Abstract
3α,21-Dihydroxy-5α-pregnan-20-one (5α-THDOC) and 3α-hydroxy-5α-pregnan-20-one (3α,5α-P) have full efficacy as allosteric modulators of [35S]t-butylbicyclophosphorothionate ([35S]TBPS) binding to sites on the γ-aminobutyric acid (GABA) type A receptor complex (GRC). Relative to 3α,5α-P and 5α-THDOC, 3α,21-dihydroxy-5β-pregnan-20-one (5β-THDOC) has limited efficacy as an allosteric modulator of [35S]TBPS binding. Interactions between 3α,5α-P, 5α-THDOC and 5β-THDOC were examined to determine whether these neuroactive steroids share a common site for modulation of the GRC. The concentration-response curves for both 3α,5α-P and 5α-THDOC modulation of [35S]TBPS binding to brain and recombinantly derived GRCs are shifted rightward in the presence of various concentrations of 5β-THDOC. Similarly, 5β-THDOC modulates GABA-evoked Cl− currents with low efficacy and inhibits the potentiation of GABA-evoked Cl− currents by 3α,5α-P. Furthermore, behavioral studies reveal that 5β-THDOC antagonizes 3α,5α-P-induced loss of the righting reflex in mice at a dose that has no effect alone. These results represent the first demonstration of antagonist-like actions of a neuroactive steroid on the GRCs at levels ranging from the receptor to animal behavior and suggest the existence of partial agonist neurosteroids.
3α,5α-P and 5α-THDOC are reduced metabolites of progesterone and deoxycorticosterone, respectively. They are among the most potent endogenously occurring neuroactive steroids with high specificity for the GRC (Gee et al., 1987, 1988; Harrison et al., 1987; Lan et al., 1990; Majewska et al., 1986; Morrow et al., 1987, 1990; Peters et al., 1988; Turner et al., 1989). In contrast to hormonal steroids, these neuroactive steroids activate a membrane-bound steroid site to exert rapid and reversible effects on GABA-gated Cl− channel conductance. The remarkable pharmacological potency of these steroids and the unique structure-activity requirements of their receptors distinguish them from the hormonal steroids and their cytosolic receptors. Consistent with their GABA-agonist like activity at the GRC, these neuroactive steroids share similar pharmacological effects with benzodiazepines and barbiturates, including anesthetic, sedative-hypnotic, anxiolytic and anticonvulsant actions (Gee et al., 1995). 3α,5α-P and 5α-THDOC allosterically modulate the GRC in a manner reminiscent of a ligand that potentiates GABA action. For example, these steroids inhibit [35S]TBPS and potentiate [3H]muscimol and [3H]flunitrazepam binding (Gee et al., 1988; Majewska et al.; 1986).
Endogenously occurring neuroactive steroids have been reported to reach concentrations in the brain well within the range necessary to potentiate the actions of GABA (Gee et al., 1987; Majewskaet al., 1986; Paul et al., 1991; Purdy et al., 1991). These findings have provided the impetus to determine the possible physiological role of these compounds. As a part of this effort, studies have revealed neuroactive steroids with a varying range of efficacies as modulators of the GRC (Gee et al., 1988;Gee and Lan, 1991). Neuroactive steroids with partial agonist and antagonist activity have been reported (Gee and Lan, 1991; McCauleyet al., 1995). The former may have significant implications for the development of pharmacological agents with therapeutic value, whereas the latter are of interest as tools for evaluation of the physiological importance of endogenous neuroactive steroids. Previous studies with 5α-THDOC and its stereoisomer 5β-THDOC revealed important differences in potency, efficacy and regional selectivity at the GRC, accounted for by only a difference in the spatial orientation of the steroid A-ring (Gee and Lan, 1991). The chemical structures showing this difference in spatial orientation are depicted in figure1. In contrast to the 5α-reduced analogs, 5β-THDOC has limited efficacy and, under certain conditions, antagonist actions at the GRC in vitro. This interesting pair of neuroactive steroids provided the means to evaluate in detail the site and mechanism of action of 5β-THDOC in vitro and in vivo.
