Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

GABA_A receptor is the main type of inhibitory receptor in the brain and is a member of the superfamily of ligand-gated ion channels that includes the strychnine-sensitive glycine receptor, the 5-hydroxytryptamine₃ subtype of the serotonin receptor, and the nicotinic acetylcholine receptor (Betz, 1990). The binding of GABA to GABA_A receptors induces the opening of an intrinsic Cl channel with consequent hyperpolarization of the cell. The subunit composition of the pentameric GABA_A receptors determines their specific physiological and pharmacological properties (Rudolph et al., 2001). The GABA_A receptor is a prominent target of certain neuroactive steroids, which act as potent endogenous allosteric (positive or negative) modulators of receptor activity (Park-Chung et al., 1999). Such compounds are also thought to be of potential therapeutic benefit. Indeed, neuroactive steroids have been used as intravenous anesthetics and have been shown to exert anxiolytic, hypnotic, anticonvulsant, and antiepileptic effects in animals or humans (Gasior et al., 1999).

Ganaxolone (3-hydroxy-3-methyl-5-pregnan-20-one) is a synthetic compound that is structurally related to the endogenous neurosteroid allopregnanolone (3-hydroxy-5-pregnan-20-one) (Gasior et al., 1999). Like its endogenous analog, ganaxolone is a potent anticonvulsive and antiepileptic agent; however, it is more stable than allopregnanolone as a result of its -methyl group, which prevents its metabolism and oxidation of the 3-hydroxy moiety (Carter et al., 1997). Ganaxolone, also like allopregnanolone, enhances GABA_A receptor function. It thus inhibits the binding of t-[³⁵S]butylbicyclophosphorothionate and promotes the binding of [³H]flunitrazepam and [³H]muscimol to GABA_A receptors present in brain membranes. Electrophysiological studies with recombinant GABA_A receptors have also shown that ganaxolone potentiates receptor function to the same extent as does allopregnanolone (Carter et al., 1997). In addition, ganaxolone is more potent than is valproate, diazepam, or phenobarbital in blocking the development of seizures induced by repeated administration of pentylenetetrazol in mice (Beekman et al., 1998). Moreover, long-term treatment with ganaxolone in rats does not appear to induce tolerance to its anticonvulsant activity (Reddy and Rogawski, 2000).

In addition to potentiating GABA_A receptor function, allopregnanolone modulates the expression of genes encoding various GABA_A receptor subunits. Indeed, physiological or pharmacological exposure of GABA_A receptors to allopregnanolone or other neuroactive steroids induces changes in receptor subunit composition that are associated with modification of receptor function (Concas et al., 1998; Follesa et al., 1998, 2000). Similar effects have been observed after long-term exposure to or abrupt discontinuation of drugs that target GABA_A receptors (Follesa et al., 2001, 2002; Papadeas et al., 2001). Long-term exposure of cerebellar granule cells to progesterone, which is converted to allopregnanolone by the enzyme 5-reductase, results in a decrease in the abundance of 1, 3, 5, and 2 subunit mRNAs, whereas withdrawal of this steroid induces an increase in the abundance of the 4 subunit mRNA (Follesa et al., 2000). Up-regulation of the expression of the 4 subunit, which is also induced by withdrawal of receptor modulators such as ethanol (Devaud et al., 1997) and benzodiazepines (Follesa et al., 2002), has been consistently associated with an increase in neuronal excitability (Smith et al., 1998).

