Neuropharmacology 44 (2003) 4955

www.elsevier.com/locate/neuropharm

On the putative physiological role of allopregnanolone on GABAA

receptor function
G. Puia a,1, J.-M. Mienville a,2, K. Matsumoto a,3, H. Takahata b, H. Watanabe c, E. Costa a,
A. Guidotti a,
a

The Psychiatric Institute, Department of Psychiatry, University of Illinois at Chicago, 1601 W. Taylor St. MC912, Chicago, IL 60612, USA
b
Tohoku Pharmaceutical University, Sendai, Japan
c
Department of Pharmacology, Institute of Natural Medicine, Toyama Medical and Pharmaceutical University, Toyama, Japan
Received 27 June 2002; accepted 26 September 2002

Abstract
To obtain definitive evidence for a physiological allosteric modulatory role for endogenous brain ALLO on GABAA receptor
function, we studied GABAA receptor activity under conditions in which the concentration of endogenous brain ALLO was decreased
by about 80% for longer than 5 h following the administration of SKF 105111- 17-17-[bis (1methylethyl) amino carbonyl] androstane-3,5-diene-3-carboxylic acid (SKF), a potent inhibitor of 5-reductases Type I and II. We used the in situ patchclamp technique
to record GABA-evoked currents and spontaneous inhibitory postsynaptic currents (sIPSCs) from pyramidal neurons in neocortical
slices of vehicle- or SKF-treated mice. The potency, but not the efficacy, of exogenously applied GABA was decreased in slices
from mice treated with SKF. When neocortical slices were treated in vitro for 3 h with 10M SKF, ALLO was also reduced (25
30%) and in addition, the GABA doseresponse curve was shifted to the right; however this shift was not as marked as the shift
in the slices obtained from mice treated with SKF, in keeping with the smaller decrease of the ALLO content in these slices.
Furthermore, direct application of ALLO to these slices shifted the doseresponse curve of GABA back toward a non-SKF treated
profile. We then analyzed GABAergic sIPSCs in neocortical slices obtained from vehicle or SKF-treated mice. Mean decay time
and charge transfer were significantly reduced by SKF treatment. The decay of sIPSCs was best fitted by two exponentials, but
only the fast component was decreased in the SKF group. Direct application of ALLO (100nM) normalizes the sIPSC kinetics in
slices from ALLO depleted mice. No changes were detected in the amplitude or frequency of sIPSCs. These data demonstrate that
endogenous ALLO physiologically regulates spontaneously induced Cl current by acting on a specific recognition site, which is
probably located on GABAA receptors (a receptor on a receptor), thereby prolonging inhibitory currents by facilitating conformational transition of the GABA-gated Cl channel to an open state.
2003 Elsevier Science Ltd. All rights reserved.
Keywords: Neurosteroid; GABAA receptor; SKF 105111; 5-reductase; In situ patchclamp; Brain slice

1. Introduction
Allopregnanolone (ALLO; 3,5-tetrahydroprogesterone)
is synthesized in the brain from progesterone via the
Corresponding author. Tel.: +1-312-413-4592; fax: +1-312-4134569/4544.
E-mail address: aguidotti@psych.uic.edu (A. Guidotti).
1
Present address: UMR 6548, University de Nice-Sophia AntipolisNice, France
2
Present address: UMR 6548, Universite de Nice-Sophia Antipolis,
Nice, France
3
Present address: Department of Pharmacology, Institute of Natural
Medicine, Toyama Medical and Pharmaceutical University, Toyama, Japan

action of two NADPH/NADP+-reducing enzymes, a 5

reductase type I that transforms progesterone into
5-dihydroprogesterone (5-DHP), and a 3-hydroxysteroid oxidoreductase that reduces 5-DHP into ALLO
(Karavolas and Hodges, 1991; Costa et al., 1994; Dong
et al., 2001). Although the exact cellular location of
these enzymes in the CNS has not been defined, the distribution of these enzymes, and the level and turnover
rate of 5-DHP and ALLO in the brain are not uniform,
suggesting a role for neurosteroids in specific CNS functions (Dong et al., 2001). While 5-DHP interacts with
progesterone receptors with high affinity and therefore
has a genomic action (Rupprecht and Holsboer, 1999),

