Introduction

Top Abstract Introduction Methods Results Discussion References

The existence of a neurosteroid recognition site on the GABA_A receptor complex has been demonstrated (Baulieu, 1991; Majewska, 1992; Lambert et al., 1995). Neurosteroids such as allopregnanolone and 5-THDOC (allotetrahydrodeoxycorticosterone) are synthesized in brain but do not interact with classical intracellular steroid receptors, rather they bind stereoselectively and with high affinity to the membrane-bound GABA_A receptor complex (Paul and Purdy, 1992). Neurosteroids may serve an allosteric modulatory role at the GABA_A receptor complex (Lambert et al., 1995). In vitro, neurosteroids are similar to barbiturates in enhancing BZ and GABA_A function and binding resulting in increased Cl uptake (Majewska et al., 1986; Harrison et al., 1987). At low concentrations, they enhance the ability of BZs to potentiate muscimol-stimulated Cl uptake (Morrow et al., 1990) and displace the binding of the convulsant t-butylbicyclophosphorothionate (Orchinik and McEwen, 1993; Lambert et al., 1995), whereas at high concentrations, neurosteroids directly activate the GABA_A receptor in the absence of GABA (Morrow et al., 1990). In vivo, neurosteroids are anticonvulsant (Belelli et al., 1989), myorelaxant and sedative-hypnotic (Ramsay et al., 1974; Bukusoglu et al., 1993), they share discriminative stimulus properties with BZ receptor agonists (Ator et al., 1993), and they produce effects similar to BZs and barbiturates in preclinical procedures sensitive to the effects of anxiolytic drugs (Crawley et al., 1986; Wieland et al., 1991). Recently, the neurosteroid allopregnanolone was found to decrease rat pup USV while also producing motor incoordination (Zimmerberg et al., 1994).

35 to 70 kHz USV are emitted by neonatal rodents that are separated from the dam and littermates (Gardner, 1985; Mos and Olivier, 1989; Winslow and Insel, 1991). These "distress vocalizations" or "isolation calls," potentiated in environmental conditions such as social isolation, reduced ambient temperature, hunger, rough handling, novelty and threat (Okon, 1970; Allin and Banks, 1971; Bell, 1979), are an effective stimulus for maternal behavior. The emission of distress vocalizations is not limited to rodents; they have been observed in neonatal birds (Panksepp et al., 1978), dogs (Panksepp et al., 1978) and primates (Kalin et al., 1987; Miczek et al., 1995). Because of the uniquely "stressful" contexts in which distress vocalizations are emitted, as well as their cross-species generality, these calls have provided an attractive measure in the neurobiology of anxiety and in the evaluation of anxiolytic compounds (Winslow and Insel, 1991).

Not surprisingly, diverse drugs with clinical anxiolytic effects suppress USV. GABA, acting on GABA_A receptors, remains the most investigated neurotransmitter system with respect to USV (Insel and Winslow, 1991). Agonists acting at BZ and GABA_A receptor sites are very effective in reducing USV emitted by rat and mouse pups. Specifically, chlordiazepoxide and diazepam decreased USV induced by rough handling in isolated rat pups in a flumazenil-reversible manner and at doses that did not alter locomotor behavior (Gardner, 1985; Gardner and Budhram, 1987). In addition, BZ receptor inverse agonists such as FG 7142, DMCM and -CCE, as well as pentylenetetrazole, increased the production of USV and antagonized the USV-suppressive effects of diazepam (Insel et al., 1986; Gardner and Budhram, 1987; Nastiti et al., 1991). The GABA receptor agonists muscimol and baclofen were also found to decrease USV (Gardner, 1985; Nastiti et al., 1991). These studies reveal that separation-induced USV are sensitive to the effects of compounds that influence GABAergic transmission through BZ, GABA_A and perhaps barbiturate receptor sites. The discovery that neurosteroids allosterically modulate GABAergic function provided the impetus for our investigation involving allopregnanolone and its behavioral effects including USV emission and sedation in rat pups.

	Methods

Top Abstract Introduction Methods Results Discussion References

Subjects. Seven-day-old (the date of parturition being designated as postnatal day 1) male and female Long-Evans rat pups weighing 6 to 18 g were bred on site from pairs (Charles River Laboratories, Wilmington, MA) in a vivarium with controlled temperature (21 ± 1°C), humidity (40-50%) and an inverted 12-hr light-dark cycle. All animals were housed in large 45.7 × 45.7 × 71.1 cm stainless steel breeding cages with pine shavings and free access to rodent food and water. Litters containing fewer than four pups were not tested, and each litter contributed not more than three pups to each treatment. All experiments were conducted during the dark phase of the light-dark cycle.