We have now investigated the effects of long-term exposure to and subsequent withdrawal of ganaxolone on the abundance of GABA_A receptor subunit mRNAs in cultured cerebellar granule cells. In addition, we compared the effects of this drug on the function of recombinant 122L and 222L GABA_A receptors with those of the natural steroid allopregnanolone.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Microinjection of cDNAs into Xenopus Oocytes and Electrophysiological Recording. Xenopus laevis oocytes (stage V to VI) were isolated as described previously (Lin et al., 1992). The cDNAs for the 1, 2, 2, and 2L subunits of the human GABA_A receptor were subcloned into the pCDM8 vector. A mixture of plasmids encoding the 1, 2, and 2L. or the 2, 2, and 2L receptor subunits (total of 1.5 ng of cDNA in 30 nl in a 1:1:1 ratio) was injected into the nucleus of oocytes, as described (Colman, 1984). Electrophysiological measurements were performed with oocytes 1 to 4 days after injection. Oocytes expressing 122L or 222L receptors were placed in a chamber (capacity, ~100 µl) and perfused (2 ml/min) with modified Barth's solution (MBS), consisting of 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO₃, 10 mM Hepes-NaOH (pH 7.5), 0.82 mM MgSO₄, 0.33 mM Ca(NO₃)₂, and 0.91 mM CaCl₂, with the use of a roller pump (Cole-Parmer Instruments, Chicago, IL). The animal pole of each oocyte was impaled with two glass electrodes (0.5 to 10 M) filled with 3 M KCl, and the cells were subjected to a voltage clamp at 70 mV (oocyte clamp OC-725C; Warner Instruments, Hamden, CT). Oocytes were exposed for 30 s to GABA dissolved in MBS. Ganaxolone (kindly provided by R. H. Purdy) and allopregnanolone (Sigma-Aldrich, St. Louis, MO) were first dissolved in dimethyl sulfoxide (DMSO) and then diluted in MBS; the final DMSO concentration to which oocytes were exposed was 1% and did not affect the response to GABA. Ganaxolone and allopregnanolone were each applied for 60 s alone before coapplication with GABA for 30 s. Oocytes were exposed to MBS alone for 5 or 20 min between applications of drugs at concentrations of 0.3 or >0.3 µM, respectively.

Cell Culture. Primary cultures of cerebellar neurons enriched in granule cells were prepared from the cerebellum of 8-day-old rats, as described by Follesa et al. (2000). After culture for 8 days, these cells express functional GABA_A receptors (Bovolin et al., 1992; Follesa et al., 2000) with a subunit composition similar to that apparent for the cerebellum during postnatal development, but different from that for the cerebellum, of adult rats (Laurie et al., 1992). Cells were plated (2.5 × 10⁶ cells/2 ml) in 100-mm dishes that had been coated with poly-L-lysine (10 µg/ml) (Sigma-Aldrich) and were cultured in basal Eagle's medium (Invitrogen, Carlsbad, CA) supplemented with 10% heat-inactivated fetal bovine serum (Invitrogen), 2 mM glutamine, gentamicin (100 µg/ml), antibiotic-antimycotic solution (10 ml/l) (Sigma-Aldrich), and 25 mM KCl; such a high concentration of K⁺ was necessary to induce persistent depolarization, which promotes the survival of granule cells. Cytosine arabinofuranoside (final concentration, 10 µM) (Sigma-Aldrich) was added to cultures 18 to 24 h after plating to inhibit the proliferation of non-neuronal cells. Cells were maintained in culture for a total of 8 days, and long-term treatment with ganaxolone was therefore initiated accordingly. Ganaxolone was dissolved in DMSO and diluted sequentially in culture medium to a final concentration of 1 µM; control cells were treated with solvent alone at the same dilution (0.1%) as that experienced by the drug-treated cells. The culture medium was replaced every day with fresh medium containing the appropriate addition. Animals were treated in accordance with the Declaration of Helsinki and the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the U.S. National Institutes of Health.