0028-3908/03/$ - see front matter 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0028-3908(02)00341-6

50

G. Puia et al. / Neuropharmacology 44 (2003) 4955

ALLO can regulate GABAA receptor function mainly by

a non-genomic mechanism; i.e., though allosteric positive modulation of GABA-gated Cl current (Harrison
et al., 1987; Mienville and Vicini, 1989; Puia et al.,
1990, 1993; Lambert et al., 1995; Zhu and Vicini, 1997;
Brussaard et al., 1999).
The definition of the physiological role of endogenous
ALLO has been hampered by the lack of specific highaffinity antagonists for the steroid recognition site
expressed by GABAA receptors. To overcome this limitation, we used a potent selective inhibitor of type I 5
reductase, the rate-limiting step enzyme for ALLO
biosynthesis in the brain (Dong et al., 2001). Recently,
we reported that after systemic administration of 1717-[bis (1methylethyl) amino carbonyl] androstane-3,5diene-3-carboxylic acid (SKF), there is a large and rapid
decrease in ALLO brain content associated with a
decrease in the duration of the righting-reflex loss elicited by muscimol or other GABA mimetic drugs (i.e.,
pentobarbital and diazepam) (Matsumoto et al., 1999;
Pinna et al., 2000; Guidotti et al., 2001). These changes
were paralleled by a decrease in muscimol-elicited Cl
current in neocortical brain slices prepared from SKFtreated mice (Pinna et al., 2000).
In this report, we further investigated the effects of
the SKF-induced decrease in ALLO brain content on
GABAA receptor function ex vivo by determining the
responsiveness to GABA of GABAA receptors expressed
in the somata of neocortical pyramidal neurons, as well
as the changes in intensity of spontaneously-induced
GABAergic inhibitory postsynaptic Cl current (sIPSC).

2. Materials and methods

2.1. Brain slice preparation
SKF-Na (47 mol/kg; dissolved in 1% DMSO) or
vehicle were administered as a single injection i.p. to
15/18-day-old male Swiss albino mice in a constant volume of 100 l/10 g body weight. Mice were decapitated
1, 3, 6, or 24 h after injection and fronto-parietal neocortical slices were prepared with a vibratome. The slices
used for the electrophysiological experiments were 200
m thick and those used for the ALLO measurements
were 300 m thick. As an alternative treatment route,
we also incubated slices from untreated animals at 22
24 C for 3 h in the same extracellular medium used
for electrophysiological studies with or without 10 M
SKF. This medium composition in mM was: 120 NaCl,
5 KCl, 1.25 NaH2PO4, 2 MgCl2, 2 CaCl2, 26 NaHCO3,
1 Na pyruvate and 10 dextrose. It was supplemented
with MEM vitamins and amino acids and continuously
gassed with 95% O2/5% CO2. Before any experiment,
these slices were washed thee times in extracellular
medium without SKF.

