Skip to main content
PMC home page
The Journal of Pharmacology and Experimental Therapeutics logoLink to The Journal of Pharmacology and Experimental Therapeutics
. 2009 Jun 29;330(3):876–883. doi: 10.1124/jpet.109.151845

Behavioral Effects of γ-Hydroxybutyrate, Its Precursor γ-Butyrolactone, and GABAB Receptor Agonists: Time Course and Differential Antagonism by the GABAB Receptor Antagonist 3-Aminopropyl(diethoxymethyl)phosphinic Acid (CGP35348)

Wouter Koek 1, Susan L Mercer 1, Andrew Coop 1, Charles P France 1
PMCID: PMC2729800  PMID: 19564487

Abstract

γ-Hydroxybutyrate (GHB) is used therapeutically and recreationally. The mechanism by which GHB produces its therapeutic and recreational effects is not entirely clear, although GABAB receptors seem to play an important role. This role could be complex, because there are indications that different GABAB receptor mechanisms mediate the effects of GHB and the prototypical GABAB receptor agonist baclofen. To further explore possible differences in underlying GABAB receptor mechanisms, the present study examined the effects of GHB and baclofen on operant responding and their antagonism by the GABAB receptor antagonist 3-aminopropyl(diethoxymethyl)phosphinic acid (CGP35348). Pigeons were trained to peck a key for access to food during response periods that started at different times after the beginning of the session. In these pigeons, GHB, its precursor γ-butyrolactone (GBL), and the GABAB receptor agonists baclofen and 3-aminopropyl(methyl)phosphinic acid hydrochloride (SKF97541) decreased the rate of responding in a dose- and time-dependent manner. CGP35348 shifted the dose-response curve of each agonist to the right, but the magnitude of the shift differed among the agonists. Schild analysis yielded a pA2 value of CGP35348 to antagonize GHB and GBL [i.e., 3.9 (3.7–4.2)] that was different (P = 0.0011) from the pA2 value to antagonize baclofen and SKF97541 [i.e., 4.5 (4.4–4.7)]. This finding is further evidence that the GABAB receptor mechanisms mediating the effects of GHB and prototypical GABAB receptor agonists are not identical. A better understanding of the similarities and differences between these mechanisms, and their involvement in the therapeutic effects of GHB and baclofen, could lead to more effective medications with fewer adverse effects.

γ-Hydroxybutyric acid (GHB) is an endogenous molecule, a marketed therapeutic drug, and a drug of abuse. GHB is a putative neuromodulator (Maitre, 1997) involved in the regulation of sleep and used clinically to treat narcolepsy (Fuller and Hornfeldt, 2003) and alcoholism (Poldrugo and Addolorato, 1999). GHB is also used recreationally (Gonzalez and Nutt, 2005). The precise mechanism by which GHB exerts its various effects is unknown.

GHB binds to specific sites in brain (Benavides et al., 1982) and to GABAB receptors (Mathivet et al., 1997). At present, there is little evidence that specific GHB binding sites mediate the in vivo effects of GHB (Wong et al., 2004). Instead, many studies suggest that GABAB receptors are particularly important for various behavioral effects of GHB, including hypolocomotion (Kaupmann et al., 2003), catalepsy (Carter et al., 2005), ataxia (Goodwin et al., 2005), loss of righting (Carai et al., 2001), decreased operant responding (Goodwin et al., 2005), and discriminative stimulus effects (Winter, 1981; Colombo et al., 1998; Carter et al., 2003, 2009; Koek et al., 2004, 2006). All of these effects of GHB are also produced by the prototypical GABAB receptor agonist baclofen (Carter et al., 2003, 2004, 2005), consistent with the involvement of GABAB receptors in the effects of GHB.

Although GABAB receptors probably mediate behavioral effects that GHB has in common with baclofen, there is growing evidence that the underlying GABAB receptor mechanisms are not identical. One line of evidence is from studies that examined the interactions of GHB and baclofen with antagonists at the N-methyl-d-aspartate (NMDA) subtype of glutamate receptors. The NMDA antagonist dizocilpine (MK-801) enhances GHB-induced catalepsy in rodents (Sevak et al., 2004, 2005; Koek and France, 2008). The cataleptic effects of GHB are enhanced not only by MK-801 but also by other drugs with NMDA antagonist activity, such as phencyclidine (PCP) and ketamine. However, these NMDA antagonists do not affect the cataleptic effects of baclofen (Koek and France, 2008). Similar interactions have been observed in drug discrimination studies; PCP enhances the discriminative stimulus effects of GHB but not of baclofen (Koek et al., 2007a). Differential enhancement of the effects of GHB and baclofen by NMDA antagonists suggests that the GABAB receptor mechanisms involved in the effects of GHB and baclofen may not be identical.