Probe Preparation. Total RNA was extracted from rat brain (Follesa et al., 1998) and subjected to reverse transcription with SuperScript reverse transcriptase (Invitrogen) in the presence of oligo(dT). The resulting cDNA (1 to 10 ng) was amplified by the polymerase chain reaction, as described by Follesa et al. (1998), with 2.5 U of TaqDNA polymerase (PerkinElmer Instruments, Norwalk, CT) in 100 µl of standard buffer [100 mM Tris-HCl (pH 8.3), 500 mM KCl, 15 mM MgCl₂, 0.01% gelatin] containing 1 µM each of specific sense and antisense primers and 200 µM of each deoxynucleoside triphosphate. The primer pairs for the various subunits of the GABA_A receptor were designed to include cDNA sequences with the lowest degree of homology among the different subunits (Follesa et al., 1998). The reaction was performed in a thermal cycler (Ericomp, San Diego, CA) for 30 cycles of 94°C for 45 s, 60°C for 1 min, and 72°C for 1 min, with a final extension at 72°C for 15 min (Follesa et al., 1998). The reaction products were separated by electrophoresis on a 1.8% low-melting point agarose gel, visualized by staining with ethidium bromide, excised from the gel, purified, and inserted into the pAMP 1 cloning vector (Invitrogen). Escherichia coli NM522 cells were transformed with the resulting plasmids, which were subsequently purified from the bacteria, and the cDNA inserts were sequenced with a Sequenase DNA sequencing kit (USB, Cleveland, OH). The determined nucleotide sequences were 100% identical to the respective previously published sequences (Follesa et al., 1998). Plasmids containing the cDNA fragments corresponding to the various GABA_A receptor subunits were linearized with restriction enzymes (Follesa et al., 1998) and used as templates, together with the appropriate RNA polymerase (SP6 or T7), to generate [-³²P]citosine triphosphate-labeled cRNA probes for RNase protection analysis.

RNase Protection Assay. RNase protection analysis was used as a sensitive technique for semiquantitative detection of mRNA (Zinn et al., 1983; Lee and Costlow, 1987) and was performed as described by Follesa et al. (1998). We determined the abundance of mRNAs encoding the 1, 2, 3, 4, and 5 subunits of the GABA_A receptor as well as of those corresponding to the two splice variants of the 2 subunit (2L and 2S). Total RNA was extracted from cultured cerebellar granule cells and quantitated by measurement of absorbance at 260 nm. In brief, 25 µg of total RNA was dissolved in 20 µl of hybridization solution containing 150,000 cpm of ³²P-labeled cRNA probe for a specific receptor subunit mRNA (specific activity, 6 × 10⁷ to 7 × 10⁷ cpm/µg) and 15,000 cpm of ³²P-labeled cyclophilin cRNA (specific activity, 1 × 10⁶ cpm/µg). Cyclophilin is expressed widely among tissues, including the brain, and its gene is most likely regulated in an "on or off" manner (Milner and Sutcliffe, 1983; Danielson et al., 1988); cyclophilin mRNA was thus used as an internal standard for our measurements. The hybridization reaction mixture was incubated overnight at 50°C and then subjected to digestion with RNase, after which the remaining RNA-RNA hybrids were detected by electrophoresis (on a sequencing gel containing 5% polyacrylamide and urea) and autoradiography. The amounts of GABA_A receptor subunit and cyclophilin mRNAs were determined by scanning of the corresponding bands on the autoradiogram with a densitometer (model GS-700; Bio-Rad, Hercules, CA); this instrument was calibrated to detect saturated values, which were automatically excluded, so that the intensity of all bands measured was in the linear range. Data were normalized by dividing the optical density of the protected fragment for each receptor subunit mRNA by that of the protected fragment for cyclophilin mRNA. The amount of receptor subunit mRNA was therefore expressed in arbitrary units. Since the maximal effects were observed at 6 h after withdrawal we did not perform the time course study for the 2 and 5 subunits.

