Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Zolpidem is an imidazopyridine ligand that binds to benzodiazepine (BZ) recognition sites and is used for treatment of sleep disorders (for reviews, see Rush, 1998; Doble, 1999). The receptor binding profile of zolpidem is distinct from that of typical BZs in that it displays highest affinity at -aminobutyric acid A receptors expressing 1 subunits (BZ/1 receptors) but low-to-moderate affinity at receptors expressing 2, 3, and 5 subunits (Pritchett and Seeburg, 1990; Hadingham et al., 1993; Cox et al., 1995). In contrast, typical BZs (e.g., diazepam, triazolam) interact with all of the BZ sites with nearly equal affinities (for review, see Lüddens et al., 1995). The behavioral effects produced by zolpidem also appear to differ from those of typical BZs. In particular, zolpidem is relatively ineffective in tests predictive of anxiolytic activity (for review, see Sanger et al., 1994), raising the possibility that action at BZ/1 receptors does not contribute to the anxiolytic activity of BZ ligands (cf. Rudolph et al., 1999; Löw et al., 2000; McKernan et al., 2000).

In contrast to the findings with zolpidem, investigations using the relatively selective BZ/1 antagonist -carboline-3-carboxylate-t-butyl ester (-CCt) have implicated a role for the BZ/1 site in mediating anxiolytic effects (Shannon et al., 1984; Belzung et al., 2000; Huang et al., 2000). For example, -CCt antagonized the effects of diazepam in several tests of anxiolytic activity (Shannon et al., 1984; Griebel et al., 1999). The reason for this discrepancy between a BZ/1-preferring agonist and antagonist is not clear, although -CCt binding to 2-, 3-, or 5-containing receptors at high doses remains a possibility. An alternative explanation has been suggested based on other behavioral effects of zolpidem (Griebel et al., 1999). In this regard, zolpidem and other BZ/1 agonists engender marked sedation and impairment of motor activity (for review, see Sanger et al., 1994). This observation has led to the hypothesis that anxiolytic-like effects of BZ/1-selective agonists are masked, or counteracted, by effects on motoric function (cf. Griebel et al., 1999).

Assessment of whether BZ/1-selective agonists produce anxiolytic effects may be facilitated by the use of procedures in which anxiolytic-related behavior can be separated from motor behavior. One such approach uses "distress" or "maternal separation" vocalizations emitted by neonatal animals when placed in contexts associated with stressful events (Winslow and Insel, 1991; Miczek et al., 1995). For example, neonatal rodents emit ultrasonic vocalizations (USVs) under conditions such as maternal separation, reduced ambient temperature, hunger, and rough handling (Okon, 1970; for review, see Miczek et al., 1995). USVs are suppressed by both BZs and anxiolytic 5-hydroxytryptamine (5-HT) agonists (Miczek et al., 1995; Olivier et al., 1998a,b; Fish et al., 2000). Suppression of USVs by anxiolytics can be dissociated from motoric impairment; for example, USV suppression by BZs and 5-HT agonists often occurs in the absence of alterations in motor activity (Olivier et al., 1998a,b; Fish et al., 2000). Accordingly, suppression of USVs in neonatal rodents may prove useful in discerning the role of BZ receptor subtypes in mediating the anxiolytic-like effects of BZ agonists. In support of this idea, several studies have shown suppression of USVs in rat pups by BZ ligands (Insel et al., 1986; Gardner and Budhram, 1987; Vivian et al., 1997; Olivier et al., 1998a). In particular, Olivier et al. (1998a) demonstrated a clear suppression of USVs in rat pups following zolpidem administration that was not accompanied by impairment of motor activity, consistent with the hypothesis that BZ/1-selective agonists may possess anxiolytic-like effects under certain conditions.

