 6); @2y2: (+)-DHP: 102 ~ 5% (n= 5); (+)-BI-1: 100 ~ 5% (n = 6); a6/32y2:(+)-DHP: 102 ~ 6% (n = 6); (+)-BI-1: 90 ~ 5% (n = 5)]. These results suggest that differences in the GABA EC50 may account for differences between HEK-293 cells transected with al/32y2and a6/i’2y2subunits.When the GABA concentration was decreased to 0.5PM in cells transected with a6/32y2,to take into account the four-fold lower EC50 at these receptors, both (+)-DHP DISCUSSION and (+)-BI-1 produced a degree of enhancement similar to that observed in al/32y2 transected cells in the To date, 16 GABAA receptor subunits (al-6, fll-4, presence of 2PM GABA (Fig. 6). We also found that yl-3, d and pi-2) have been cloned (Macdonald and 1 pM (+)-DHP decreased the EC50 for GABA in Olsen, 1994; Sieghart, 1995). Five of these subunits are hippocampalneurons and the two sets of HEK-293 cells believed to combine to form functional GABAA

1166

C. F. Zorumski et al.

receptors. Although the stoichiometry of most native GABA~ receptors is uncertain,it is clear that differences in subunit compositioncontribute significantlyto differences in GABAA receptor pharmacology (Sieghart, 1995).While considerableinformationhas been obtained about the sites at which GABA, benzodiazepines, loreclezole and picrotoxin act, less is understood about the molecular determinants of steroid actions on these receptors. Studies using recombinant GABAAreceptors have not yet provided unequivocal evidence for subunit selectivityin the actionsof neurosteroids,althoughsome subunit combinationsappear to be more or less sensitive to the effects of these agents (Puia et al., 1993) and & subunits may be important in certain steroid actions (Lambert et al., 1995).It also remainsunclearwhether all of the effects of steroids on GABAA receptors can be accounted for by effects on the receptor protein or whether effects on membrane lipids contribute to the modulationof responses. In the absence of a specificbinding assay, one way to establish that neurosteroids act on GABAA receptor proteins is to demonstrate enantioselectivity of their actions. Previous structure–activitystudies of the putative steroid recognition site on GABAA receptors have used diastereomersto determine how changes in steroid structure affect gating and potentiation of GABAA responses (Harrison et al., 1987) and the evidence for a distinct steroid site has been mostly indirect (Turner et al., 1989). However, diastereomers have different physical properties which may affect not only their interactionswith receptors,but also the way they perturb the lipid environment of membrane bound GABAA receptors. Thus, diastereomers cannot be used to establishthat altered receptorfunctionresultsfrom either indirect effects due to membrane perturbation or direct effects resulting from binding to the protein receptor. This problem is overcome by the use of enantiomers which have identical physical properties and therefore perturb membranesin an identicalway. Thus, differential modulationof GABAAreceptor function by enantiomers can be attributed to differences in the direct binding interactionsof enantiomerswith the proteinreceptor.The present study extends prior work from our laboratories (Wittmer et al., 1995) and provides strong evidence for the existenceof enantioselectivesteroid sites on GABAA receptors in both native (hippocampal)and recombinant receptors expressed in HEK-293 cells. Furthermore, benz[e]indenes,which appear to augment GABA actions by effects at the steroid site, exhibit similar enantioselectivity. Our previous studies in hippocampal neurons have demonstrated that neurosteroidsand benz[e]indenes are effective modulators of GABA responses (RodgersNeame et al., 1992; Covey et al., 1993; Hu et al., 1993). However, the effects of the steroids show considerable cell-to-cell and region-to-regionvariability (Gee and Lan, 1991).This variability maybe due to the subunit heterogeneity of GABAA receptors, differences