Statistical Analysis. Data are presented as means ± S.E.M. The statistical significance of differences was assessed by analysis of variance followed by Scheffe's test. A p value <0.05 was considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of Ganaxolone and Allopregnanolone on GABA_A Receptor Function. The effects of ganaxolone and allopregnanolone on GABA_A receptor function were compared by electrophysiological recording from recombinant human 122L (Fig. 1A) and 222L (Fig. 1B) receptors expressed in Xenopus oocytes. As previously described (Carter et al., 1997), both ganaxolone and allopregnanolone (0.01 to 10 µM) potentiated Cl currents induced by GABA at an EC_5-10 (concentration of GABA that induced a peak current with an amplitude of 5 to 10% of the maximal current observed with 1 mM GABA; it was determined for each oocyte and was ~5 to 8 µM). No significant difference was apparent between the actions of the two drugs at either receptor subtype; the EC₅₀ values for ganaxolone and allopregnanolone were 0.5 ± 0.05 and 0.4 ± 0.07 µM at 122L receptors and 1.6 ± 0.3 and 0.9 ± 0.1 µM at 222L receptors, respectively. Data are the means ± S.E.M. of values obtained from three to five oocytes.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Concentration-dependent potentiation by ganaxolone and allopregnanolone of GABA-induced Cl currents in Xenopus oocytes expressing recombinant human 122L (A) or 222L (B) GABAA receptors and representative tracing of ganaxolone or allopregnanolone effects obtained from oocytes expressing 122L (C) or 222L (D). Ganaxolone or allopregnanolone was applied alone for 60 s before coapplication with an EC5-10 of GABA for 30 s. Data are means ± S.E.M. of values obtained from three to five oocytes and are expressed as the percentage of potentiation of the GABA-evoked response.

Effects of Long-Term Exposure to and Subsequent Withdrawal of Ganaxolone on the Abundance of GABA_A Receptor Subunit mRNAs. We next evaluated the effects of long-term exposure to ganaxolone on the abundance of GABA_A receptor subunit mRNAs by an approach similar to that previously used to demonstrate such effects of progesterone (Follesa et al., 2000). RNase protection analysis thus revealed that exposure of cultured cerebellar granule cells to ganaxolone (1 µM) for 5 days had no significant effect on the amounts of 1, 2, 3, 4, 5, 2L, and 2S subunit mRNAs (Fig. 2A). Data are the means ± S.E.M. of values obtained from three independent experiments. We then determined whether withdrawal of ganaxolone after long-term treatment affects receptor subunit mRNA abundance. After exposure of the cells to ganaxolone (1 µM) for 5 days, they were incubated in the absence of the drug for 6 h. Withdrawal of ganaxolone resulted in a marked increase in the amounts of the 2 (+132%) and 5 (+75%) mRNAs and a smaller (but significant) increase (+17%) in the amount of the 4 subunit mRNA (Fig. 2B). In contrast, ganaxolone withdrawal induced a decrease in the amounts of the 1 (18%), 2L (35%), and 2S (35%) mRNAs and had no effect on that of the 3 mRNA. Data are the means ± S.E.M. of values obtained from three independent experiments.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Effects of long-term treatment with (A) and subsequent withdrawal of (B) ganaxolone on the abundance of GABAA receptor subunit mRNAs in cultured cerebellar granule cells. The abundance of the indicated receptor subunit mRNAs was determined by RNase protection assay both after incubation of cells with 1 µM ganaxolone for 5 days and after subsequent incubation in the absence of drug for 6 h. Data are means ± S.E.M. of values from three independent experiments and are expressed as the percentage of change relative to control cultures not exposed to ganaxolone. , p < 0.01 versus control.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Time course of changes in the abundance of GABAA receptor 1 (A), 4 (B), 2L (C), and 2S (D) subunit mRNAs induced by ganaxolone withdrawal in cultured cerebellar granule cells. Cells were incubated for 5 days in the presence of 1 µM ganaxolone and then for the indicated times in its absence, after which the amounts of receptor subunit mRNAs were determined by RNase protection assay. Data are means ± S.E.M. of values from two independent experiments and are expressed as the percentage of change relative to control cultures not exposed to ganaxolone. , p < 0.05; , p < 0.01 versus control.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Consistent with previous electrophysiological data (Carter et al., 1997), we have shown that the synthetic neuroactive steroid ganaxolone potentiates GABA-evoked Cl currents with an efficacy and potency similar to those of allopregnanolone at recombinant human 122L and 222L GABA_A receptors. In spite of their similarities in structure and pharmacology, however, ganaxolone and allopregnanolone differ in their abilities to modulate the abundance of GABA_A receptor subunit mRNAs in cultured cerebellar granule cells. Thus, long-term exposure of these cells to ganaxolone had no significant effect on the amounts of 1, 2, 3, 4, 5, 2L, or 2S subunit mRNAs. We previously obtained similar results with the hypnotic compounds zaleplon and zolpidem (Follesa et al., 2002), both of which exhibit high selectivity for GABA_A receptors containing the 1 subunit (Follesa et al., 2001). Our electrophysiological results, in agreement with those previously described by others (Carter et al., 1997), indicate that, unlike zaleplon or zolpidem, ganaxolone is not selective for GABA_A receptors that contain the 1 subunit.