As with rats, neonatal mice emit USVs during maternal separation and cold ambient temperatures, and these USVs are suppressed by BZ agonist administration (Benton and Nastiti, 1988, 1991; Fish et al., 2000). The present study examined the ability of the BZ/1 agonist zolpidem to suppress USVs in mouse pups, in comparison to the suppression engendered by the nonselective BZs diazepam and triazolam (Ducic et al., 1993; Hadingham et al., 1993). The role of BZ/1 receptors was examined by conducting antagonism studies using -CCt, in comparison with the nonselective BZ antagonist flumazenil. To evaluate the specificity of the effects of BZ ligands on USVs, the effects of these drugs on motor incoordination and thermoregulation were measured concurrently. The goal of these studies was to use USVs as a measure of anxiolytic activity not dependent on motor behavior to more clearly discern whether the BZ/1 receptor subtype plays a fundamental role in the anxiolytic effects of typical and atypical BZ agonists.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. Seven-day-old CFW mouse pups (N = 450) from litters that had 7 to 13 pups were used in the experiments. They were bred on site and housed with their parents (Charles River Breeding Labs, Wilmington, MA) with free access to rodent chow (Purina, St. Louis, MO) and tap water in clear polycarbonate cages (28 × 17 × 14 cm) and pine shavings as floor covering. The mice were maintained in a vivarium with controlled humidity and temperature (35-40%, 21 ± 1°C) and a 12-h light/dark cycle (lights on at 8:00 AM). All procedures were conducted with the approval and under the supervision of the Tufts University Institutional Animal Care and Use Committee. The animals were cared for according to the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, 1996).

Apparatus and Measurements. The mouse pups were tested in a separate procedure room. USVs were recorded in a sound-attenuated chamber (49.5 × 38 × 34 cm) fitted with a temperature-controlled platform (23 × 23 cm; 19 ± 1°C). The surface of the platform was divided by a 2- × 2-cm grid. The chamber was illuminated by a 10-W red light and equipped with a one-way mirror inserted in the door that permitted the measurement of locomotor behavior. As described previously (Fish et al., 2000), USVs were detected with a condenser microphone (type 4135; Bruel & Kjaer, Naerum, Denmark), preamplifier (type 2633; Bruel & Kjaer), and amplifier (type 2610; Bruel & Kjaer) in conjunction with a filter (Krohn-Hite model 3550R; nominal settings: 30-100 kHz, band-pass). This equipment provided a flat frequency response in the 30- to 65-kHz range.

Procedure. On test days the entire litter was separated from their parents and transferred with home cage shavings to an incubator that maintained nest temperature around 34°C. After ca. 20 min, pups were weighed and screened for baseline calling rate when individually placed onto the cold platform of the test chamber for 30 s. Only pups that produced more than six calls (USV) per 30 s and weighed between 3.5 and 5.5 g (ca. 90% of the animals) were assigned randomly to the treatment groups. The test procedure began with recording rectal temperature with a thermo-probe that was lubricated with mineral oil, inserted ca. 7 mm, and kept in place until the temperature reading was stable for at least 3 s. After the pup was injected with the assigned drug or vehicle treatment, it was returned to the incubator for a 10-min postinjection interval. Immediately before and after the separation test, two additional measurements of rectal temperature were taken. To measure the number of USVs, the pup was placed onto the cool surface (19 ± 1°C) for 4 min. Concurrently, the frequency of two additional endpoints was recorded: grid crossing and "rolls". A grid crossing was counted when half of the pup's body crossed into the next grid and a "roll" was counted whenever the back of the pup made contact with the surface of the test chamber. The experimental sessions were conducted between 9:00 AM and 5:00 PM. No differences in baseline rates of vocalization were observed according to time of day. CO₂ inhalation was used to euthanize the pups at the end of the test session.

Drugs. Zolpidem, triazolam, and diazepam were purchased in base form from Research Biochemicals International (Natick, MA). Flumazenil base was a gift from Hoffman-LaRoche Pharmaceuticals (Nutley, NJ). -CCt base was synthesized as described in detail previously (Cox et al., 1995). All drugs were suspended with the aid of sonication in a solution containing 85% distilled water, 14% propylene glycol, and 1% Tween 80, and were administered via the s.c. route using a 27-gauge syringe and in a volume of 1 ml/100 g of body weight. Triazolam, zolpidem, diazepam, or vehicle was injected 10 min before the separation test. For antagonism, flumazenil or -CCt was given concurrently with the injection of the agonist. These injection intervals were derived based on intervals previously used in studies with rat pups (Vivian et al., 1997; Olivier et al., 1998a) and on the results from pilot experiments with mouse pups.

