Introduction

Top Abstract Introduction Materials Results Discussion References

The ligand-gated GABA_A receptors are an important site of action for a variety of centrally acting drugs such as benzodiazepines, barbiturates, anesthetics, neurosteroids, picrotoxin and TBPS-like cage convulsants (Olsen and Venter, 1986; Ticku, 1991; Majewska, 1992; Gee et al., 1995; Mehta and Ticku, 1995). This receptor is an oligomeric complex with multiple binding sites which bear an allosteric relationship with each other. The molecular biological studies have demonstrated that the GABA_A receptors are composed of multiple subunits, and a minimum combination of ₂ is required to form functional receptors (Pritchett et al., 1989), and it is likely that the stoichiometry of GABA_A receptors may be 2:2:1 subunits (Chang et al., 1996). However, the exact stoichiometry of the GABA_A receptors in vivo is unknown.

There is evidence that several neurosteroids found in the brain alter the central nervous system excitability and function by nongenomic mechanisms. These neurosteroids include pregnenolone, DHEA, pregnenolone sulfate and DHEAS. These neurosteroids modulate GABA function in both positive and negative manners. Progesterone and its active metabolite, 5-pregan-3-ol-20-one (53), are positive modulators of GABA_Aergic transmission (Harrison et al., 1984; Morrow et al., 1987; Majeswka, 1992; Yu and Ticku, 1995a, 1995b). These neurosteroids enhance the binding of benzodiazepine and GABA agonists and inhibit the binding of TBPS. Although these effects appear to be similar to barbiturate modulation of GABA receptors (reviewed in Ticku, 1991; Gee et al., 1995), these positive modulators of GABAergic responses bind to a distinct site (Turner et al., 1989; Gee et al., 1995). In contrast, pregnenolone sulfate and DHEAS are negative modulators of GABA responses (Majewska et al., 1990a, b; Demirgoren et al., 1991; Spivak, 1994). The latter studies have also demonstrated that DHEAS inhibited GABA responses in a noncompetitive manner. Furthermore, pregnenolone sulfate, but not DHEAS, has been reported to inhibit TBPS binding to brain membranes (Majewska et al., 1989). It has also been suggested that DHEAS may inhibit GABA responses by binding to the barbiturate site of the GABA receptor complex. In addition, DHEA does not inhibit the binding of DHEAS to brain membranes, suggesting that they may also bind to distinct sites (Demirgoren et al., 1991). Despite the lack of this interaction, both DHEA and DHEAS have been shown to inhibit GABA-induced currents (Demirgoren et al., 1991).

In summary, these and other observations in the literature suggest that the positive and negative neurosteroids bind to distinct sites on the GABA_A receptor complex. We have investigated the interaction of DHEAS with various binding sites associated with the GABA_A receptor complex using brain membrane homogenates, and examined its effects on GABA-induced [³⁶Cl] influx in cultured mammalian cortical neurons. Our results indicate that DHEAS, but not DHEA, interacts competitively with the picrotoxin/TBPS site, and this interaction may be responsible for the noncompetitive inhibition of the GABA responses reported in the literature.

	Materials

Top Abstract Introduction Materials Results Discussion References

[³⁶Cl]-HCl was purchased from ICN Radiochemicals (Irvine, CA). [³H]-flunitrazepam, [³⁵S]TBPS and [³H]GABA were purchased from New England Nuclear. GABA, poly-L-lysine hydrobromide (molecular weight > 300,000), DHEA, DHEAS, 5-fluoro-2-deoxyuridine, uridine and picrotoxinin were purchased from Sigma Chemical Co., St. Louis, MO. Bicinchoninic acid protein assay reagents were purchased from Pierce Chemicals (Rockford, IL). MEM and horse serum were purchased from Gibco (Santa Clara, CA). Fetal bovine serum was purchased from Atlanta Biologicals (Norcross, GA). Male Sprague Dawley rats and female and male C57 BL/6 mice (18-20g) were purchased from Harlan (Indianapolis, IN).