Long-term exposure of cerebellar granule cells to allopregnanolone results in a decrease in the amounts of the mRNAs encoding the 2L and 2S subunits of the GABA_A receptor (Follesa et al., 2000). A reduction in the amount of the 2L subunit mRNA in rat cerebral cortex and hippocampus has also been observed during pregnancy when the abundance of allopregnanolone in the brain is substantially increased (Concas et al., 1998). A decrease in the expression of  subunits may result in a down-regulation of GABA_A receptor function and, in particular, in a reduced efficacy of receptor modulators, such as benzodiazepines, that are active only at receptors containing a  subunit. Long-term exposure of cultured cortical neurons to allopregnanolone induces a significant decrease in the abundance of 2, 3, 2, and 3 subunit mRNAs (Yu et al., 1996). The modifications in the GABA_A receptor gene expression induced by chronic exposure to allopregnanolone are consistent with the evidence that this steroid gives tolerance to its anticonvulsant activity (Czlonkowska et al., 2001). Given that chronic treatment of cerebellar granule cells with ganaxolone had no effect on the amounts of any of the  or  subunit mRNAs examined, such treatment would not be expected to affect GABA_A receptor sensitivity to positive modulators or to lead to the development of tolerance. This conclusion is consistent with the observation that long-term administration of ganaxolone, contrary to allopregnenolone (Czlonkowska et al., 2001), does not induce tolerance to its anticonvulsant activity (Reddy and Rogawski, 2000). Ganaxolone, however, does induce cross-tolerance to the anticonvulsant effect of diazepam (Reddy and Rogawski, 2000), an effect that cannot be explained by our present results but may possibly be due to posttranslational modification of receptor subunits. Chronic treatment of mice with zaleplon, zolpidem, or imidazenil, none of which modifies GABA_A receptor subunit gene expression after long-term treatment in cultured cerebellar granule cells (Follesa et al., 2002), also does not induce tolerance (Ghiani et al., 1994; Sanger et al., 1996). It is quite surprising the finding that ganaxolone does not alter the GABA_A receptor gene expression upon long-term exposure of cultured cerebellar granule cells in spite of being very similar, for chemical and electrophysiological properties, to the endogenous neurosteroid allopregnenolone. We cannot explain this evidence with the present results. Given that allopregnanolone is chemically less stable than ganaxolone, we believe that the effects on the GABA_A receptor gene expression observed during allopregnenolone long-term treatment of cerebellar granule cells are due to oxidation of allopregnanolone into 5-dihydropyridine, which can interact with the cytosolic progesterone receptor and regulate through a genomic action the GABA_A receptor gene expression (Rupprecht et al., 1993). This hypothesis, however, is rather improbable because progesterone by itself is not capable of modifying the GABA_A gene expression (Follesa et al., 2000). Indeed, our results may explain the differences between these two steroids in the development of tolerance to their anticonvulsant activities (Reddy and Rogawski, 2000; Czlonkowska et al., 2001).

We did not examine the possible effects of long-term exposure of cerebellar granule cells to ganaxolone on the abundance of GABA_A receptor  subunit mRNAs. Both  and  subunits are required for the formation of functional GABA_A receptors; they contribute to the binding site for GABA and other agonists (Boileau et al., 2002) and might also be important for neuroactive steroid action (Rick et al., 1998). Abrupt discontinuation of long-term exposure to ganaxolone resulted in an increase in the abundance of 2, 4, and 5 subunit mRNAs and a decrease in that of 1, 2L, and 2S mRNAs in cultured cerebellar granule cells. The changes in the amounts of the 1, 4, 2L, and 2S mRNAs were reversible, with the amounts having returned to control levels by 24 h after drug withdrawal.