2.2. Quantitation of ALLO content in brain slices

To measure ALLO (Uzunov et al., 1996), brain tissue
was homogenized in 10 volumes of distilled water containing 2 fmol/ml of [3H]ALLO to monitor recovery. The
supernatant was extracted with ethylacetate and after
lyphilization was purified with HPLC. The HPLC-eluted
fraction containing ALLO was derivatized with heptafluorobutyric acid anhydride (HFBA) after addition of
the deuteriated internal standard (Dong et al., 2001).
ALLO was analyzed at the mass-spectrometer with multiple ion detection in the negative ion chemical ionization mode (NICI) using methane as the reaction gas
(Uzunov et al., 1996). The standard curve for ALLO was
prepared by combining different known quantities of
authentic ALLO with a constant amount of alfaxalone
(3 fmol) as the internal standard. The detection limit for
ALLO was approximately 10 fmol; the standard curve
was linear between 1 and 103 fmol (Matsumoto et al.,
1999; Pinna et al., 2000).
2.3. Electrophysiology
Electrophysiological recordings were performed
between 1 and 3 h following SKF injection. Slices were
placed in a recording chamber mounted on the stage of
an upright microscope (Nikon Eclipse), visualized by IRDIC microscopy though a 40 water-immersion objective, and perfused with extracellular medium at 0.5
ml/min. Pyramidal neurons were identified by their morphology and characteristic regular spiking response to
depolarizing pulses (Pinna et al., 2000). GABA-evoked
currents and sIPSCs were recorded with a patchclamp
amplifier (EPC7, List Electronic) from pyramidal neurons in cortical layer II/III using the tight-seal voltage
clamp technique in the whole-cell configuration
(Matsumoto et al., 1999; Pinna et al., 2000). Electrodes
(24 M) were filled with an intracellular medium
(solutes in mM) 125 KCl, 4 MgCl2, 10 HEPES, and 5
NaATP. Current traces elicited by exogenously applied
GABA were analyzed using pClamp8 software (Axon
Instruments). When recording sIPSCs, 10 M CNQX
and 5 M MK801 were added to the extracellular solution to block excitation of AMPA, kinate and NMDA
receptors. Spontaneous activity was stored on video tape
with a VR100/VCR system (Instrutech). Selected segments containing sIPSCs were then digitized with
pClamp8, and given a coded file name for blind analysis
with Mini Analysis software (Synaptosoft Inc.,
Leonia, NJ).
2.4. Drugs
SKF-Na was synthesized using the method described
by Holt et al., 1990. The purity of the crystallized compound was checked by determining its melting point:

G. Puia et al. / Neuropharmacology 44 (2003) 4955

230245C, which is in the range of that reported by

Holt et al. (1990). Moreover, NMR 1H chemical shifts
correspond to the chemical shifts reported by Holt et al.
(1990) (i.e., C4H=7.16 vs 7.15; C6H=5.99 vs 5.88;
C18H3=0.94 vs 0.92; C19H3=0.82 vs 0.81Calculated
values of the test compound vs the data reported by Holt
et al., 1990). GABA (Sigma) was applied though a
multi-channel delivery system converging onto a 100m-diameter (i.d.) micromanifold with its tip placed
~200 m from the recorded cell. This setup allowed fast
application (t~40 ms) and removal of GABA
(Matsumoto et al., 1999; Pinna et al., 2000). Bicuculline
methiodide (Sigma), allopregnanolone (Steraloids, Wilton, NH), MK801 and NBQX (Research Biochemicals
Inc., Natick, MA) were applied via the bath superfusion.

3. Results
3.1. Effects of in vivo treatment with SKF
We first measured ALLO content in neocortical slices
prepared 1, 3, 6, and 24 h after i.p. administration of 47
mol/kg of SKF to mice. Fig. 1 shows that the decrease
of neocortical ALLO was already large (~70%) at 1 h
after SKF treatment and reached a maximum of approximately 80% between 1 and 5 h post SKF injection.
To test whether the decrease in endogenous brain
ALLO alters Cl currents elicited by GABA at GABAA
receptors, we recorded Cl current intensities from layer
II/III pyramidal neurons in cortical slices from mice
treated 23 h earlier with either vehicle or SKF at a time
when the SKF effect on ALLO levels had peaked. We
measured GABA-gated Cl current intensities elicited in

51

these neurons by GABA applications (0.3300 M; Fig.