Apparatus. Ultrasonic vocalizations were detected using a condensor microphone (Bruel and Kjær model 4135) suspended 10 cm above a 22 × 22 × 4.5 cm aluminum pan. Signals were preamplified (Bruel and Kjær model 2633), filtered (Khron-Hite model 3550R) and passed through a measuring amplifier (Bruel & Kjær model 2610) which provided a flat frequency response between 20 to 60 kHz. The amplifier output was monitored with an oscilloscope (Goldstar model 9020A) and concurrently connected to a customized signal detection system. This MacIntosh II-based system digitized and frequency-filtered (band-pass: 32-52 kHz) the signal to determine the onset and offset of each sound pulse. Spurious signals (i.e., scratches on the aluminum pan and audible "squeals") were eliminated by using an additional algorithm that rejected signals less than 0.04 sec in duration or those that were separated by less than 0.06 sec. When correlating measurements from this automated sound detection system with taped playback of sound pulses, the reliability estimates exceeded r = 0.95.

Procedure. Individual litters were separated from their mothers and placed in a 20 × 18 × 12 cm polycarbonate cage containing shavings from the home cage. Pups were brought to the experimental room with a heat source that maintained body temperature at approximately 33°C. After a 30-min adaptation period, pups were weighed, numbered and evaluated for ultrasound production by placing them individually onto the aluminum pan marked with a 5 × 5 cm grid and maintained at 20 ± 1°C. 70% of the pups emitted USV within 2 min and they were subsequently assigned randomly to the appropriate treatment. After the appropriate injection-test interval had elapsed (see "Drugs" below), pups were individually placed onto the 20°C surface for a 2-min observation period during which the rate and duration of USV were recorded concurrently with motor activity by direct observation of grid crossings (operationally defined as any traversing of a grid by the pup's head and both front paws). In addition, body temperature was measured before drug administration and immediately before the vocalization test, inserting the probe 1.5 cm into the rectum of the pup until the temperature had stabilized (approximately 10 sec).

Drugs. Diazepam (Hoffman LaRoche, Nutley, NJ; 0.01, 0.03, 0.1, 0.3, 0.6, 1, 3, 6, 10 mg/kg suspended in a solution containing 85% distilled water, 14% propylene glycol and 1% Tween 80), muscimol (Sigma Chemical Co., St. Louis, MO; 0.003, 0.01, 0.03, 0.1, 0.3, 0.6, 1 mg/kg dissolved in 0.9% sodium chloride), and pentobarbital (Sigma; 0.3, 1, 3, 6, 10, 17, 30, 60 mg/kg dissolved in 0.9% sodium chloride) were administered 30 min before the test. Alprazolam (Upjohn, Kalamazoo, MI; 0.003, 0.01, 0.03, 0.1, 0.3, 0.6, 1, 3, 6 mg/kg suspended in a solution containing 85% distilled water 14% propylene glycol and 1% Tween 80) was administered 20 min before the test. Allopregnanolone (Sigma; 1, 3, 10, 17, 30 mg/kg dissolved in a 20% aqueous solution of 2-hydroxypropyl--cyclodextrin) was administered 10 min before the test.

Data analysis. The mean rate of grid crossings and changes in body temperature were transformed into percent of control; these transformed values and the mean rate and duration of USV were analyzed with individual one-factor (DRUG) between-subjects analysis of variances. Time course data were analyzed with a two-factor (DRUG, TIME) between-subjects analysis of variance. When comparisons were significant, posthoc Tukey t tests were performed. For analysis of the selectivity of the behavioral effects of the tested drugs, Pearson product moment correlation coefficients were calculated for USV rate, grid crossings and changes in body temperature. Alpha was 0.05, two-tailed. For analysis of the antagonism and interaction data, and after test for parallelism revealed no deviances from parallelism between dose-effect curves, ED₅₀s (i.e., the dose at which USV were suppressed to 50% of vehicle control) and 95% confidence intervals for USV rate were calculated from first order regression equations. Nonoverlapping confidence intervals were accepted as significant.

	Results

Top Abstract Introduction Methods Results Discussion References

USV were monotonic pulses, ranging from 35 to 45 kHz in frequency and 0.04 to 0.3 sec in duration. During the 2-min vehicle isolation tests, the rate of USV was 92.77 ± 1.53 (mean ± S.E.M.) calls/min, the total call duration was 10.88 ± 0.17 sec, the pup's locomotor behavior was 6.84 ± 0.38 grid crosses/min, and the pup's body temperature decreased by 1.46 ± 0.09°C (base-line body temperature: 32.57 ± 0.08°C). These observations remained consistent across different control vehicle treatments (distilled water, saline, propylene glycol and Tween 80, 20% cyclodextrin vehicles). Because the suppressive effects of the tested compounds on the duration of USV closely paralleled their effects on the rate of USV, only the effects on USV rate will be presented in detail. All agonist effects on the rate of USV are summarized in figure 1.