in the lipid environment of the receptor and phosphorylation status of the receptors (Lambert et al., 1995). Differences in GABA’s affinity and efficacy at various GABAA receptors may account for some or all of this variability because steroids potentate responses to low concentrations of GABA but have no effect on peak responsesto saturating agonist concentrations. To begin addressing questions concerning the heterogeneity of steroid effects, we examined the actions of DHP and BI-1 enantiomersin HEK-293 cells transected with al~272 and u6~2y2GABAAreceptor subunits.The cYlf12y2combination appears to be a major GABAA receptor expressed in brain (Fritschy et al., 1992), and is present in hippocampal interneurons (Gao and Fritschy, 1994). The u6/32y2receptor appears to be restricted to cerebella granule cells (Liiddens et al., 1990), but is of particular interest because it has unique pharmacological properties that influence the actions of benzodiazepine agonists (Liiddens et al., 1990) and y-butyrolactones (Mathews et al., 1994).Our studies show that the effects of DHP and BI-1 are not significantlydependent on the variants present in these GABAA receptor subunit combinations, provided that differences in the GABA EC50are taken into account. Hippocampal neurons have heterogeneous GABAA receptors,based both on physiologicalmeasures (Schonrock and Bormann, 1993) and in situhybridization studies(Wisden et al., 1992).Therefore, the variabilityin the effects of steroids may be due to the heterogeneous subunit composition of hippocampal GABAA receptors and the resulting differences in GABA EC50S.However, we observedvariability in responsesto (+)-DHP and (+)BI-1 in the transected HEK-293 cells that is similar to the variability observed in hippocampal neurons, suggesting that cell-to-cell differences in subunit expression and/or assembly, even in cells with a limited number of subunittypes, may contributeto differencesin responses. Both the potentiation of GABA responses and the direct gating of chloride channels by neurosteroids are enantioselective. However, it is presently unclear whether both of these effects are mediated by the same site on GABAAreceptors or whether there is more than one steroid recognition site. Studies examining the effects of steroids on chloride flux (Morrow et al., 1990)and TBPS binding, a marker for the picrotoxinsite (Hawkinson et al., 1994),have suggestedthe presence of high and low affinitysteroid sites. If the direct gating and GABA potentiating effects of steroids are mediated by actions at two distinct sites, our data suggest that both sites exhibit enantioselectivity. Behavioraland clinicaleffects of neurosteroidsinclude anxiolytic, anesthetic and anticonvulsant actions (Lambert et al., 1995). The steroid and benz[e]indene enantiomershave allowed us to begin examiningwhether the behavioral effects of these agents also show enantioselectivity. In studies using tadpole and mouse assays, we found that the anesthetic effects of DHP and BI-1 are similarto the GABA~ receptor effects in that the

Steroid enantiomers and GABA~ receptors

are significantlyless effective than the (+)-enantiomer~ in some cases by at least an order of magnitude (Wittmeret al., 1995). Such findings help to strengthen the relationship between GABAergic and clinical drug effects. (–)-enantiomers

Acknowledgements—This work was supportedby NIH Grants GM47969 and NS14834, Research Scientist Development Award MHO0964,and a fellowship from the Bantly Foundation. The authors thank Ann Benz, Jianxin Que, Xia Zhou and Kathleen Reinagel for technical assistance, and Jim Krause, Steve Mennerick, Amy Zarrin, Alex Evers and Joe Henry Steinbach for helpful discussions.

REFERENCES Barker J. L., Harrison N. L., Lange G. D. and Owen D. G. (1987) Potentiation of y-aminobutyricacid-activated chloride conductanceby a steroid anesthetic in culturedrat spinal neurones.J. Physiol. (Lend.) 386: 485–501. Callachan H., Cottrell G. A., Hather N. Y., Lambert J. J., Nooney J. M. and Peters J. A. (1987) Modulation of the GABA~ receptor by progesteronemetabolizes.Proc. R. ,SOC. Lend. 231: 359–369. Chen C. and Okayama H. (1987) High efficiency transformation of mammalian cells by plasmid DNA.A401ec.Cell Biol. 7: 2745-2752. Covey D. F., Hu Y., Bouley M. G., Holland K. D., RodgersNeame N. T., Isenberg K. E. and Zorumski C. F. (1993) Modulationof GABA~ receptor function by benz[e]indenes and phenanthrenes.J. med. Chem. 36: 627–630. Ducic I., Caruncho H. J., Zhu W. J., Vicini S. and Costa E. (1995) y-Aminobutyric acid gating of Cl- channels in recombinantGABA~receptors.J. Pharmac. exp. Ther. 272: 438&445. Fritschy J. M., Benke D., Mertens S., Oertel W. H., Bachi T. and Mohler H. (1992) Five subtypes of type A yaminobutyric acid receptors identified in neuronsby double and triple immunofluorescencestainingwith subunit-specific antibodies.Proc. natn. Acad. Sci. U.S.A. 89: 672&6730. Gao B. and Fritschy J. M. (1994) Selective allocation of GABA~ receptors containing the U1subunit to neurochemically distinct populationsof hippocampalinterneurons.Eur. J. Neurosci. 6: 837-853. Gee K. W. and Lan N. C. (1991)y-ArninobutyricacidAreceptor complexes in rat frontal cortex and spinal cord show differential responses to steroid modulation. Molec. Pharmac. 40: 995–999. Grieco P. A., Fukamiya N. and Miyashita M. (1976) An asymmetric approach to the synthesis of 12-methylprostaglandins.J. Chem. Soc. Chem. Commun.573–575. Harrison N. L., Majewska M. D., Barrington J. W. and Barker J. L. (1987) Structure–activity relationships for steroid interaction with the y-aminobutyric acid~ receptor complex.J. Pharmac. exp. Ther. 241: 346-353. HawkinsonJ. E., KimbroughC. L., McCauleyL. D., Bolger M. B., Lan N. C. and Gee K. W. (1994)The neuroactive steroid 3c+hydroxy-5&pregnan-20 -one is a two-component modulator of ligand binding to the GABA~ receptor. Eur. J. Pharmac. 269: 157–163. Hu Y., Zorumski C. F. and Covey D. F. (1993) Neurosteroid analogues: Structure-activity studies of benz[e]indenemod-