2). The responses elicited by 0.330 M GABA were
greatly reduced (by about 70%) in brain slices obtained
from mice receiving in vivo SKF treatment, whereas
those elicited by higher (near saturation) GABA concentrations (100300 M) were not significantly different
from the control (Fig. 2A).
3.2. Effects of in vitro treatment with SKF
Neocortical slices preincubated with 10 M SKF for
3 h also showed a significant reduction (ranging from
20 to 35%) of ALLO content compared with vehicletreated slices (vehicle: 5.4 0.1 pmol/g; SKF:
4.0 0.04 pmol/g, p 0.05, n 6). The amplitude of
the Cl current elicited by low concentrations of GABA
(0.31 M) was also significantly reduced in SKFtreated slices compared with vehicle (Fig. 2B). However,
the shift to the right in doseresponse relationships was
not as marked as for SKF treatment in vivo, as responses
elicited by 3 (Fig. 2B) and 10 M GABA (not shown)
were not significantly different (p 0.1) in SKF- vs
vehicle-treated slices. A potential direct effect of SKF
on GABAA receptors should be ruled out because in previous experiments (Matsumoto et al., 1999; Pinna et al.,
2000), we reported that the Cl current intensities elicited by coapplication of SKF (non-preincubation) with
GABA or muscimol were not different from those
obtained with the GABA receptor agonist alone. However, to rule out the possibility that the effect of SKF
on GABAA receptors is due to a mechanism other than
the decrease of ALLO, we applied ALLO to slices in
which the response to GABA was shifted to the right
following pretreatment with SKF.
Fig. 2B shows that bath application of 10 M ALLO
to SKF treated slices shifted back toward the non-SKF
treated slice the GABA doseresponse curve, suggesting
that the action of SKF on GABAA receptor response is
not direct but is mediated via a decrease of ALLO
tissue content.
3.3. Spontaneous IPSCs from SKF-treated mice have
faster kinetics compared to control mice

Fig. 1. Time-dependent changes in ALLO content of mouse neocortex slices (300 m thick) following a single intraperitoneal injection
of SKF (47 mol/kg) at time 0. Mice were killed 1, 3, 6, and 24 h after
SKF administration and ALLO content was measured as described in
Section 2.2. Each data point is the mean SEM of six animals.
p 0.001 compared to the value at t 0 (ANOVA).

If a decrease in endogenous ALLO has this important

an effect on GABA-gated channel function, it should
also modify the action of GABA released from presynaptic terminals. To test this hypothesis, sIPSCs were
recorded in slices from SKF-treated vs control animals.
Bath application of 10 M bicuculline methiodide, a
selective GABAA receptor antagonist, totally abolished
sIPSCs (n 4; not shown). Representative sIPSCs
recorded in the two experimental conditions are illustrated in Fig. 3 (left panels). Recordings including
between 100 and 300 events were analyzed with respect
to kinetics, amplitude, charge transfer (area), and fre-

52

G. Puia et al. / Neuropharmacology 44 (2003) 4955

Fig. 2. Decrease in endogenous ALLO reduces GABAA receptor sensitivity to exogenously applied GABA. (A) Responses to GABA in frontoparietal cortical slices prepared 23 h following in vivo SKF treatment. For each slice five to six neurons were recorded. Each data point is the
mean SEM of six mice. p 0.05 and p 0.01 compared to corresponding values obtained in slices from vehicle-treated mice. The insert
in A is an enlargement of the responses to 0.3 and 1 M GABA. Open Bars=vehicle-treated mice; Closed Bars=SKF- (47mol/kg ip) treated mice.
(B) Responses to GABA in slices treated with SKF in vitro (35% depletion of ALLO in this experiment) in the presence or absence of exogenously
applied ALLO (10 M). Cortical slices (200 m thick) from the same brain were incubated for 3 h in extracellular medium (see Section 2.3) with
or without 10 M SKF. In these slices, GABA doseresponse testing was conducted before and 10 min after bath application of ALLO. Each data
point is the mean SE of the current recorded from five to six neurons per slice. p 0.01 when values of SKF treated slices were compared
with the corresponding values obtained in control slices or in SKF-treated slices after the application of ALLO. Results of a typical experiment
repeated with similar results in thee other brains. Similar results were also obtained studying the doseresponse curve of muscimol (for details see
Pinna et al., 2000).