Withdrawal of progesterone also induces a decrease in the amounts of 1 and 2L subunit mRNAs in cultured cerebellar granule cells (Follesa et al., 2000). In addition, withdrawal of agonists of the benzodiazepine binding site, such as zaleplon, zolpidem, imidazenil, and diazepam, also results in a decrease in the abundance of 1, 2L, and 2S subunit mRNAs in these cells (Follesa et al., 2001). Electrophysiological studies have revealed that down-regulation of the expression of these subunits generally results in a reduced functional response of GABA_A receptors to ligands of the benzodiazepine binding site (Follesa et al., 2000). Ganaxolone withdrawal might therefore be expected to result in a reduction in the sensitivity of GABA_A receptors to benzodiazepines and to other agonists of the benzodiazepine binding site. This prediction is further supported by our observation that ganaxolone withdrawal induced a small but significant increase in the amount of the 4 subunit mRNA, given that recombinant GABA_A receptors containing the 4 subunit are insensitive to benzodiazepines (Mohler et al., 2002). Nevertheless, our data also show that ganaxolone withdrawal increases the amounts of 2 and 5 subunit mRNAs, and receptors containing these subunits, like those containing 1, are sensitive to benzodiazepines (Mohler et al., 2002). Moreover, the 2 subunit, which is not widely expressed in the brain, is increased in abundance in brain regions in which the 1 subunit is absent or present at low levels (Mohler et al., 2002). The increase in the amount of the 2 subunit mRNA induced by ganaxolone withdrawal might thus reflect a mechanism for compensation of the down-regulation of the amount of the 1 subunit mRNA. Given that GABA_A receptors containing the 1 subunit are alone responsible for the sedative effect and contribute to the anticonvulsant effect of diazepam, whereas those containing the 2 subunit mediate the anxiolytic and muscle relaxant actions of this drug (Mohler et al., 2002), our in vitro data suggest that, during ganaxolone withdrawal, the sedative and anticonvulsant effects, but not the anxiolytic or muscle relaxant actions, of benzodiazepines might be impaired.

An increase in the expression of the 4 subunit after discontinuation of long-term treatment with positive modulators of the GABA_A receptor has been associated with the development of withdrawal effects such as anxiety, seizure susceptibility, and behavioral hyperexcitability (Smith et al., 1998). Such up-regulation of the 4 subunit might thus be important for the development of drug dependence. The increase in the abundance of the 4 subunit mRNA induced by ganaxolone withdrawal (+17%) in the present study is relatively small compared with those previously observed in cultured cerebellar granule cells after discontinuation of allopregnanolone treatment (+145%) (P.F., unpublished data) in rat brain after progesterone withdrawal (Smith et al., 1998) or in cultured cerebellar granule cells after withdrawal of full or partial agonists of the benzodiazepine binding site (Follesa et al., 2001). Discontinuation of long-term treatment with ganaxolone might thus be expected to induce withdrawal effects in animals that are markedly less pronounced than are those elicited by discontinuation of such treatment with other positive modulators of GABA_A receptor function.

In conclusion, the lack of effect of long-term exposure of cultured cerebellar granule cells to ganaxolone on GABA_A receptor subunit gene expression may explain the failure of this drug to induce tolerance to its anticonvulsant action in rats (Reddy and Rogawski, 2000). Our data further support the notion that ganaxolone might prove to be effective for the long-term treatment of epilepsy. Moreover, the changes in GABA_A receptor gene expression induced by ganaxolone withdrawal suggest that discontinuation of chronic treatment with this drug may result in a withdrawal syndrome, the severity of which remains to be established with behavioral studies in appropriate animal models.