Membrane preparation. Membranes were prepared as described previously (Maksay and Ticku, 1985), using male Sprague Dawley rats. Briefly, the rats were decapitated and cerebral cortex and cerebellum dissected. Tissues were stored at -70°C until use. The tissue was thawed and homogenized in 0.32 M sucrose (20 ml/g of tissue) and centrifuged at 1000 g for 10 min. The supernatant was then centrifuged at 140,000 × g for 30 min. The pellet was resuspended in ice- cold distilled-deionized water and subsequently centrifuged at 140,000 × g for 30 min. The pellet was then resuspended and centrifuged at 140,000 × g for 30 min three times in ice-cold Tris-HCl buffer (50 mM pH 7.4). After the final centrifugation step, the pellets were stored frozen (-70°C) overnight. On the day of the assay, the tissue was thawed and washed two more times with buffer as before and then resuspended for use in assays.

[³H]Flunitrazepam binding assays. [³H]Flunitrazepam binding was measured using filtration methods as described previously (Thyagarajan et al., 1983). Briefly, aliquots (0.45-0.6 mg/ml) of cerebral cortex or cerebellar membrane preparation were resuspended in 50 mM Tris-HCl, pH 7.4, and incubated with 1 nM [³H]Flunitrazepam at 4°C for 90 min in the absence and presence of neurosteroids. Nonspecific binding was measured using 10 µM Ro15-1788. DHEA was dissolved in a final concentration of 1% DMSO. After incubation, 300-µl aliquots were filtered on Whatman GF/B filters and rapidly washed twice with 50 mM Tris-HCl, pH 7.4. Individual filters were incubated for 4 hr with 3.5 ml scintillation cocktail (Beckman Ready Protein, Fullerton, CA), before the measurement of radioactivity.

[³H]GABA binding assays. [³H]GABA binding was measured by a centrifugation assay as described previously (Burch et al., 1983; Thyagarajan et al., 1983). Briefly, aliquots of rat brain cerebral cortex or cerebellum membrane (P₂ + P₃ fraction) preparation were resuspended in 50 mM Tris-HCl, pH 7.4, with or without DHEAS and incubated at 4°C for 10 min with 4 nM [³H]GABA. Nonspecific binding was determined in the presence of 100 µM GABA. After incubation, samples were centrifuged at 50,000 × g for 10 min. The pellets were then washed twice with 50 mM Tris-HCl, pH 7.4, and excess buffer was carefully removed from vials, tissue solubilized and radioactivity determined as described (Burch et al., 1983).

[³⁵S]TBPS binding assays. [³⁵S]TBPS binding was measured by filtration as described (Maksay and Ticku, 1985). Tissue aliquots (0.45-0.6 mg/ml) were resuspended in 50 mM Tris-HCl pH 7.4 + 150 mM KCl and incubated with or without DHEAS and 4 nM [³⁵S]-TBPS for 180 min at room temperature. Nonspecific binding was determined by the addition of 100 µM picrotoxinin. Aliquots (300 µl) were then filtered and washed twice with 50 mM Tris-HCl + 150 mM KCl pH 7.4 and processed as described for [³H]-flunitrazepam binding experiments. For saturation analysis, tissue aliquots were incubated with or without DHEAS (50 µM) before adding increasing concentrations of [³⁵S]TBPS (0.5-120 nM).

Dissociation experiments. [³⁵S]TBPS dissociation experiments were performed as a variation of a previously described method (Maksay and Ticku, 1985). Tissue samples were incubated with 3 nM [³⁵S]TBPS at 25°C for 3 hr. Nonspecific binding was determined using 100 µM picrotoxinin. At various time intervals, aliquots (400 µl) of the membrane suspension incubated with [³⁵S]TBPS were added to vials containing either 100 µM picrotoxinin, 1000 µM DHEAS or 1000 µM pentobarbital to initiate dissociation. The samples were then filtered and washed twice and processed as described for [³⁵S]TBPS binding assays. For all binding assays, protein content was estimated using the Bicinchonic acid assay.