A second line of evidence that the GABAB receptor mechanisms underlying the effects of GHB and baclofen may be different is from antagonism studies. The GABAB receptor antagonist CGP35348 antagonizes the discriminative stimulus effects of GHB and baclofen, consistent with the involvement of GABAB receptors, but is less potent in antagonizing these effects of GHB than those of baclofen (Carter et al., 2006). Recently, we reported that CGP35348 was also less potent in antagonizing the cataleptic effects of GHB than those of baclofen (Koek et al., 2007b). Together, these findings suggest a possible role for GABAB receptor subtypes or different interactions with the same GABAB receptor in the behavioral effects of GHB and baclofen.

A detailed characterization of antagonist actions requires complete dose-response curves of the agonist in the presence of several doses of the antagonist. Such data are often analyzed by Schild regression, which yields information about the nature of the antagonism and the potency of the antagonist. This method compares the pattern of antagonism to that predicted by the simple competitive model (i.e., the agonist and antagonist compete for the same recognition sites on the receptor). If the Schild regression has a slope of unity, this is consistent with simple competitive antagonism; deviations from unity can signify noncompetitive antagonism, nonequilibrium steady states, or receptor-population heterogeneity (Kenakin, 1997). A limitation of the studies examining antagonism of GHB and baclofen that we have conducted to date is that their results could not be analyzed by Schild regression. The present study, aimed at remedying this limitation, is an effort to characterize in detail the antagonism by CGP35348 of behavioral effects of GHB, its precursor γ-butyrolactone (GBL), and the GABAB receptor agonists baclofen and SKF97541. The behavioral measure used was decreased operant responding, assessed in a procedure that provided information about the time course of agonist and antagonist effects. The results confirm preliminary findings of differential antagonism by CGP35348 of GHB and baclofen, and suggest that the underlying GABAB receptor mechanisms are different, which may have implications for their different profiles of preclinical and clinical activities.

Materials and Methods

Animals. Ten adult white Carneau pigeons (Columbia livia; Palmetto Pigeon Plant, Sumter, SC) were individually housed under a 12:12-h light/dark cycle. They had free access to water and were maintained between 80 and 90% of their free-feeding weight by food (Purina Pigeon Checkers, St. Louis, MO) received during experimental sessions and supplemental postsession feedings (Purina Pigeon Checkers or mixed grain). All subjects had drug discrimination histories (Koek et al., 2006) and had not received any drug for at least one month before the start of the current study. Animals were maintained and experiments were conducted in accordance with the Institutional Animal Care and Use Committee (The University of Texas Health Science Center at San Antonio) and with the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, 1996).

Apparatus. Experiments were conducted in sound attenuating, ventilated chambers (BRS/LVE, Laurel, MD) equipped with a response key that could be illuminated by a red light. After completion of each fixed ratio, the key light was extinguished for 4 s, during which time a white light illuminated the hopper where food (Purina Pigeon Checkers) was available. Chambers were connected by an interface (MED Associates Inc., St. Albans, VT) to a computer that used MED-PC IV software (MED Associates Inc.) to monitor and control inputs and outputs and to record the data.

Procedure. The procedure was similar to that described in detail by Schlinger and Poling (1988). In brief, pigeons trained to peck the response key for access to food were exposed to 12-h overnight sessions that were conducted four times per week, Monday through Thursday. During each session, the key was illuminated during 10 response periods, each starting at a different time after the beginning of the session (i.e., 0, 15, 30, and 60 min and 2, 4, 6, 8, 10, and 12 h). When the key was illuminated, 20 responses resulted in 4-s access to food (i.e., fixed ratio 20). A response period ended after five food presentations or 5 min, whichever occurred first. Between response periods the key light was off, and responses had no programmed consequence. Response periods began with a brief (0.25 s) operation and illumination of the hopper (i.e., a brief auditory and visual stimulus).