Materials and Methods
Tissue Preparation
The brains from male Sprague-Dawley rats (160–200 g; Simonsen Laboratories, Gilroy, CA) were removed immediately after sacrifice, and the frontal cortex from each animal was dissected over ice. A P2 homogenate was prepared for radioligand binding assays as described previously (Gee et al., 1987). Briefly, the tissue was gently homogenized (with a Teflon pestle) as a 10% (w/v) suspension in 0.32 M sucrose, followed by centrifugation at 1000 × g for 10 min at 0° to 4°C. The supernatant was collected and centrifuged at 9000 × g for 20 min at 0° to 4°C. The resultant pellet was washed three times in 100 volumes of ice-cold PBS (50 mM sodium/potassium phosphate, 200 mM NaCl, pH 7.4) by centrifugation at 9000 × g for 10 min and resuspended as a 10% (w/v) homogenate for immediate use in binding assays.
Stable Cell Lines
Stable expression of GABAA receptor subunits.
Human alpha-1 and gamma-2L GABAA receptor subunits were cloned into mammalian expression vectors containing the constitutive enhancer/promoter of the immediate early gene of human cytomegalovirus, pCDM8 (InVitrogen, San Diego), or pcDNA1 (InVitrogen). The human beta-2 cDNA was obtained from the rat beta-2 sequence (Ymer et al., 1989) through mutation of the codon for amino acid 347 from Asn to Ser (Hadingham et al., 1993) and then ligation into the pcDNA1 expression vector.
All plasmid DNA for transfection was prepared using two-cycle cesium chloride gradient centrifugation. The transfection of the HEK 293 cells (CRL 1573; American Type Culture Collection, Rockville, MD) follows the protocol reported previously (Ishiura et al., 1982). The expression plasmid pY3 was cotransfected with the GABA subunits for stable cell selection with the antibiotic hygromycin B. The cells were incubated in a 3% CO2, 35°C humidified incubator for 16 to 20 hr with Dulbecco’s modified Eagle’s medium (Life Technologies, Grand Island, NY) containing 10% fetal bovine serum. The Dulbecco’s modified Eagle’s medium was then changed, and the cells were incubated in a 5% CO2, 37°C humidified incubator.
Selection was started 48 hr later by replacing the medium with complete medium plus 200 μg/ml hygromycin B (Calbiochem, San Diego, CA). After 2 weeks, resistant colonies were trypsinized in cloning cylinders and transferred to 12-well plastic plates. Individual cell lines were expanded and maintained in the same medium. All cell lines were analyzed for the presence of GABA receptor mRNAs by reverse transcription of total RNA followed by polymerase chain reaction usingalpha-1-, beta-2- andgamma-2L-specific primers. Cell lines shown positive for thealpha, beta and gamma subunit mRNA transcripts were then analyzed for the presence of GABAAreceptor complexes by their ability to bind [35S]TBPS.
Membrane preparation.
Cells were harvested by removing the incubation media and replacing with 1 ml of 10× trypsin-EDTA solution (Life Technologies). After 5-min incubation with gentle agitation, 9 ml of serum-containing media was added, and the cells were released from the flask by gentle pipetting up and down of the culture medium. The culture medium was then removed by low-speed centrifugation at 1000 × g for 10 min and rinsed twice with cold 200 mM NaCl/50 mM Na-K phosphate, pH 7.4, buffer (PBS). Cell membranes were disrupted using a Polytron (Brinkmann Instruments, Westbury, NY) at setting 10 for 20 sec. The cell homogenate was centrifuged at 9000 × g for 20 min, and the pellet rinsed once before resuspension with cold PBS in the desired volume for the [35S]TBPS binding assay.
[35S]TBPS binding assay.