Analysis of Data. The measures analyzed were rate of USVs (calls/4 min), grid crossings/4 min, change in body temperature, and a derived score of motor "incoordination". The latter incoordination score consisted of the ratio of frequency of rolls per grid crosses during the 4-min session (the higher the ratio of rolls to grid crosses, the larger the degree of incoordination). USVs, change in body temperature, and incoordination were analyzed for effects of dose by separate one-way between-subjects ANOVA. If the ANOVA was significant, each treatment condition was compared with the appropriate vehicle control by a Dunnett's test ( = 0.05). To calculate the potency of drug effects, the data for USVs and incoordination were converted to percentage of vehicle control effects, and the dose of the agonist that produced 50% of the maximum possible effect for this agonist, or ED₅₀, was obtained. ED₅₀ values were calculated from first order regression of the linear portion of the dose-response function. Statistical reliability among ED₅₀ values was determined by calculating 95% confidence intervals (CIs) of each ED₅₀ based on the coefficient of variation from the regression analysis. Nonoverlapping 95% CIs were considered to be reliably different.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Seven-day-old mouse pups treated with vehicle emitted a mean of 367 ± 19.7 (S.E.M.) USVs after separation from the litter and exposed for 4 min to the 19°C temperature. During this test, the pups accumulated 31.7 ± 1.9 grid crosses with few indications of incoordination (0.04 ± 0.01 incidences of rolling on the back per grid cross). The body temperature of these pups decreased during the 4-min test by 4.3 ± 0.1°C. Mouse pups were returned to the litter immediately after the test and their core temperature recovered within 10 min.

Suppression of USVs. Zolpidem, triazolam, and diazepam dose dependently decreased the rate of USVs [Fig. 1A; F(4,97) = 14.3, F(4,98) = 21.8, and F(4,122) = 27.3, p values <0.05, respectively]. Dunnett's tests revealed for zolpidem that doses of 1.0 mg/kg or higher produced reliable decreases in USVs. For triazolam, the effective dose range was 0.3 mg/kg or higher, whereas for diazepam, the effective dose range was 3.0 mg/kg or higher.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Effects of the benzodiazepine receptor antagonists flumazenil (0.1 mg/kg) and -CCt (3.0 mg/kg) on USVs in 7-day-old mouse pups following maternal separation and administration of vehicle (Veh) or zolpidem (A), triazolam (B), and diazepam (C). Data are means ± S.E.M. of USVs emitted after tests with vehicle, expressed as percentage of vehicle control (USVs after vehicle control = 100%). Each point represents the data from an individual group of mice. The horizontal dashed line represents 100% of vehicle control, and asterisks represent a statistically reliable difference from the vehicle control condition, Dunnett's tests, p < 0.05.

Incoordination (Rolls per Grid Cross). The effects of the three BZ agonists on incoordination paralleled the effects of these drugs on USVs. In this regard, zolpidem, triazolam, and diazepam increased incoordination [Fig. 2; F(4,102) = 77.6, F(4,101) = 58.7, and F(4,121) = 76, respectively). Dunnett's tests revealed reliable increases in incoordination for the following doses of agonist: zolpidem, 1-10 mg/kg; triazolam, 0.03-0.3 mg/kg; diazepam, 1-10 mg/kg. Concurrent injection with 0.1 mg/kg flumazenil resulted in rightward and approximately parallel shifts to the right of the dose-response functions for zolpidem, triazolam, and diazepam (doses necessary to reliably increase incoordination after flumazenil: zolpidem, 1-100 mg/kg; triazolam, 0.1-3.0 mg/kg; diazepam, 30-300 mg/kg, Dunnett's test, p < 0.05). In contrast, the dose-response functions for zolpidem, triazolam, and diazepam were not altered by 3.0 mg/kg -CCt (Fig. 2).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Effects of the benzodiazepine receptor antagonists flumazenil and -CCt on motor incoordination (number of rolls per grid cross) in 7-day-old mouse pups following maternal separation and administration of vehicle (Veh) or zolpidem (A), triazolam (B), and diazepam (C). Data are mean number of rolls per grid cross ± S.E.M., expressed as percentage of vehicle control (motor incoordination after vehicle control = 100%). Other details as in Fig. 1.