Fig. 3. Comparison of GABAergic sIPSCs recorded in pyramidal neurons from vehicle (A) and SKF (47 mol/kg i.p.) (B) treated mice. Left
panels: superimposed sIPSCs representative of each condition. Right panels: corresponding scattered histograms of sIPSC decay time fitted with
a lognormal function (Brussaard et al., 1999). Decay time was calculated by Mini Analysis software as the time for sIPSC to decay to 33% of
peak amplitude. Insets: averaged sIPSCs fitted with a biexponential function. Calibration: 50 ms, 100 pA (A), 200 pA (B). Data displayed were
selected from recordings made in slices from six vehicle-treated and six SKF-treated animals (see also Table 1).

quency. Mean decay time of sIPSCs recorded in slices

from SKF-treated mice was significantly shorter than
control (Fig. 3, right panels; Table 1). The decay phase
of averaged sIPSCs from both vehicle- and SKF-treated
mice was best fitted by a biexponential function with a
fast (tfast) and a slow (tslow) component (Fig. 3, insets).

Table 1 summarizes the results obtained with sIPSCs,

which indicate that the shortening of mean decay time
observed in the SKF group appears to be due mainly to a
selective reduction of tfast. The area under sIPSC, which
represents the charge transferred though GABAA receptor channels, was also reduced in slices obtained from

G. Puia et al. / Neuropharmacology 44 (2003) 4955

53

Table 1
Comparison of sIPSC kinetic parameters and charge transfer (area) between SKF- (47mol/kg ip) and vehicle-treated micea
In vivo treatment

tfast (ms)

tslow (ms)

% slow

Mean decay time (ms) Charge transfer (pA. ms)

Vehicle (29)b
SKF 105111 (25)b

13.30.8
9.10.6c

39.93
35.13

352
372

14.11
11.10.6c

a
b
c

4517524
2946314c

Each value is the meanSEM

The number of neurons studied is indicated in parentheses
p0.05, p0.01

SKF-treated mice. We did not detect significant differences between groups for sIPSC amplitude (vehicletreated: 225 29 pA, n 30; SKF-treated: 216 30
pA, n 22) or frequency (vehicle-treated: 7.5 1.2 Hz;
SKF-treated: 7.7 1.0 Hz).
Bath application of 100 nM ALLO normalizes sIPSC
kinetics in slices from ALLO-depleted mice. Values of
tfast before and after ALLO were: control, 10.9 0.5
and 13.7 1.2 ms (p 0.05, n 9); SKF: 9.2 0.4
and 11.4 0.7 ms (p 0.01, n 6; paired t-tests).
Values for charge transfer before and after ALLO were:
control, 837 74 and 1007 145 pAms (p 0.07);
SKF, 628 43 and 848 89 pAms (p 0.05).
4. Discussion
Here we demonstrate that in neocortical pyramidal
neurons, a SKF-induced decrease in ALLO content
results in a reduction of GABAA gated Cl current intensity, shifting the doseresponse curve for exogenously
applied GABA to the right and attenuating the charge
transfer of GABA-mediated sIPSCs, mainly though a
decrease in their fast component of decay. The fact that:
(a) SKF treatment was more effective in decreasing
ALLO levels and in reducing the amplitude of GABA
gated Cl current intensity following in vivo rather than
in vitro treatment (shown in Figs. 2 and 3); and (b) reversal of SKF effects were obtained coadministering ALLO
with GABA (these results) or muscimol (Pinna et al.,
2000), supports the inference that endogenous ALLO
exerts a physiological modulatory action at GABAA
receptors and thus may regulate inhibitory tone in the
neocortex by optimizing the effect of GABA at
GABAA receptors.
It is noteworthy that following the SKF-induced
decrease of neocortical ALLO content, the potency but
not the efficacy of exogenously applied GABA was
reduced (Fig. 2). A similar shift in potency, without
changes in efficacy, was observed when muscimol,
which is a more stable (not taken up by reuptake mechanisms, not rapidly metabolized) and more specific (has
no action at GABAB receptors) GABA receptor agonist
than GABA itself, was used (Pinna et al., 2000). This
suggests that the effect of SKF treatment on the action