Cell culture. Cerebral hemispheres were dissected from 14-day-old C57BL/6 fetuses and dissociated by trituration using a Pasteur pipette as described previously (Yu and Ticku, 1995a, 1995b). Dissociated cells were then resuspended in medium (MEM 10/10) containing 80% Eagle's MEM pH 7.4, glucose (33.3 mM), NaHCO₃ (26.2 mM), 10% horse serum (heat inactivated 30 min at 56°C) and 10% fetal bovine serum. Cells were plated in 1.5 ml of media on 25 mm round coverslips that were coated with poly-L-lysine and placed into an incubator at 37°C with 5% CO₂. After 24 hr, 1 ml of the media was removed from each coverslip and replaced with 1 ml of medium containing 10% horse serum (MEM 10) with a mixture of 5-fluro-2-deoxyuridine and uridine (2 mg/ml of 5-fluro-2-deoxyuridine and 5 mg/ml uridine) at a final concentration of 10 mg/ml. On the third day in culture, 1 ml of media was removed from each coverslip and 1 ml of MEM 10 was added. Coverslips were used on the seventh or eighth day of culture.

Chloride influx. Chloride influx was measured as described previously (Mehta and Ticku, 1992). Briefly, the coverslips were washed twice in 35-mm dishes containing 2 ml of HEPES buffered saline (in mM: NaCl, 136; KCl, 5.4; MgCl₂, 1.4; CaCl₂, 1.2; NaH₂PO₄, 1.0; HEPES, 20) for 3 min each time. The coverslips were then drained well and incubated in various concentrations of DHEAS for 3 min. The coverslips were again drained well and transferred to 2 ml of a solution containing [³⁶Cl] HCl (2 mCi/ml), GABA (10 µM), nipecotic acid (100 µM), DHEAS and HEPES buffered saline for 5 sec, as described (Mehta and Ticku, 1992). Quickly, the coverslips were immersed in 1000 ml ice-cold stop solution pH 7.4 (in mM: NaCl, 150; KCl, 5.4; MgCl₂, 1.4; CaCl₂, 1.2; NOH₂PO₄, 1.0; pentylene tetrazole, 1.0 mM and HEPES, 5.0, adjusted to pH 7.4 with Tris base) for 2 to 3 sec and then immersed into another 1,000 ml of ice-cold stop solution for 7 sec. The coverslips were then well drained, cut in half, and placed into scintillation vials containing 1.5 ml of 0.2 N NaOH. After 1 hr, 0.5 ml from each vial was removed for protein determination and 300 µl of 1 N HCl was added to each vial and mixed well. Hydrofluor scintillation fluid (10 ml) was then added to each vial and vortexed thoroughly. After 1 hr, the vials were then counted to determine radioactivity.

Data analysis. Data for inhibition studies were analyzed using Delta Graph 4.0 software (Macintosh) with user-defined curve fitting. Saturation and Scatchard data were analyzed using nonlinear regression by EBDA and Ligand programs included in KELL software (Macintosh), and dissociation data were analyzed using Kinetic (KELL).

	Results

Top Abstract Introduction Materials Results Discussion References

Effects of DHEAS and DHEA on the binding of ligands to the GABA_A receptor complex. In the initial studies, we examined the effects of DHEAS and DHEA on the binding of [³H]flunitrazepam binding to rat brain cerebral cortex and cerebellum. Table 1 shows that DHEAS up to 100 µM did not alter the binding of this benzodiazepine ligand to either cortical or cerebellar membrane homogenates. Higher concentrations (e.g., 500 µM) of DHEAS did inhibit [³H]flunitrazepam binding, which apparently is a nonspecific effect. Similarly, DHEA up to 100 µM did not alter the [³H]flunitrazepam binding in these regions. Although there was decreased binding in the presence of DHEA, it was due solely to the vehicle (DMSO) used to dissolve this drug (table 1).