Body Temperature. When separated from the parents and maintained at the 34°C nest temperature, neither vehicle nor BZ agonist treatment reliably altered body temperature (data not shown). During the 4-min test session, when the pups were exposed to the ambient temperature of 19°C, the body temperature of untreated mice was reduced by 4.3°C (±0.2). This hypothermia was dose dependently augmented by zolpidem [F(4,100) = 15.9, p < 0.05], triazolam [F(4,99) = 6.8, p < 0.05], and diazepam [F(4,121) = 7.2, p < 0.05; see Fig. 3 for all three drugs]. All doses of zolpidem increased hypothermia, whereas only the highest three doses of triazolam alone and the highest two doses of diazepam alone enhanced hypothermia (Dunnett's tests, p < 0.05; Fig. 3). After 0.1 mg/kg flumazenil, the dose-response function of zolpidem was shifted to the right, with doses of 30 and 100 mg/kg required to augment hypothermia (Dunnett's tests, p < 0.05; Fig. 3, top). Similarly, 3.0 mg/kg -CCt shifted the zolpidem dose-response function to the right (effective zolpidem dose range 10-100 mg/kg; p < 0.05). As with zolpidem, treatment with flumazenil shifted the dose-response functions for both triazolam and diazepam to the right (effective doses after flumazenil treatment: 1.0, 3.0 mg/kg triazolam; 300 mg/kg diazepam; p values <0.05; Fig. 3). In contrast to zolpidem, -CCt did not alter the hypothermic effects engendered by triazolam and diazepam.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Effects of the benzodiazepine receptor antagonists flumazenil and -CCt on hypothermia in 7-day-old mouse pups following maternal separation and administration of vehicle (Veh) or zolpidem (A), triazolam (B), and diazepam (C). Data are mean rectal temperature ± S.E.M., expressed as percentage of vehicle control (temperatures after vehicle control = 0%). Other details as in Fig. 1.

Effects of Increasing Doses of Flumazenil and -CCt. The 0.3-mg/kg dose of triazolam and the 10-mg/kg dose of zolpidem were equieffective in suppressing USVs. Therefore, these doses of zolpidem and triazolam were chosen to examine further the differential antagonism of USVs and incoordination. As can be seen in Fig. 4A, increasing doses of flumazenil resulted in an increase in the rate of USVs for both zolpidem [F(4,55) = 20.4, p < 0.05] and triazolam [F(4,56) = 14.0, p < 0.05]. Flumazenil was equipotent in reversing the suppression of USVs produced by zolpidem and triazolam. The suppression of USVs engendered by zolpidem also was reversed by increasing doses of -CCt [F(4,48) = 16.4, p < 0.05], whereas the suppression of USVs engendered by triazolam were not altered by -CCt up to a dose of 100 mg/kg.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Effects of increasing doses of the benzodiazepine receptor antagonists flumazenil and -CCt on maximum suppression of ultrasonic vocalizations (A), maximum increase in motor incoordination (B), and maximum augmentation of hypothermia (C), engendered by zolpidem (10 mg/kg) or triazolam (0.3 mg/kg). Other details as in Figs. 1 and 3. , triazolam (0.3) + flumazenil; , zolpidem (10) + flumazenil; , triazolam (0.3) + -CCt; and , zolpidem (10) + -CCt.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Consistent with the general findings observed with other neonatal rodents, mouse pups emitted USVs following separation from their dam and littermates (Vivian et al., 1997; Olivier et al., 1998a,b; Fish et al., 2000). Previous research additionally has demonstrated that compounds effective as anxiolytics in people, such as BZs and 5-HT agonists, suppress USVs in both rat and mouse pups (Vivian et al., 1997; Fish et al., 2000). In the present study, the typical BZs triazolam and diazepam suppressed USVs emitted by 7-day-old mouse pups, supporting the idea that attenuation of distress vocalizations in neonatal rodents may be a valid and sensitive procedure for identifying and evaluating potential antianxiety agents. Interestingly, the BZ/1 receptor agonist zolpidem also suppressed USVs effectively and in a dose-dependent manner, similar to typical BZs (Fish et al., 2000; present study). This observation is consistent with the results of an earlier report by Olivier et al. (1998a), in which zolpidem dose dependently attenuated USVs emitted by neonatal rats after maternal separation.