[35S]TBPS (2 nM, 60–120 Ci/mmol; New England Nuclear, Boston, MA) was incubated with 100-μl aliquots of cortical P2 homogenate or cell membrane containing GRCalpha-1-beta-2-gamma-2L subunits in the presence and absence of various concentrations of steroids. All test drugs were dissolved in DMSO (Sigma Chemical Co, St Louis) and added to the incubation mixture in 5-μl aliquots. The incubation mixture was brought to a final volume of 1 ml with assay buffer. Nonspecific binding was defined as binding in the presence of 2 μM TBPS (Research Biochemicals, Natick, NH). The binding assays were performed in the presence or absence of 5 μM GABA (IC50value for GABA inhibition of [35S]TBPS binding under the condition used). The incubation (90 min, 25°C) was terminated by rapid filtration through glass fiber filters (No. 32; Schleicher & Schuell, Keene, NH). The filters were washed three times with 3 ml of ice-cold phosphate buffer, and filter-bound radioactivity was quantified by liquid scintillation spectrophotometry. Protein concentration was determined according to the method of Lowry et al. (1951). The dose-response data were evaluated by computerized nonlinear regression (InPlot; GraphPAD, San Diego, CA) using a one-component (three-parameter) model to generate IC50values (Boxenbaum et al., 1971). The data collected from the receptor binding assays were analyzed by ANOVA and Newman-Keuls (P < .05) when warranted (Winer, 1971).
Electrophysiological Studies
Preparation of cRNA.
Preparation of cRNAs for thealpha-1, beta-2 and gamma-2L subunits was performed as described previously (Hadingham et al., 1993; Ishiura et al., 1982; Ymer et al., 1989). cRNA was diluted to 1 μg/μl with DEPC-treated water and stored in 1- to 2-μl aliquots at −80°C until injection. Stocks of cRNA were thawed, mixed and diluted as noted below in H2O immediately before injection.
Xenopus laevis oocyte expression system.
Mature female X. laevis (Xenopus I, Ann Arbor, MI) were immersed in 0.15% 3-aminobenzoic ethyl ester (MS-222; Sigma Chemical, St. Louis, MO) until fully anesthetized (30–45 min), and two to four ovarian lobes were surgically removed and placed into Barth’s medium containing (in mM) 88 NaCl, 1 KCl, 0.41 CaCl2, 0.33 Ca(NO3)2, 0.82 MgSO4, 2.4 NaHCO3 and 5 HEPES, pH adjusted to 7.4 with NaOH. With slight modifications of established procedures (Woodward et al., 1994), oocytes (developmental stages V and VI) were plucked from the ovary and enzymatically defolliculated by treatment for 45 to 60 min with collagenase (0.5 mg/mL, Boehringer-Mannheim, Indianapolis, IN). After brief vortex-mixing to dislodge epithelia, theca and most of the follicular layer, oocytes were rinsed extensively with fresh Barth’s medium and incubated overnight in Barth’s medium supplemented with gentamycin (0.1 mg/ml). Individual oocytes were microinjected with a 5:1:1 ratio of cRNA encoding the GABAA receptor subunitsalpha-1, beta-2 and gamma-2L (∼5 ng of the alpha-1 subunit and ∼1 ng each of beta-2 and gamma-2L/oocyte). After injection, oocytes were maintained in Barth’s supplemented with gentamycin (0.1 mg/ml) at 15° to 18°C.
Electrical recordings at a holding potential of −70 mV were made from oocytes using two-electrode voltage-clamp (Dagan TEV-200) at 7 to 11 days after injection. Individual oocytes were placed on a nylon mesh in a standard 35-mm tissue culture dish and continually perfused with frog Ringer’s solution containing (in mM) 115 NaCl, 2 KCl, 1.8 CaCl2 and 5 HEPES, pH adjusted to 7.4 with NaOH. Oocytes were perfused with fresh Ringer’s solution and exposed to GABA and steroids via a gravity-driven perfusion system that consisted of a three-barrel linear array of capillary tubes (Hawkinsonet al., 1996).
Drugs.
Neurosteroids were dissolved at 10 mM in DMSO (Sigma) and further diluted to make a series of DMSO stock solutions over the range of 0.001 to 10 mM. Working solutions were made by dilution of these DMSO stock solutions into Ringer’s solution just before application, with final DMSO concentrations of 0.1% to 0.3%. At this dilution, DMSO alone had little or no measurable effects on the GABA control responses. DMSO stocks were stored at room temperature in the dark for ∼7 days without apparent changes in potency. The neurosteroids 3α,5α-P and 5β-THDOC were synthesized by CoCensys, Inc. (Irvine, CA); other reagents were from Sigma.
Experimental design and data analysis.
GABA concentration-response data were obtained through successive exposures to increasing concentrations of GABA, until an apparent maximal current was reached (3–10 mM GABA). These data were analyzed using a PC-based graphing program (Origin, Microcal, Inc.). The following logistic equation was used to fit individual concentration-response data, wheren is the slope, EC50 is the concentration of GABA that produces a half-maximal response, I is the current at a given concentration of GABA and Imax is the maximal current in response to GABA.