View larger version (K): [in this window] [in a new window]   Fig. 1.   The effects of selected benzodiazepine (alprazolam, filled circles; diazepam, filled squares), GABAA (muscimol, filled triangles), barbiturate (pentobarbital, filled inverted triangles) and neurosteroid (allopregnanolone, filled diamonds) agonists on ultrasonic vocalizations in rat pups. Control values: alprazolam and diazepam (polyethylene glycol and Tween 80, open circles); muscimol and pentobarbital (saline, open triangles); allopregnanolone (2-hydroxypropyl--cyclodextrin, open diamonds). Lines depict first order regression equations, asterisks denote significant differences (P < .05) from control, and error bars denote 1 S.E.M.

Allopregnanolone. Allopregnanolone dose-dependently decreased USV [10-30 mg/kg; F(5,60) = 19.66, P < .05), locomotor activity [3-30 mg/kg; F(5,60) = 10.22, P < .05], and body temperature [30 mg/kg; F(5,60) = 3.04, P < .05; fig. 2A]. Pretreatment with flumazenil (0.1 mg/kg), bicuculline (2 mg/kg) or picrotoxin (1 mg/kg) failed to alter the suppressive effects of allopregnanolone on USV (table 1) and locomotor activity (table 2). Pretreatment with these BZ-GABA_A receptor complex antagonists prevented the decrease in body temperature due to allopregnanolone (table 3), although in general, drug effects on body temperature were quite variable.

View larger version (K): [in this window] [in a new window]   Fig. 2.   A, The effects of allopregnanolone (open diamonds) and antagonists acting at the benzodiazepine (flumazenil, filled circles), GABAA (bicuculline, filled triangles) and barbiturate (picrotoxin, filled inverted triangles) receptor sites on ultrasonic vocalizations in rat pups. B, The time course of the suppressive effects of allopregnanolone (3 mg/kg, filled diamonds) on ultrasonic vocalizations in rat pups. Lines depict first order regression equations (A), asterisks denote significant differences (P < .05) from control, and error bars denote 1 S.E.M.

Alprazolam and diazepam. Alprazolam [0.1-1 mg/kg; F(6,99) = 8.01, P < .05] and diazepam [0.3-3 mg/kg; F(5,60) = 5.41, P < .05; fig. 3] dose-dependently decreased USV. Alprazolam increased locomotor behavior at 0.03 mg/kg [F(6,99) = 4.55, P < .05]. At the doses tested, diazepam produced quite variable effects on locomotor behavior (N.S.), and did not alter body temperature. Alprazolam's effects on body temperature were not measured.

View larger version (K): [in this window] [in a new window]   Fig. 3.   A, The effects of alprazolam alone (open circles), in conjunction with flumazenil (0.1 mg/kg, filled circles), or in conjunction with allopregnanolone (3 mg/kg, filled circles with crosses) on ultrasonic vocalizations in rat pups. B, The effects of diazepam alone (open squares), in conjunction with flumazenil (0.1 mg/kg, filled squares), or in conjunction with allopregnanolone (3 mg/kg, filled squares with crosses) on ultrasonic vocalizations in rat pups. Lines depict first order regression equations, asterisks denote significant differences (P < .05) from control, and error bars denote 1 S.E.M.

Muscimol. Muscimol dose-dependently decreased USV [0.1-0.6 mg/kg; F(4,49) = 14.65, P < .05; fig. 4A]. Muscimol also dose-dependently decreased locomotor behavior [0.3-0.6 mg/kg; F(4,49) = 6.22, P < .05]. Body temperature was unaffected by muscimol up to the highest dose tested.

View larger version (K): [in this window] [in a new window]   Fig. 4.   A, The effects of muscimol alone (open triangles), in conjunction with bicuculline (2 mg/kg, filled triangles), or in conjunction with allopregnanolone (3 mg/kg, filled triangles with crosses) on ultrasonic vocalizations in rat pups. B, The effects of pentobarbital alone (open inverted triangles), in conjunction with picrotoxin (1 mg/kg, filled inverted triangles), or in conjunction with allopregnanolone (3 mg/kg, filled inverted triangles with crosses) on ultrasonic vocalizations in rat pups. Lines depict first order regression equations, asterisks denote significant differences (P < .05) from control, and error bars denote 1 S.E.M.