Once responding stabilized under the fixed-ratio schedule during each of the 10 response periods (i.e., no visible trend was evident for at least five consecutive sessions), subjects received an intramuscular injection of physiological saline before each session. Monday and Wednesday sessions were always preceded by a saline injection. If responding during a saline session did not differ by more than 20% from responding during the previous saline session, an antagonist and/or agonist was given before the next session (i.e., on Tuesday or Thursday). Otherwise, a saline session was conducted. Agonists were given immediately before the session, and the antagonist was given 10 min before an agonist.

Data Analysis. Response rates were calculated for each response period by dividing the number of responses by the duration (in seconds) of the period. For each animal, response rates during the periods that started at different times after drug administration were expressed as a percentage of the corresponding control values obtained during the previous vehicle session. The response rates during drug test sessions, expressed as percentage control, were averaged across animals, and mean values ± S.E.M. were plotted as a function of dose and time after drug administration.

To calculate doses needed to produce 50% of the maximal response (ED50) and their 95% confidence limits, the linear portion of the dose-response curves was analyzed by log-linear regression (Tallarida, 2000) of data from individual subjects by use of GraphPad Prism version 5.01 for Windows (GraphPad Software, San Diego, CA; www.graphpad.com), with the following equation: effect = slope × log(dose) + intercept. The linear portion comprised the data points at doses immediately below and above 50% and included not more than one dose with an effect larger than 80% and not more than one dose with an effect smaller than 20%. For each dose-response curve fitted to more than two doses, the replicates test (also called the lack-of-fit test; Draper and Smith, 1998) implemented in GraphPad Prism was used to examine whether the log dose-response data used in the log-linear regression deviated from linearity. F ratio tests in GraphPad Prism were used to compare dose-response curves with respect to their slopes and intercepts. For example, a nonsignificant F ratio for slopes and a significant F ratio for intercepts show that dose-response curves are parallel but occupy different positions on the dose axis.

To examine the onset and duration of agonist effects, ED50 values for each agonist were plotted as a function of time after drug administration. Differences among ED50 values were analyzed by use of the F ratio test implemented in GraphPad Prism and with the common slope calculated by GraphPad Prism to constrain the fit of the parallel-line assay (Tallarida, 2000), as detailed elsewhere (Koek et al., 2006). Dose-response data obtained at the time that the ED50 values were lowest (i.e., when the agonist seemed to reach peak effect) were used to examine antagonist effects. For each agonist, differences among its dose-response curves to decrease response rate in the presence of different doses of antagonist were analyzed by simultaneously fitting straight lines to the linear portion of the dose-response curves. Differences among the slopes and intercepts of the curves were analyzed with the F ratio test, and ED50 values and potency ratios were calculated by parallel-line analysis (Tallarida, 2000). For each agonist, potency ratios were used to calculate an apparent pA2 value for the antagonist according to the methods described by Arunlakshana and Schild (1959). The Schild plots were analyzed with the F ratio test to examine whether the slopes of the Schild regressions deviated significantly from -1, and then to examine differences among pA2 values by comparing the following models of increasing complexity (i.e., increasing number of parameters): 1) the same pA2 value for all agonists; 2) the same pA2 value for GHB and SKF97541, the same pA2 value for GHB and GBL; and 3) an individual pA2 value for each agonist. Thus, Schild plots were used to examine whether the effects of the antagonist differed among agonists. In addition, Schild analyses of data obtained at various times after the administration of the antagonist were used to examine the duration of antagonist activity.

Drugs. GHB, GBL, and (±)-baclofen were purchased from Sigma-Aldrich (St. Louis, MO). CGP35348 (sodium salt) and SKF97541 hydrochloride were synthesized at the University of Maryland. All compounds were dissolved in physiological saline (0.9% NaCl), with the exception of GHB, which was dissolved in sterile water. All compounds were injected intramuscularly in a volume of 0.1 to 1.0 ml. Doses are expressed as the form of the drug listed above.