Behavioral studies.
Male Swiss Webster mice (23–28 g, Simonsen Laboratories, Gilroy, CA) were used in the LRR studies. Mice were maintained under a 12-hr light/dark cycle with food and waterad libitum. In a given experiment, mice were randomly assigned to each test group. Testing was performed between 9:00 a.m. and 4:00 p.m. 3α,5α-P was tested at doses of 0, 1.25, 2.5, 5 and 7.5 mg/kg i.v. (injection volume is 30 μl) in 20% hydroxypropyl-β-cyclodextrin (Research Biochemicals, Natick, MA). At 30 sec after injection, each mouse was placed on its back. The observation period continued for 2 min after i.v injection. Any mice failing to right itself by returning to an upright position within 2 min was scored as having lost its righting reflex. Mice in the 5β-THDOC + 3α,5α-P group were administered 5β-THDOC (1 mg/kg i.v.) immediately before the injection of 3α,5α-P at 1.25, 2.5, 5, 7.5 or 12.5 mg/kg i.v., respectively. Mice receiving 5β-THDOC alone did not show LRR at any time up to 30 min after the injection. The ED50 values were determined according to the method ofLitchfield and Wilcoxon (1949), and the significance of the potency ratio in the presence or absence of 5β-THDOC was determined by the χ2 test.
Results
Effect of 5β-THDOC on 3α,5α-P and 5α-THDOC on the modulation of [35S]TBPS binding to rat cortex and alpha-1-beta-2-gamma-2Lβ subunit containing GRCs.
In the presence of 5 μM GABA, 3α,5α-P, 5α-THDOC and 5β-THDOC inhibited [35S]TBPS binding in rat cortex with IC50values of 29, 99 and 145 nM respectively (fig. 2). Consistent with previous results (Gee and Lan, 1991), the endogenous neuroactive steroids 3α,5α-P and 5α-THDOC inhibit [35S]TBPS binding in rat cortex with apparent full efficacy (i.e., 100% inhibition). As shown as in figure 2, 5β-THDOC inhibited [35S]TBPS binding in rat cortex with certain limited efficacy. In light of the difference in efficacy between 5β-THDOC and the apparent full-efficacy neuroactive steroids 3α,5α-P and 5α-THDOC, it was of interest to study the interactions among 5β-THDOC, 3α,5α-P and 5α-THDOC in the modulation of [35S]TBPS binding. Fixed concentrations of 5β-THDOC of ≤3 μM caused a parallel rightward shift in the dose-response curves for both 3α,5α-P and 5α-THDOC displacement of [35S]TBPS binding (fig. 3, A and B). The magnitude of the rightward shift for both curves was dependent on the concentration of 5β-THDOC. When the data from these dose-response curves were subjected to Schild analysis (Schild, 1949), the pA 2 values were 6.7 for 3α,5α-P +5β-THDOC and 7.04 for 5α-THDOC + 5β-THDOC (fig. 4). The slopes of the Schild plots were 0.99 for 3α,5α-P +5β-THDOC and 0.98 for 5α-THDOC +5β-THDOC (fig. 4). On the basis of the pA 2 values, the apparent potencies of 5β-THDOC as an antagonist of 3α,5α-P and 5α-THDOC were ∼200 and 90 nM, respectively. These values are in reasonable agreement with 5β-THDOC IC50 values of 145 nM against [35S]TBPS binding when measured directly under similar assay conditions. The results are consistent with the hypothesis that 5β-THDOC is a partial agonist at the neurosteroid site recognized by 3α,5α-P and 5α-THDOC.