Zolpidem often has been shown to be largely ineffective in rodent models predictive of anxiolytic activity (Depoortere et al., 1986; for review, see Sanger et al., 1994). A possible conclusion from these observations is that stimulation of the BZ/1 receptor does not play a primary role in modulating the anxiolytic effects of BZs and related drugs. This hypothesis recently has received support from studies with transgenic mice, in which the 1, 2, and 3 subunits have been rendered insensitive to BZ agonists by targeted mutation (Rudolph et al., 1999; Löw et al., 2000; McKernan et al., 2000). Indeed, the anxiolytic-related effects of diazepam were not altered by the 1 or 3 subunit mutation, but were abolished by the 2 subunit mutation, clearly implicating a role for this latter subunit in the modulation of antianxiety effects of diazepam (Rudolph et al., 1999; Löw et al., 2000; McKernan et al., 2000). These recent findings with transgenic animals also have implicated the BZ/1 receptor subtype as the primary site of action contributing to motor and sedative effects of BZ ligands, consistent with the possibility that any anxiolytic activity of drugs with prominent activity at BZ/1 receptors may be masked by a concomitant increase in 1 receptor-mediated sedation (cf. Sanger et al., 1994). In the present study, zolpidem suppressed USVs in a manner similar to the suppression engendered by typical BZs, suggesting that zolpidem may have anxiolytic effects that often are not expressed due to this compound's action at the BZ/1 receptor.

Although suggestive of an anxiolytic action of zolpidem, the possibility of other effects of zolpidem either indirectly or directly affecting the ability of mouse pups to emit USVs should be considered. Two possible mechanisms underlying the suppression of USVs produced by zolpidem were explored by the antagonism experiments in the present study: the possibility of a mechanism of action not involving BZ receptor stimulation, and the possibility that zolpidem's effects on USVs were mediated by BZ receptors other than the BZ/1 subtype. With regard to the former possibility that zolpidem's suppression of USVs was not mediated by BZ receptors, the nonselective BZ antagonist flumazenil attenuated the suppression of USVs produced by zolpidem and the typical BZs. Higher doses of these BZ agonists reversed the blockade by flumazenil, indicative of surmountable pharmacological antagonism and a clear involvement of BZ receptors in the modulation of zolpidem-engendered suppression of distress vocalizations. The suppression of USVs produced by the highest doses of zolpidem and triazolam was antagonized in a dose-dependent manner by further increases in the flumazenil dose, providing additional support for the observation of pharmacological antagonism with this ligand. Collectively, these findings indicate that the suppression of USVs engendered by zolpidem, triazolam, and diazepam were modulated by BZ receptors, and further suggest that the zolpidem-engendered suppression of USVs was not due to action at receptors other than BZ receptors.

In contrast to the results with flumazenil, the BZ/1 receptor-selective antagonist -CCt displayed a distinct pattern of blockade of BZ agonist-engendered suppression of USVs. In this regard, -CCt dose dependently and reversibly antagonized the effects of zolpidem on USVs, whereas the effects of triazolam and diazepam were not altered over a wide range of -CCt doses. These findings provide clear evidence that the suppression in USVs engendered by zolpidem was modulated by a mechanism of action distinct from typical BZs. Moreover, because of -CCt's approximately 20-fold selectivity for BZ/1 receptors over other BZ receptor subtypes (Huang et al., 2000) these findings suggest that zolpidem suppressed USVs via a BZ/1 mechanism.