Results

Under control conditions, key peck responses occurred at a rate that varied not more than 20% within pigeons, and that varied from 0.83 to 1.94 responses per second between pigeons. Baclofen decreased the rate of responding in a dose- and time-dependent manner (Fig. 1). Responding was decreased to less than 50% of control between 30 min and 4 h after the injection of baclofen (Fig. 1, ▴, ▾, ♦, and •), but not at shorter or longer intervals (Fig. 1, ○, □, ▵, ▿, and ⋄). The dose-response data obtained at 30 min and 1, 2, and 4 h after 5.6, 10, and 17.8 mg/kg baclofen were analyzed by log-linear regression. The four dose-response curves, which did not deviate from linearity (replicates test, P > 0.20), had a common slope [F(3,84) = 0.31, P = 0.82] but not a common ED50 [F(3,87) = 2.73, P = 0.049]. The ED50 of baclofen to decrease responding was lowest at 60 min after injection [i.e., 8.6 (95% confidence limits: 7.1–10) mg/kg; Fig. 2; Table 1] and was not different at 30 and 120 min [F(1,43) ≤ 1.69, P ≥ 0.20] but was higher at 240 min [F(1,43) = 8.28, P = 0.0062]. Based on these results, baclofen seemed to reach peak effect 30 to 120 min after injection. The other drugs reached peak effect at 30 to 120 (SKF97541) or at 15 to 30 min (GHB, GBL). Thus, at 30 min after injection, all drugs were maximally active.

Fig. 1.

Fig. 1.

Open in a new tab

Effects of baclofen, administered intramuscularly immediately before the session, on fixed ratio 20 responding in a one-key, food-reinforced procedure (n = 5–9). Each 12-h session consisted of 10 5-min response periods starting at different times after the beginning of the session. For each of these response periods, the rate of key peck responding is plotted as a function of dose. Symbols represent mean ± S.E.M.; S.E.M. values are contained by the symbol.

Fig. 2.

Fig. 2.

Open in a new tab

Effects of baclofen, SKF97541, GHB, and GBL, administered intramuscularly immediately before the session, on fixed ratio 20 responding in a one-key, food-reinforced procedure in pigeons (n = 4–9). The potency to decrease the rate of key peck responding is plotted as a function of time from the beginning of the session. Circles represent ED50 and 95% confidence limits. For each drug, open circles indicate ED50 values that differed significantly from the lowest observed value, and closed circles indicate ED50 values that did not. Squares represent interval estimates of ED50 values.

TABLE 1.

ED50 values (in milligrams per kilogram) for effects of agonists on food-reinforced responding in pigeons at different injection-test intervals and pA2 values of CGP35348 [in – log (mol/kg)] to antagonize these effects Numbers in parentheses are the 95% confidence limits.

Agonist
Time
CGP35348
pA2
0 3.2 10 32
min mg/kg
Baclofen 30 9.4 (7.4–12) 14 (10–N.D.) 22 (16–30) 33 (24–47) 4.46 (4.10–4.82)
60 8.6 (7.1–10) 15 (12–N.D.) 18 (14–22) 31 (25–40) 4.49 (3.93–5.05)
120 9.1 (7.9–10) 14 (12–17) 19 (16–22) 34 (28–40) 4.46 (4.07–4.85)
240 12 (9.6–15) 17 (13–N.D.) 22 (17–28) 36 (27–48) 4.34 (3.95–4.73)
SKF97541 15 0.077 (0.047–0.14) 0.12 (0.068–N.D.) 0.19 (N.D.) 0.42 (0.20–N.D.) 4.46 (4.11–4.81)
30 0.046 (0.034–0.063) 0.079 (0.057–0.11) 0.15 (N.D.) 0.20 (0.14–0.30) 4.63 (4.17–5.09)
60 0.046 (0.036–0.059) 0.078 (0.061–0.098) 0.11 (0.085–0.15) 0.18 (0.14–0.24) 4.54 (4.09–4.98)
120 0.051 (0.041–0.064) 0.079 (0.064–0.098) 0.11 (0.085–0.14) 0.22 (N.D.-30) 4.52 (4.21–4.82)
240 0.091 (0.067–0.13) 0.10 (0.074–0.16) 0.15 (0.10–N.D.) 0.30 (0.18–N.D.) 4.18 (3.92–4.43)
GHB 15 150 (120–190) 200 (160–250) 250 (190–N.D.) 250 (190–N.D.) 4.14 (3.26–5.02)
30 150 (130–180) 170 (140–200) 240 (210–270) 260 (220–N.D.) 3.97 (3.47–4.47)
60 160 (130–200) 180 (150–210) 280 (230–N.D.) 340 (280–400) 3.96 (2.38–5.54)
GBL 30 79 (67–93) 100 (82–130) 94 (76–120) 130 (N.D.-170) 3.91 (2.98–4.85)
60 100 (81–130) 110 (86–140) 140 (130–150) 150 (120–N.D.) 3.72 (3.24–4.19)
Open in a new tab