Previous studies have shown regional differences in the potency and efficacy of 5β-THDOC, and other neuroactive steroids, as modulators of the GRC. These differences may result in part from the heterogeneity of GABAA receptors (Gee and Lan, 1991; Morrow et al., 1990; Wisden et al., 1992). To obviate the influence of GRC heterogeneity on the response to 5β-THDOC, the effects of this steroid was studied on a recombinantly expressed GRC. Similar to the observations in brain, 5β-THDOC produced a rightward parallel shift in the 3α,5α-P/[35S]TBPS displacement curve using membranes from stably transfected HEK cells expressing GRCs containing the alpha-1-beta-2-gamma-2L subunits (fig. 5). The apparent potency of 3α,5α-P was decreased from 29 to 208 nM (IC50) by 1 μM 5β-THDOC and to 636 nM by 3 μM 5β-THDOC. The effect of 1 and 3 μM 5β-THDOC alone on [35S]TBPS binding toalpha-1-beta-2-gamma-2L receptors was ∼61% and ∼75%, respectively. The magnitude of the reductions in potency approximate the changes seen in studies using rat cortical homogenates. The findings from the recombinantly expressed GRCs support the hypothesis that 5β-THDOC is a competitive antagonist of 3α,5α-P in modulation of [35S]TBPS binding to the GRC.
Effects of 5β-THDOC and 3α,5α-P onalpha-1-beta-2-gamma-2L-containing GRCs expressed in the X. laevis oocyte
Injection of the cRNA encoding the GABAA receptor subunits alpha-1, beta-2 and gamma-2L into oocytes resulted in strong expression of GABA-evoked currents. Maximal responses to GABA (3–10 mM) were 2640 ± 160 nA, ranging between 1900 and 3600 nA (n = 11). As described previously in alpha-1-beta-1-gamma-2L receptors (Hawkinson et al., 1996), responses in oocytes expressing alpha-1-beta-2- gamma-2L receptors were insensitive to blockade by 100 μM ZnCl2, whereas alpha-1-beta-1-mediated responses were blocked by >80% (data not shown). This indicates that the receptors used for steroid assays were ternary, containing gamma-2L subunits, although the precise subunit stoichiometry is unknown. The EC50 value for GABA in these oocytes was 35 ± 3 μM, with a slope of 1.2 ± 0.05, which is within the range normally observed for this subunit combination.
The steroids 5β-THDOC and 3α,5α-P both induced potentiation of GABA control currents (5% of GABA maximal currents) (Fig.6). For 5β-THDOC, potentiation of GABA control responses was detectable at concentrations as low as 10 nM. Maximal potentiation was evoked by 3 μM 5β-THDOC (48 ± 5%, expressed as a percentage of maximal GABA currents). A slight decrease in the magnitude of potentiation was observed at 10 μM. The EC50value for 5β-THDOC-induced potentiation was 670 ± 240 nM, with a slope of 1.2 ± 0.27 (n = 4). In comparison, potentiation induced by 3α,5α-P resulted in greater efficacy of potentiation and higher apparent potency. For 3α,5α-P, the threshold for detecting potentiation was ∼3 nM, and maximal potentiation was observed at 10 μM (89 ± 5% of GABA maximum). The EC50 value for 3α,5α-P was 220 ± 19 nM, and the slope was 1.5 ± 0.12 (n = 4). As observed with other neuroactive steroids, direct steroid-activated currents were observed at concentrations of >1 μM (Fig. 6). Maximum steroid currents were 14 ± 3% of GABA maximal currents for 3α,5α-P (10 μM) but only 1 ± 0.2% of the GABA maximum for 5β-THDOC at 10 μM (Fig. 6).
The blocking effects of the low-efficacy agonist 5β-THDOC on modulation induced by the full agonist 3α,5α-P were also tested. As illustrated in figure 7 (left), coapplication of 5β-THDOC (3 μM) with 3α,5α-P (1 μM) resulted in significantly less potentiation of GABA currents than that observed with 3α,5α-P alone. This blockade could to a large degree be surmounted by increasing the concentration of 3α,5α-P to 10 μM. Combined data from 10 such experiments are shown in figure 7 (right); these data suggest that 5β-THDOC has characteristics consistent with those of a moderate potency partial agonist for the neurosteroid site on the GRC.
Behavioral studies.