of GABA at GABAA receptors does not depend on the

alteration of GABA reuptake mechanisms or metabolism. The fact that SKF (47mol/kg) treatment does not
affect maximum GABA-gated Cl conductance and that
exogenously applied ALLO shifted the doseresponse of
GABA back to a non-SKF treated profile further suggests that this drug fails to produce a significant change
in GABAA receptor number or structure but rather, by
rapidly decreasing endogenous brain ALLO content,
allosterically modulates the action of GABA at
GABAA receptors.
Binding studies have demonstrated that the affinity of
3
H-muscimol and 3H-flunitrazepam to GABA and
benzodizepine recognition sites, respectively, is
enhanced by neurosteroids at concentrations likely to
occur physiologically (Belelli et al., 1990; Lan et al.,
1990; Majewska, 1990; Nguyen et al., 1995). Thus, the
decrease in the response to smaller doses (0.330 M)
of GABA in brain slices with a reduction in the amount
of ALLO elicited by SKF treatment in vivo or in vitro
could be due to a decrease of the positive allosteric
action of ALLO at the GABAA receptor with a consequent change in the affinity of GABA for its recognition site.
These considerations, however, are not relevant to the
effect of neurosteroids on synaptic inhibitory signals,
because when brief pulses of GABA are released synaptically from presynaptic nerve terminals to postsynaptic
GABAA receptors, these receptors became largely saturated by GABA (Edwards et al., 1990), and in this case
the effects of the decrease in ALLO on synaptic current
are characterized by a reduction of the fast decay component and charge transfer of synaptic GABAA Cl currents. It has been reported that, in addition to an allosteric action similar to that of benzodiazepines, ALLO,
like pentobarbital, may potentiate GABA action at
GABAA receptors, directly prolonging the channel opening duration (Lambert et al., 1995), hence modulating
the action of saturating concentrations of GABA at synapses.
In a very few previous reports, the effect of neurosteroids applied in physiologically relevant concentrations
on synaptic GABA-mediated transitions in brain slices
was investigated. In one pharmacological study, however, it was demonstrated that the direct application of