Effects of DHEAS and DHEA on the binding of [³H]GABA to its receptor sites. Because DHEA did not affect any of the binding sites within the concentration-range tested (tables 1 and 2), further experiments were performed with DHEAS. Figure 1 shows that DHEAS produced a concentration-dependent inhibition of [³H]GABA binding in cerebral cortex and cerebellum. The IC₅₀ values of DHEAS to inhibit [³H]GABA binding were in a similar range in these regions (see legend to fig.1).

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Concentration-dependent inhibition of [3H]GABA (4 nM) binding by DHEAS to rat brain cerebral cortex and cerebellar membrane homogenates. The values represent the mean ± S.D. of three experiments, each done in triplicate. The IC50 values for DHEAS inhibition of [3H]GABA binding were 109 ± 22 and 138 ± 21 µM in cerebral cortex and cerebellum, respectfully (n = 3).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Saturation (A) and Scatchard (B) analysis of [3H]GABA binding to rat brain cerebral cortex membrane homogenates in the absence and presence of 200 µM DHEAS. The KD and max values were obtained by nonlinear regression by EBDA and ligand programs included in KELL software (Macintosh), and are summarized in table 3.

Effects of DHEAS on the binding of [³⁵S]TBPS to the picrotoxin site of the GABA_A receptor complex. Figure 3 shows that DHEAS inhibited the binding of [³⁵S]TBPS to the picrotoxin site in a concentration-dependent manner in both cerebral cortical and cerebellar membranes preparations, and DHEAS was equipotent in both these regions. The IC₅₀ values are listed in table 2. DHEAS was more potent in inhibiting TBPS binding than GABA binding.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Concentration-dependent inhibition of [35S]TBPS (4 nM) binding to rat cerebral cortical and cerebellar membranes by DHEAS. The values are the mean ± S.D. of three separate experiments, each done in triplicate. The IC50 value inhibition of [35S]TBPS were 46 ± 7 in cerebral cortex and 59 ± 9 in cerebellum (n = 3).

View larger version (20K): [in this window] [in a new window]   Fig. 4.   A representative Scatchard analysis of [35S]TBPS binding to rat brain cerebellar cortical membrane homogenates in the absence and presence of 50 µM DHEAS. The KD and max values are summarized in table 4.

Effects of DHEAS on the dissociation kinetics of [³⁵S]TBPS binding. The rationale for these studies stems from our previous studies demonstrating that drugs that interact with the TBPS/picrotoxin in a competitive manner produce monophasic dissociation of [³⁵S]TBPS, whereas allosteric ligands such as GABA and barbiturates, which also inhibit TBPS binding, greatly accelerate the dissociation profile, suggesting that they bind to distinct sites (Maksay and Ticku, 1985). The dissociation of [³⁵S]TBPS by DHEAS and other drugs was studied as described in the "Methods" and elsewhere (Maksay and Ticku, 1985). Figure 5 shows that picrotoxin dissociated [³⁵S]TBPS binding in a monophasic manner, and similar to that observed with nonradioactive TBPS, suggesting a competitive interaction, as reported previously (Maksay and Ticku, 1985). DHEAS, such as picrotoxin, also dissociated [³⁵S]TBPS in a monophasic manner (fig. 5). In contrast, pentobarbital, which binds to a distinct allosteric site, greatly accelerated the dissociation profile of [³⁵S]TBPS (fig. 5). GABA also greatly accelerated the dissociation profile (data not shown). Table 5 gives the half-lives for the monophasic dissociation for picrotoxin and DHEAS, and the half-life for biphasic dissociation profile observed with pentobarbital. These results suggest that DHEAS interacts with the TBPS/picrotoxin in a competitive manner, and in a manner distinct from barbiturates (fig. 6).

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Representative dissociation of [35S]TBPS binding as induced by DHEAS (), picrotoxin () and pentobarbital (). The dissociation was initiated as described in "Methods." The values are the mean of duplicate determinations that varied  5%. The kinetic constant of dissociation are summarized in table 6.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Inhibition of GABA (10 µM)-induced [36Cl] influx by DHEAS in mouse cortical and cultured neurons. The values are mean ± S.D. of three separate experiments, each using three coverslips, as described in "Methods."