Pentobarbital. Pentobarbital produced biphasic effects of USV (fig. 4B) and locomotor behavior and dose-dependently decreased body temperature. Pentobarbital (1 mg/kg) tended to increase USV and locomotor behavior, although higher doses monotonically decreased USV [10-30 mg/kg; F(6,64) = 17.00, P <.05], locomotor behavior [30 mg/kg; F(6,64) = 10.06, P < .05], and body temperature [30 mg/kg; F(6,64) = 7.67, P < .05].

	Discussion

Top Abstract Introduction Methods Results Discussion References

Rat pup USV were sensitive to clinically effective anxiolytics and compounds acting at various sites on the GABA_A receptor complex. These effects were mediated by receptors at distinct sites of the GABA_A receptor complex because flumazenil, which acts at the BZ site, antagonized the effects of alprazolam and diazepam, and bicuculline, which acts at the GABA_A site, antagonized the effects of muscimol. Furthermore, the increased potencies to suppress USV demonstrated with alprazolam, diazepam and pentobarbital in the presence of allopregnanolone may involve the proposed positive allosteric modulation by neurosteroids at the GABA_A receptor complex (Orchinik and McEwen, 1993; Lambert et al., 1995). Finally, the suppression of USV by allopregnanolone was not influenced by the antagonists flumazenil, bicuculline or picrotoxin; these results are consistent with the view that neurosteroids act at a site distinct from BZ, GABA_A and barbiturate receptor sites on the GABA_A receptor complex (Gee et al., 1995).

Allopregnanolone has been shown previously to possess sedative-hypnotic effects (Crawley et al., 1986; Bitran et al., 1991; Orchinik and McEwen, 1993). In our experiments, allopregnanolone produced its USV-suppressive, myorelaxant and hypothermic effects in the same dose range; it is therefore difficult to dissociate its behavioral effects that might also influence USV. Although the current knowledge of the behavioral effects and site of action of neurosteroids is incomplete (Purdy et al., 1990; Mellon, 1994), the USV-suppressive and sedative activity observed in our experiments parallel the anxiolytic-like and sedative-hypnotic effects demonstrated with preclinical methods sensitive to anxiolytic compounds (Britton et al., 1991; Bitran et al., 1993; Zimmerberg et al., 1994; Picazo and Fernandez-Guasti, 1995).

Antagonists acting at BZ, GABA_A and convulsant sites were ineffective in altering allopregnanolone's USV-suppressive effects, and these effects are indicative of separate mechanisms or sites of action for BZs and neurosteroids (Morrow et al., 1990; Lambert et al., 1995). BZ antagonists were previously ineffective in altering the in vitro and in vivo effects of neurosteroids. Flumazenil failed to inhibit the potentiation of GABA-elicited membrane currents and depolarizing responses by alphaxolone (Cottrell et al., 1987) and similarly failed to inhibit the potentiation of muscimol-stimulated Cl uptake by THDOC (Morrow et al., 1990). CGS 8216 failed to antagonize allopregnanolone's anxiolytic-like activity in a light/dark transition test (Wieland et al., 1991), and flumazenil only partially blocked the alphaxolone-induced increases in punished behavior (Britton et al., 1991).

The interactions of neurosteroids with the GABA antagonist bicuculline are more complex. In vitro, bicuculline inhibited the displacement of TBPS and reversed the membrane current elicited by high concentrations of alphaxolone (Cottrell et al., 1987; Olsen and Sapp, 1995); in contrast, bicuculline-insensitive augmentation of TBPS binding has also been demonstrated with lower concentrations of alphaxolone. The results from in vivo investigations are equivocal. Whereas bicuculline and the convulsant picrotoxin reduced the anxiolytic-like effects of 3-hydroxy-4-pregnan-20-one in response to a predator odor (Kavaliers et al., 1994), these GABA_A receptor complex antagonists failed to suppress the alphaxolone-induced increases in punished behavior (Britton et al., 1991). These latter results and the failure of bicuculline and picrotoxin to alter allopregnanolone's USV-suppressive effects in our experiments are consistent with separate sites of action for neurosteroids, GABA_A and convulsants on the GABA_A receptor complex, and should prompt future in vivo investigations of this receptor complex using selective neurosteroid antagonists. It must also be noted that the results from our set of experiments involving picrotoxin are less compelling, as this antagonist failed to alter pentobarbital's and allopregnanolone's USV-suppressive effects.