Both 3α,5α-P and 5α-THDOC has been reported to have central nervous system depressant actions, including sedative-hypnotic and anesthetic effects mediated through modulation of the GRC (Gee et al., 1995). Based on the in vitrostudies, 5β-THDOC could potentially have antagonist actions against full-efficacy neuroactive steroid such as 3α,5α-P in vivo. Consequently, the LRR in mice was used as a behavioral measure of GRC-mediated central nervous system depression. 3α,5α-P induced dose-dependent LRR with an ED50 value of 3.5 mg/kg (table 1). 5β-THDOC at 1 mg/kg (a dose that did not produce LRR alone) increased the ED50 value for 3α,5α-P induced LRR to 7 mg/kg. The dose at which 5β-THDOC induced the LRR was >2 mg/kg i.v. The solubility of 5β-THDOC above this dose was limited, which prevented further evaluation of its ability to induce an LRR. These behavioral findings are consistent with 5β-THDOC having antagonist action in vivo at the same site as 3α,5α-P on the GRC.
Discussion
Previous studies have shown that 5β-THDOC has limited efficacy in the modulation of [35S]TBPS binding and36Cl uptake in the rat cortex and no efficacy in the spinal cord (Gee and Lan, 1991). 5β-THDOC was also shown to antagonize 3α,5α-P modulation of [35S]TBPS binding in rat spinal cord. Based on these data, it was proposed that 5β-THDOC was either a partial agonist or a receptor subtype-selective ligand. In the present study, the basis for the apparent limited efficacy of 5β-THDOC at the GRC was investigated in detail both in vitro and in vivo.
In binding assays, 5β-THDOC behaves in a manner consistent with interaction with a site recognized by 5α-THDOC and 3α,5α-P. The pA 2 values derived from Schild analysis are in the same range as the IC50 values for 5β-THDOC inhibition of [35S]TBPS binding. These data suggest that 3α,5α-P, 5α-THDOC and 5β-THDOC act at the same site in rat cortex. However, interpretation of the data rat cortex is complicated by receptor heterogeneity (Laurie et al., 1992; Wisdenet al., 1992). GABAA receptor subtypes have different pharmacological properties (Draguhn et al., 1990;Smart, et al., 1991). In particular, subunit composition of the GRC produces apparent changes in the allosteric modulatory effects of neuroactive steroids (Lan et al., 1990, 1991; Puiaet al., 1991; Shingai et al., 1991). Expression of GABAA receptor subunits of known composition generates a homogeneous system for studying the interactions between 5β-THDOC and 3α,5α-P. Using a system with defined subunit composition rules out the possibility that the apparent limited efficacy of 5β-THDOC in cortical homogenates results from the modulation of a subpopulation of [35S]TBPS binding sites. Consistent with the results observed in the cortex, 5β-THDOC (1 and 3 μM) causes a rightward parallel shift of the 3α,5α-P/[35S]TBPS dose-response curves without changing the degree of maximum inhibition. The magnitude of the shift in the IC50 values for 3α,5α-P-induced by 5β-THDOC is close to that observed in cortical homogenates. Combined, the data from cortical homogenates and expressed GRCs of known subunit composition provide strong support that 5β-THDOC is a partial agonist acting at the same site as 3α,5α-P.
Electrophysiological studies similarly suggest that 5β-THDOC has partial agonist activity in the modulation of GABA currents mediated byalpha-1-beta-gamma-beta-2-containing GRCs expressed in oocytes. Coapplication of 5β-THDOC with 3α,5α-P results in significantly less potentiation of GABA currents than is observed with 3α,5α-P alone. Most importantly, this inhibition could be surmounted by increasing the concentration of the full agonist. In combination with the binding data, these observations strongly suggest that the antagonist activity of 5β-THDOC is mediated by the neurosteroid binding site at which 3α,5α-P evokes full-efficacy modulation.
In vivo, 5β-THDOC (1 mg/kg) induces a 2-fold increase in the ED50 value for 3α,5α-P induction of LRR in mice. This dose of 5β-THDOC alone did not produce LRR. Thus, the antagonist actions of 5β-THDOC can be observed both in vitro andin vivo. However, the potency of 5β-THDOC in blocking 3α,5α-P induction of LRR was observed to be greater than thatin vitro. The possible explanations for the greater potencyin vivo are additional factors such as different level of endogenous GABA or enhanced bioavailability of 5β-THDOC.