Effect of DHEAS on GABA-induced [³⁶Cl^-]influx in mammalian cortical neurons. To determine if DHEAS was a GABA antagonist, we examined its effects on GABA-induced [³⁶Cl] influx in well-characterized mammalian cortical neurons. These neurons have all the properties of the GABA_A receptor complex, and provide an in vitro model to study GABA receptor pharmacology (e.g., Mehta and Ticku, 1992; Hu and Ticku, 1994). Figure 6 shows that DHEAS produced a concentration-dependent inhibition of GABA-induced (10 µM, IC₅₀ value) [³⁶Cl] influx in cortical neurons. These results are consistent with the notion that DHEAS is a GABA_A antagonist. In contrast, DHEA was much less potent in inhibiting GABA-induced [³⁶Cl] influx by 13 ± 5% at 10 µM (n = 4), and by 33 ± 12% at 100 µM (n = 4) concentrations.

	Discussion

Top Abstract Introduction Materials Results Discussion References

There is strong evidence that neurosteroids modulate GABA_Aergic transmission in the central nervous system. Numerous studies have demonstrated that metabolites of progesterone, such as allopregnanolone (5-pregan-3-ol-20-one, 53), are positive modulators of GABAergic transmission. These interactions include enhancement of GABA and benzodiazepine agonist binding, inhibition of TBPS binding and potentiation of GABA responses measured by electrophysiological and [³⁶Cl] measurements (Harrison et al., 1987; Majewska et al., 1986; Morrow et al., 1987; Gee et al., 1988; Lambert et al., 1990; Turner et al., 1989; Yu and Ticku, 1995a, 1995b). The site for the positive neurosteroids is more distinct than other modulatory sites on the oligomeric GABA_A receptor complex. In contrast, the site of action for the sulfate metabolite of DHEA, DHEAS, which is a noncompetitive antagonist of GABA responses (Demirgoren et al., 1991; Spivak, 1994), has not been well characterized. In electrophysiological studies, DHEAS has been shown to inhibit GABA responses in cultured neurons of ventral midbrain of fetal rats and ventral mesenaphalon in a noncompetitive manner with an IC₅₀ value of 13 ± 3 µM (Demirgoren et al., 1991; Spivak, 1994).

We studied the interaction of DHEAS and DHEA with various sites of the GABA receptor complex in rat brain membranes in order to identify their potential site of action. Additionally, we investigated the effect of DHEAS on GABA-induced [³⁶Cl] influx in mammalian cortical cultured neurons to confirm its GABA antagonistic activity. Our studies demonstrated that DHEA and DHEAS did not interact with the binding of [³H]flunitrazepam to either cerebral cortex or cerebellar membranes. This is consistent with a previous observation regarding a lack of interaction between DHEAS and the benzodiazepine binding sites (Majewska et al., 1990a; Majewska 1992). In contrast, DHEAS, but not DHEA, inhibited the binding of GABA and TBPS to their binding sites both in cerebral cortex and cerebellum. DHEAS inhibited GABA and TBPS binding in a concentration-dependent manner; however, DHEAS was more potent in inhibiting TBPS binding than GABA binding (table 3). The reason for this difference is not clear, but may be related to different modes of interaction of DHEAS with these sites.

To further investigate these interactions, we investigated the effect of DHEAS on the saturation isotherms of GABA and TBPS binding in cerebral cortex. DHEAS, without significantly affecting the K_D values of the high and low-affinity GABA_A receptor binding sites, decreased the B_max values of both the high and low affinity sites in cerebral cortex. These results suggest that DHEAS inhibits GABA binding in an allosteric or noncompetitive manner. In contrast, DHEAS inhibited the TBPS binding in a competitive manner, because it decreased the affinity of TBPS for its binding sites. These results suggest that DHEAS binds to the TBPS/picrotoxin site of the GABA receptor complex, and this would be consistent with the reported noncompetitive inhibition of DHEAS of GABA responses (Majewska et al., 1990a, 1990b; Spivak, 1994). These results are also consistent with other noncompetitive antagonists of GABA that also bind to this site and include picrotoxin, TBPS and related cage convulsants and pentelenetetrazol (reviewed in Ticku, 1991).