The BZ receptor full agonists alprazolam and diazepam suppressed USV in a behaviorally specific manner, i.e., the suppression of USV was not secondary to the sedative or myorelaxant properties of these compounds (Gardner, 1985). The reversal of alprazolam's and diazepam's suppressive effects on USV by flumazenil suggests that central BZ receptors were the site of action for this effect. Of particular interest was the interaction between allopregnanolone and the BZ agonists. Allopregnanolone enhanced alprazolam's and diazepam's USV-suppressive effects and this increase in potency was not secondary to allopregnanolone's sedative or hypothermic properties, as neither the rate of grid crossings nor body temperature were altered. In vitro, neurosteroids have been found to enhance GABA_A receptor function at a site distinct from the BZ receptor (Macdonald and Olsen, 1994). At low concentrations, neurosteroids stimulated [³H]flunitrazepam binding, increased the duration of the Cl channel opening and increased the likelihood of longer open-channel burst durations (Harrison et al., 1987); these effects are further augmented by BZs, revealing a separate site of action for each (Gee et al., 1988; Morrow et al., 1990). In our experiments, the increased potency of the BZ agonists to suppress USV is consistent with the proposal that neurosteroids serve as positive allosteric modulators at the GABA_A receptor complex; it is also true that this leftward shift in BZ function may reflect an additive relationship.

The GABA_A receptor agonist muscimol suppressed USV and locomotor behavior, and similar to allopregnanolone, suggests that the reduction in USV by muscimol might have been secondary to its sedative effects (Nastiti et al., 1991). Pretreatment with the GABA_A receptor antagonist bicuculline reversed muscimol's suppressive effects on USV and locomotor activity and confirms the involvement of central GABA_A receptors in the USV-suppressive, muscle relaxant or sedative effects of muscimol (Gardner, 1985; Shephard, 1987; Benton and Nastiti, 1988; Nastiti et al., 1991). The results from the interaction experiments with allopregnanolone and muscimol were unexpected. In vitro, muscimol-stimulated Cl uptake was robustly enhanced with a variety of neurosteroids, including allopregnanolone, alphaxolone, THDOC, THDOC 21-mesylate, and 3-hydroxy-pregn-4-en-20-one; neurosteroids increased muscimol binding to GABA_A receptor sites (Turner et al., 1989; Morrow et al., 1990). In contrast, allopregnanolone failed to potentiate the USV-suppressive effects of muscimol in our set of experiments and should prompt future in vivo investigations of the interaction between neurosteroid and GABA_A receptor agonists. One explanation for the discrepancy between the in vitro and current in vivo results may be in the use of a relatively immature system. Although present and functional at birth, BZ, GABA_A and convulsant recognition sites increase in number and change in their central distribution up to 3 to 4 wk postnatally (Aldinio et al., 1980; Palacios and Kuhar, 1981). Additionally, it has been discovered that the subunit composition of the GABA_A receptor complex changes early in postnatal development (Gambarana et al., 1990; Frostholm et al., 1992).

Neurosteroids have been most often compared to barbiturates on the basis of their in vitro and in vivo activity. Both of these allosteric modulators of the GABA_A receptor complex increase GABA-stimulated Cl flux by increasing the duration of channel opening (Allan and Harris, 1986), they stimulate muscimol binding, and at higher concentrations they directly elicit a membrane current from chromaffin cells (Peters et al., 1988; Turner et al., 1989). Additionally, barbiturates, as with neurosteroids, have anticonvulsant properties and produce anxiolytic-like effects in preclinical methods sensitive to anxiolytics (Meert and Colpaert, 1986). Neurosteroid augmentation of barbiturate effects is well documented (Turner et al., 1989; Paul and Purdy, 1992; Macdonald and Olsen, 1994). Similar to the results from the BZ-neurosteroid experiments (described above), the presently observed enhancement of the USV-suppressive effects of pentobarbital by allopregnanolone is consistent with the proposal that neurosteroids serve as positive allosteric modulators at the GABA_A receptor complex; this enhancement is also consistent with an additive relationship.

In summary, rat pup USV were sensitive to BZ, GABAergic, barbiturate and neurosteroid compounds acting at the GABA_A receptor complex. The USV-suppressive effects of alprazolam, diazepam and pentobarbital were augmented in the presence of allopregnanolone and supports the proposal that neurosteroids serve as positive allosteric modulators at the GABA_A receptor complex. These results may also indicate an additive relationship between these compounds and should prompt further in vivo research involving multiple combinations of doses (i.e., alprazolam and allopregnanolone) and the use of a selective neurosteroid antagonist to more fully characterize allopregnanolone's effects at the GABA_A receptor complex. Finally, allopregnanolone's USV-suppressive effects were not altered by the GABA_A receptor complex antagonists flumazenil, bicuculline and picrotoxin; these results are indicative of a separate neurosteroid site of action on this receptor complex.