Earlier studies have shown that certain non-3α-hydroxylated neuroactive steroids interact with the GRC as allosteric antagonists of GABA action (Demirgoren et al., 1991; Majewska et al., 1986, 1988). These neuroactive steroids do not appear to produce their actions at the same site as 3α,5α-P (Gee et al., 1989). The existence of competitive neuroactive steroid antagonists with high affinity has not yet been unequivocally demonstrated. Although evidence has been presented to suggest that 5β-pregnan-3β-ol-20-one competitively antagonizes the action of 5β-pregnan-3α-ol-20-one as a modulator of [3H]flunitrazepam binding to the benzodiazepine receptor (Prince and Simmonds, 1992), the high concentration of steroid required to block the potentiation of benzodiazepine receptor binding, ∼60 μM in these studies, suggests that the ideal structure-activity requirements for competitive antagonists have not yet been identified. The structural leads provided by 5β-pregnan-3β-ol-20-one and 5β-THDOC may be useful avenues of approach in the search for a pure antagonist.
Collectively, the results of the present study strongly support the hypothesis that 5β-THDOC is a partial agonist at the same site as the full agonists 5α-THDOC and 3α,5α-P on the GRC. Both 3α,5α-P and 5α-THDOC are detected in the rat brain (Purdy et al., 1991). Brain levels of these steroids after swim stress reach concentrations that are sufficient to potentiate GABA action when based on the in vitro levels necessary to modulate GABA action at the GRC (Gee et al., 1987; Majewska et al., 1986;Purdy et al., 1991). Whether sufficient levels of the 5β-isomer are produced in the central nervous system to affect GRC function is uncertain, although 5β-reductase activity has been detected in the central nervous system and periphery (Mickan, 1972;Mickan and Zander, 1979). It is possible, therefore, that 5β-THDOC has a role in the endogenous modulation of the GABAAreceptor. The relative contribution of regional selectivity of 5α-THDOC and 5β-THDOC to the pharmacological profile is not currently known despite evidence suggesting that 5β-THDOC shows regional selectivity (Gee and Lan, 1991). Nevertheless, it is possible that the pharmacological profile of 5β-THDOC is a reflection of both its limited efficacy and its regional selectivity.
Especially intriguing is the question of the physiological role of a endogenous partial agonist ligand in the modulation of the GRC relative to those of full agonists. Moreover, whether apparent partial agonist neuroactive steroids of this type have unique pharmacological profiles relative to their full efficacy counterparts remains to be determined. Recently, the synthetic partial agonist neurosteroid 3α-hydroxy-3β-trifluoromethyl-5β-pregnan-20-one was synthesized and characterized in vitro (Hawkinson et al., 1996). This synthetic neuroactive steroids may be a useful tool for the evaluation of the in vivo pharmacological profile of limited efficacy neuroactive steroids because of its lower efficacy and metabolic lability (i.e., rapid metabolic degradation and elimination) relative to 5β-THDOC.
In conclusion, the present study is the first demonstration of competitive antagonism of endogenous neuroactive steroids acting on the GRC. Our findings that 5β-THDOC can antagonize both in vitro and in vivo actions of neuroactive steroids on the GRC raise the possibility that appropriate modification of the 5β-THDOC molecule may give rise to high-affinity steroids that are pure antagonists at the neuroactive steroid site on the GRC. Such antagonists will provide essential tools for the elucidation of the physiological role of endogenous neuroactive steroids.
Acknowledgments
The authors thank Dr. John Drewe and Mr. J.-S. Chen for the preparation of the stable cell line.
Footnotes
-
Send reprint requests to: Kelvin W. Gee, Ph.D., Department of Pharmacology, College of Medicine, University of California, Irvine, Irvine, CA 92697.
-
↵1 This work was supported by a grant from CoCensys, Inc.
- Abbreviations:
- 3α
- 5α-P, 3α-hydroxy-5α-pregnan-20-one
- GABA
- γ-aminobutyric acid
- GRC
- γ-aminobutyric type A receptor complex
- DMSO
- dimethylsulfoxide
- TBPS
- t-butylbicyclophosphorothionate
- 5α-THDOC
- 3α,21-dihydroxy-5α-pregnan-20-one
- 5β-THDOC
- 3α,21-dihydroxy-5β-pregnan-20-one
- LRR
- loss of the righting reflex
- ANOVA
- analysis of variance
- i.v.
- intravenous
- Received July 23, 1996.
- Accepted February 14, 1997.
- The American Society for Pharmacology and Experimental Therapeutics