Various other behavioral measures were examined to determine whether other effects associated with BZ receptor stimulation, specifically motor impairment and hypothermia, may have played a significant role in the ability of BZ agonists to suppress USVs. All of the BZ agonists in the present study engendered motor incoordination, which consisted of a rolling motion that contrasts markedly with the normal locomotor activity of mouse pups. The incoordination effect produced by zolpidem was similar to the effects of the two typical BZs on this measure. Moreover, the incoordination effects of zolpidem, similar to those of triazolam and diazepam, were not antagonized by -CCt, although increasing the dose of -CCt did result in a partial blockade of zolpidem-induced incoordination. Nonetheless, the lack of a clear distinction among the BZ ligands in both the production and blockade of incoordination suggests that the suppression in USVs engendered by zolpidem was not due to motoric impairment. Furthermore, the relative lack of sensitivity of -CCt in blocking the effects of the BZ agonists raises the possibility that the BZ/1 receptor subtype does not contribute to this particular motor incoordination measure. An intriguing speculation is that the incoordination effect observed after BZ agonist administration in mouse pups involves a predominantly spinal mechanism, based on the observation that the spinal cord contains predominantly BZ/2 and 3 receptor subtypes (McKernan and Whiting, 1996).

In contrast to incoordination, the body temperature measurements revealed interesting differences among zolpidem and the typical BZs. First, zolpidem-induced hypothermia occurred at a low dose that did not suppress USVs, whereas hypothermia generally paralleled or occurred at higher doses of triazolam and diazepam. Second, the magnitude of hypothermia was greater with zolpidem compared with the conventional BZs, although the differences in maximum effect were not striking. Finally, and perhaps most importantly, -CCt surmountably antagonized the hypothermic effects of zolpidem but not those produced by triazolam or diazepam. These results with -CCt suggest that the hypothermic effects of zolpidem are modulated primarily by BZ/1 receptor subtypes. Moreover, this latter observation parallels the findings with USV suppression, suggesting that the suppression of USVs after zolpidem may be the consequence of this compound's hypothermic effects. Although clearly a possibility, this hypothesis is difficult to reconcile with the finding that USVs emitted by neonatal rodents characteristically increase, rather than decrease, as a result of lowering body temperature (Sokoloff and Blumberg, 1997). Examination of hypothermia as a mechanism, as well as other effects of BZ agonists that could influence emission of USVs (e.g., respiratory effects, Sokoloff and Blumberg, 1997), clearly warrants further investigation.

Growing evidence has revealed that USVs emitted by infant mice separated from their dam and littermates are susceptible to pharmacological and environmental manipulations, consistent with these calls reflecting a response to stressful stimuli (Nastiti et al., 1991; Brunner et al., 1999; Fish et al., 2000). These findings, in turn, support the use of distress calls in neonatal mice as a valid and efficient procedure for examining the anxiolytic-like effects of drugs as well as the mechanisms that contribute to anxiety reduction. Moreover, anxiolytic-related effects of many drugs, including BZ agonists, have been demonstrated in a variety of species, including rats, gerbils, and nonhuman primates (for review, see Miczek et al., 1995). Because of this diverse species and pharmacological generality, combined with the increasing use of transgenic mice with targeted mutations (McKernan et al., 2000), the USV suppression procedure in neonatal mice may prove a useful model for studying the neuropharmacology of stress and anxiety (cf. Fish et al., 2000). Based on the results of the present investigation, however, it may be postulated that -CCt-sensitive suppression of USVs represents a nonanxiolytic and BZ/1 receptor-mediated effect. If this hypothesis is upheld by further study, then future assessment of the ability of BZ ligands to suppress USVs may require a demonstration of -CCt-sensitive versus -insensitive effects, to determine not only the role of BZ receptor subtypes but also the extent to which the USVs represent anxiolytic-like effects.