54

G. Puia et al. / Neuropharmacology 44 (2003) 4955

relatively small (50100 nM) concentrations of neuroactive steroids to different brain areas can modulate the
current elicited by endogenously released GABA at
inhibitory synapses by slowing the decay phase of
sIPSCs, in this way prolonging the synaptic current
(Cooper et al., 1999). This effect of neurosteroids was
attributed to a slowing of the recovery of the receptor
from the desensitized state and hence to a prolongation
of inhibitory current by this mechanism (Zhu and Vicini,
1997). Here, we found that lowering the ALLO content
with SKF results in a speed-up of channel closing and
in a reduction of charge transfer in GABA-mediated
sIPSCs. Interestingly, the effects of ALLO depletion by
SKF on sIPSC decay were selective for tfast. According
to Jones and Westbrook (1995), the initial decay phase
of GABAergic IPSC is due to fast desensitization, while
the slower phase reflects the reopening of channels from
desensitized states. An alternative model proposes that
each decay phase could correspond to a succession from
biliganded to monoliganded states of the receptor channel (Hill et al., 1998). Thus, endogenous ALLO might
selectively affect fast desensitization or preferentially
bind to a biliganded conformation of the receptor, or
both.
Taken together, the data suggest that endogenous concentrations of ALLO may have an important physiological significance in the modulation of GABAA receptormediated transmission in vivo, not necessarily by changing the affinity of GABA for its receptor, but most likely
by accelerating the transition of the channel from an
open to a closed conformational state, thereby regulating
inhibitory currents by this mechanism.
In previous studies, we demonstrated a decrease in the
behavioral response elicited by GABA receptor agonists
in rats where ALLO brain levels were reduced either by
social isolation or by SKF administration (Matsumoto et
al., 1999; Pinna et al., 2000). Our present results on: (a)
the fast rate of ALLO brain biosynthesis revealed by the
SKF experiments, and (b) the decrease in response to
exogenous GABA and sIPSC charge transfer observed
in neurons from ALLO-depleted mice, are compatible
with the proposed role of ALLO as a physiological positive modulator of GABA-Cl channel gating and represent a key result, suggesting that inhibitory tone is
impaired in animals with decreased brain levels of neurosteroids.
In conclusion, we have demonstrated that GABAA
receptor function could be dynamically regulated by
endogenous levels of ALLO. Therefore, physiological or
drug-induced variations in the synthesis and release of
this neurosteroid may have profound effects on CNS
inhibition. Indeed, several observations suggest that
changes in brain ALLO concentration may contribute to
specific physiological or pathological situations
(Guidotti and Costa, 1998; Guidotti et al., 2001). A deep

understanding of the underlying mechanism of action of

ALLO might be decisive for new therapeutic approaches.

Acknowledgements
This work was supported by RO1MH49486 and
RO1MH56890 to AG and RO1MH56500 to EC.

References
Belelli, D., Lan, N.C., Gee, K.W., 1990. Anticonvulsant steroids and
the GABA/benzodiazepine receptor-chloride ionophore complex.
Neuroscience Biobehavioral Review 14, 315322.
Brussaard, A.B., Devay, P., Leyting-Vermeulen, J.L., Kits, K.S., 1999.
Changes in properties and neurosteroid regulation of GABAergic
synapses in the supraoptic nucleus during the mammalian female
reproductive cycle. Journal of Physiology 516, 513524.
Cooper, E.J., Johnston, G.A.R., Edwards, F.A., 1999. Effects of naturally occurring neurosteroids on GABAA IPSCs during development in rat hippocampal or cerebellar slices. Journal of Physiology
521, 437449.
Costa, E., Auta, J., Guidotti, A., Korneyev, A., Romeo, E., 1994. The
pharmacology of neurosteroidogenesis. Journal of Steroid Biochemical Molecular Biology 49, 385389.
Dong, E., Matsumoto, K., Uzunova, V., Sugaya, I., Takahata, H.,
Nomura, H., Watanabe, H., Costa, E., Guidotti, A., 2001. Brain 5diphydroprogesterone and allopregnanolone synthesis in a mouse
model of protracted social isolation, 98, 28492854.
Edwards, F.A., Konnerth, A., Sakmann, B., 1990. Quantal analyses of
inhibitory synaptic transmission in the dentate gyrus of rat hippocampal slices: a patchclamp study. Journal of Physiology 430,
213249.
Guidotti, A., Costa, E., 1998. Can the antidysphoric and anxiolytic
profiles of SSRIs be related to their ability to increase brain 3, 5tetrahydroprogesterone (ALLO) availability? Biological Psychiatry
44, 865873.
Guidotti, A., Dong, E., Matsumoto, K., Pinna, G., Rasmusson, A.M.,
Costa, E., 2001. The socially-isolated mouse: a model to study the
putative role of allopregnanolone and 5-dihydroprogesterone in
psychiatric disorders. Brain Research Review 37, 110115.
Harrison, N.L., Majewska, M.D., Harrington, J.W., Barker, J.L., 1987.
Structureactivity relationships for steroid interaction with the
GABAA receptor complex. Journal of Pharmacological Experimental Therapeutics 241, 346353.
Hill, M.W., Reddy, P.A., Covey, D.F., Rothman, S.M., 1998. Contribution of subsaturating GABA concentrations to IPSCs in cultured
hippocampal neurons. The Journal of Neuroscience 18, 51035111.
Holt, D.A., Levy, M.A., Oh, H.J., Erb, J.M., Heaslip, J.I., Brandt, M.,
Lan-Hargest, Y., Metcalf, B.W., 1990. Inhibition of steroid 5reductase by unsaturated 3-carboxysteroids. Journal of Medical
Chemistry 33, 943950.
Jones, M.V., Westbrook, G.L., 1995. Desensitized states prolong
GABAA channel responses to brief agonist pulses. Neuron 15,
181191.
Karavolas, H.J., Hodges, D.R., 1991. Metabolism of progesterone and
related steroids by neural and neuroendocrine structures. In: Costa,
E., Paul, S.M. (Eds.), Neurosteroids and Brain Function. Thieme,
New York, pp. 135145.
Lambert, J.J., Belelli, D., Hill-Venning, C., Peters, J.A., 1995. Neurosteroids and GABAA receptor function. Trends in Pharmacology
Science 16, 295303.
Lan, N.C., Chen, J.-S., Belelli, D., Pritchett, D.S., Seeburg, P.H., Gee,