1167

ulators of GABA~ receptor function. 1. The effect of 6methyl-substitutionon the electrophysiologicalactivity of7substituted-benz[e]indene-3-carbonitriles. J.med. Chem. 36: 3956-3967. LambertJ. J., Belelli D., Hill-Venning C. and Peters J. A. (1995)Neurosteroidsand GABA~ receptor function. Trends Pharmac. Sci. 16: 295–303. LiiddensH., Pritchett D. B., KohlerM., Kilisch I., KeinanenK., MonyerH., Sprengel R. and SeeburgP. H. (1990)Cerebella GABAA receptor selective for a behavioral alcohol antagonist.Nature 346: 648-651. Macdonald R. L. and Olsen R. W. (1994) GABAA receptor channels.A. Rev. Neurosci. 17: 569–602. Majewska M. D., Harrison N. L., Schwartz R. D., Barker J. L. and Paul S. M. (1986) Steroid hormone metabolizes are barbiturate-like modulators of the GABA receptor. Science 232: 1004-1007. MathewsG. C., Bolos-SyA. M., Holland K. D., IsenbergK. E., Covey D. F., Ferrendelli J. A. and Rothman S. M. (1994) Developmental alteration in GABA~ receptor structure and physiological properties in cultured cerebella granule neurons.Neuron 13: 149–158. Mennerick S., Que J., Benz A. and Zorumski C. F. (1995) Passive and synaptic properties of hippocampal neurons grown in microcultures and in mass cultures. J. Neurophysiol. 73: 320–332. Morrow A. L., Pace J. R., Purdy R. H. and Paul S. M. (1990) Characterizationof steroid interactions with y-arninobutyric acid receptor-gated chloride ion channels: evidence for multiple steroid recognitionsites. Molec. Pharmac.37: 263– 270. Puia G., Ducic I., Vicini S. and Costa E. (1993) Does neurosteroid modulatory efficacy depend on GABAA receptor subunit composition? Receptors and Channels 1: 135–142. Rodgers-NeameN. T., Covey D. F., Hu Y., Isenberg K. E. and ZorumskiC. F. (1992)Effects of a benz[e]indeneon GABAgated chloride currents in cultured post-natal rat hippocampal neurons.Molec. Pharmac. 42: 952–957. RychnovskyS. D. and Mickus D. E. (1992) Synthesis of entcholesterol, the unnatural enantiomer. J. org. Chem. 57: 2732-2736. SchonrockB. and BormannJ. (1993) Functional heterogeneity of hippocampal GABA~ receptors. Eur. J. Neurosci. 5: 1042–1049. Sieghart W. (1995) Structure and pharmacology of yaminobutyric acidA receptor subtypes. Pharmac. Rev. 4’7: 181–234. Turner D. M., Ransom R. W., Yang J. S. -J. and Olsen R. W. (1989) Steroid anesthetics and naturally occurring analogs modulate the y-aminobutyricacid receptor complex at a site distinct from barbiturates.J. Pharmac. exp. Ther. 248: 960966. Wisden W., Laurie D. J., Monyer H. and Seeburg P. H. (1992) The distribution of 13 GABA~ receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon, mesencephalon.J. Neurosci. 12: 1040-1062. Wittmer L. L., Zorumski C. F., Kalkbrenner M., Evers A. S., Hu Y. and Covey D. F. (1995) Modulation of GABAA receptors by neurosteroids, benz[e]indenes, and their enantiomers.Soc. Neurosci. Abstr. 21: 1344. Wong E. T., Holstad S. G., Mennerick S. J., Hong S. E., Zorumski C. F. and Isenberg K. E. (1995) Pharmacological

1168

C. F. Zorumski et al.

and physiological properties of a putative ganglionic nicotinic receptor, u3/14,expressedin transected eucaryotic cells. MoZec.Brain Res. 28: 101–109. Wong E. T., Holstad S. G., Mennerick, S. J., Hong S. E.,

Zorumski C. F. and Isenberg K. E. (1996) Transient transection and messenger RNA analysis of a putative neuronal nicotinic acetylcholine receptor expressed in eucaryotic cells. Neurosci. Protocols in press.