To further confirm that DHEAS interacted with the picrotoxin/TBPS binding site in a competitive manner, we performed the dissociation of TBPS from its binding sites with saturating concentrations of picrotoxin, DHEAS, pentobarbital and GABA. We have used this approach in the past and demonstrated for the first time that barbiturates inhibited picrotoxin/TBPS binding allosterically, whereas convulsants such as pentylenetetrazol and TBPS interacted with the picrotoxin site competitively (Maksay and Ticku, 1985). The dissociation profile of DHEAS was monophasic and similar to TBPS/picrotoxin, whereas, pentobarbital greatly accelerated the dissociation profile of TBPS in a biphasic manner. Similarly, GABA greatly accelerated the dissociation of TBPS. These results are similar to our previous observations that barbiturates bind to a distinct site on the GABA receptor complex (Maksay and Ticku, 1985). The dissociation studies and the saturation analysis studies suggest that DHEAS interacts with the TBPS site in a competitive manner. Finally, we confirmed that DHEAS is a GABA_A receptor antagonist, because it inhibited GABA-induced [³⁶Cl]-influx in cultured cortical neurons. The IC₅₀ value of DHEAS (10 µM) to inhibit GABA-induced [³⁶Cl]-influx is similar to its IC₅₀ value (13 µM) for inhibiting electrophysiological responses of GABA (Majewska et al., 1990a, 1990b; Spivak, 1994). The reason for differences in the IC₅₀ values of DHEAS to inhibit [³⁶Cl] influx and GABA and TBPS binding is not clear, but could be due to different tissue preparations, (fetal vs. adult tissue) and/or assay conditions. The [³⁶Cl] influx assay was carried out under conditions close to physiological conditions and utilized fetal tissue, whereas binding studies were conducted in membrane homogenates of adult tissue. The possible contribution of subunit composition of GABA receptors in these preparations may also have contributed to the differences in the IC₅₀ values.

Although our studies clearly demonstrate that the picrotoxin/TBPS site on the GABA receptor complex is the site of action for DHEAS, they are in contrast with reports in literature that suggested a lack of interaction of DHEAS with the TBPS binding site, and its ability to antagonize pentobarbital enhancement of flunitrazepam binding (Majewska and Schwartz, 1987; Majewska et al., 1990a, 1990b). These authors also suggested that DHEAS acted at the barbiturate site. Our dissociation studies of TBPS binding clearly rule out that possibility.

[³H]DHEAS has also been reported to bind to two distinct sites in brain membranes (Demirgoren et al., 1991). In this study, [³H]DHEAS B_max values for the two sites were reported to be 88.4 ± 13.8 pmol/mg protein and 10.2 nmol/mg protein, respectively. These B_max values are extremely high and unlikely to be related to the GABA_A receptors, despite such a claim made in the manuscript (Demirgoren et al., 1991). [³H]DHEAS binding was inhibited by pentobarbital (IC₅₀ = 3.46 mM), but not by GABA agonists, picrotoxin or TBPS (Demirgoren et al., 1991). These [³H]DHEAS binding sites are apparently distinct and not associated with the GABA_A receptors, and it is not apparent what these binding sites represent.

In summary, DHEAS, but not DHEA, competitively inhibited the binding of [³⁵S]TBPS to rat brain membranes, and it dissociated [³⁵S]TBPS binding from its binding sites with a profile similar to other convulsants and noncompetitive antagonists of GABA receptors such as picrotoxin, TBPS and pentylenetetrazol (Maksay and Ticku, 1985; this study). The dissociation profile is also different from barbiturates. Taken together, these results suggest that DHEAS binds to the picrotoxin/TBPS binding site of the GABA_A receptor complex. The observed noncompetitive inhibition of DHEAS of GABA responses may be mediated by binding to this site.