G. Puia et al. / Neuropharmacology 44 (2003) 4955

K.W., 1990. A steroid recognition site is functionally coupled to

an expressed GABAA-benzodiazepine receptor. European Journal
of Pharmacology: Molecular Pharmacology 188, 403406.
Majewska, M.D., 1990. Steroid regulation of GABAA receptor: ligand
binding, chloride transport and behavior. In: Costa, E., Paul, S.M.
(Eds.), Steroids and Neuronal Activity. Wiley, Chichester, pp.
83106.
Matsumoto, K., Uzunova, V., Pinna, G., Taki, K., Uzunov, D.P., Watanabe, H., Mienville, J.-M., Guidotti, A., Costa, E., 1999. Permissive role of allopregnanolone content in the regulation of pentobarbital-induced righting reflex loss. Neuropharmacology 38, 955963.
Mienville, J.-M., Vicini, S., 1989. Pregnenolone sulfate antagonizes
GABAA receptor-mediated currents via a reduction of channel
opening frequency. Brain Research 489, 190194.
Nguyen, Q., Sapp, D.W., VanNess, P.C., Olsen, R.W., 1995. Modulation of GABAA receptor binding in human brain by neuroactive
steroids: species and brain regional differences. Synapse 19, 7787.
Pinna, G., Uzunova, V., Matsumoto, K., Puia, G., Mienville, J.-M.,
Costa, E., Guidotti, A., 2000. Brain allopregnanolone regulates the

55

potency of the GABAA receptor agonist muscimol. Neuropharmacology 39, 440448.

Puia, G., Ducic, I., Vicini, S., Costa, E., 1993. Does neurosteroid
modulatory efficacy depend on GABAA receptor subunit composition? Receptors and Channels 1, 135142.
Puia, G., Santi, M.R., Vicini, S., Pritchett, D.B., Purdy, R.H., Paul,
S.M., Seeburg, P.H., Costa, E., 1990. Neurosteroids act on human
recombinant GABAA receptors. Neuron 4, 759765.
Rupprecht, R., Holsboer, F., 1999. Neuroactive steroids: mechanisms
of action and neuropsychopharmacological perspectives. Trends in
Neuroscience 22, 410416.
Uzunov, D.P., Cooper, T.B., Costa, E., Guidotti, A., 1996. Fluoxetineelicited changes in brain neurosteroid content measured by negative
ion mass fragmentography. Proceedings of the National Academy
of Sciences of the United States of America 93, 1259912604.
Zhu, W.J., Vicini, S., 1997. Neurosteroid prolongs GABAA channel
deactivation by altering kinetics of desensitized states. Journal of
Neuroscience 17, 40224031.