N.D., not determined because the confidence limit could not be calculated

When injected alone at a dose of 32 mg/kg, the GABAB receptor antagonist CGP35348 did not alter the rate of responding (data not shown). The response rate, expressed as a percentage of saline control, did not significantly change during the session and varied between 98 and 109% (S.E.M., 2.5–6.3). When injected 10 min before each agonist, CGP35348 dose-dependently shifted the dose-response curve of each agonist to the right (Fig. 3, top and middle; Table 1). However, the extent and nature of these shifts seemed to differ among some of the agonists. To examine these apparent shifts, the linear portion of each dose-response curve (see Materials and Methods) was analyzed by log-linear regression. None of the dose-response data used in the regression analyses deviated from linearity [replicates test: P values ranged from a minimum of 0.13 obtained for F(1,9) = 2.73 to a maximum of 0.98 for F(2,14) = 0.02]. CGP35348 shifted the dose-response curves of baclofen and SKF97541 to the right [common intercept: F(3,49) ≥ 7.80, P ≤ 0.0002], in a parallel manner [common slope: F(3,46) ≤ 1.29. P ≥ 0.29], and to an apparently similar extent. At the same doses, CGP35348 shifted the dose-response curves of GHB [common intercept: F(3,41) = 7.53, P = 0.0004] and GBL [common intercept: F(3,47) = 3.37, P = 0.026], but apparently less extensively, and in the case of GHB, in a nonparallel manner [GHB, common slope: F(3,41) = 6.34, P = 0.0012; GBL, common slope: F(3,44) = 2.02, P = 0.12]. The antagonist effects of CGP35348 were quantified by means of Schild regression plots (Fig. 3, bottom). These plots, with a common slope [F(3,4) = 0.47, P = 0.72] not different from -1[F(1,10) = 2.04, P = 0.18], yielded the following pA2 values (Table 1): 4.46 (95% confidence limits: 4.10–4.82) for baclofen, 4.63 (4.17–5.09) for SKF97541, 3.97 (3.47–4.47) for GHB, and 3.91 (2.98–4.85) for GBL. The confidence interval of the estimated pA2 value was wider for GBL (i.e., 1.87) than for baclofen, SKF97541, and GHB (i.e., 0.72, 0.92, and 1, respectively), indicating that the Schild regression fitted the data obtained with GBL less well than those obtained with the other drugs. To examine similarities and differences among the pA2 values obtained for CGP35348 with each of the four agonists, the following models of increasing complexity (i.e., larger number of parameters) were compared by means of F tests: model 1) a common pA2 value for all agonists; model 2) a common pA2 value for baclofen and SKF97541 and a common pA2 value for GHB and GBL; and model 3) an individual pA2 value for each agonist. Model 2 fitted the data better than model 1 did [F(1,10) = 20.74, P = 0.0011], which indicates that the data could not be adequately fitted with a single pA2 value. However, adding more parameters, by assuming that the pA2 values differed for each agonist, did not further increase the fit [comparison of model 3 with model 2: F(2,8) = 0.42, P = 0.67]. Thus, model 2 was the simplest model that could be fitted to the Schild regression data obtained with all four drugs and consisted of one plot for baclofen and SKF97541 and one for GHB and GBL. Based on these plots, the pA2 value of CGP35348 was 4.54 (4.36–4.73) to antagonize baclofen and SKF97541, and was 3.94 (3.66–4.23) to antagonize GHB and GBL. Thus, CGP35348 was 4-fold less potent to antagonize the response rate decreasing effects of GHB and GBL than to antagonize those of baclofen and SKF97541.

Fig. 3.

Fig. 3.

Open in a new tab

Top and middle, attenuation by CGP35348 of response rate-decreasing effects of baclofen, SKF97541, GHB, and GBL in pigeons (n = 5–6). CGP35348 or saline was injected 10 min before each agonist, and 30 min after the agonist was injected, key peck responding was measured. For each agonist, the rate of key peck responding, expressed as a percentage of saline control, is plotted as a function of dose after pretreatment with saline (•) or different doses of CGP35348 [○, ▵, ▿, data comprising the linear portion of the dose-response curves, and fitted with log-linear regression lines; gray-filled symbols, data obtained with the same doses of CGP35348 as the corresponding open symbols, but not part of the linear portion of the dose-response curves, and not used in the regression calculations (see “Data Analysis”)]. Symbols represent mean ± S.E.M.; if not shown, S.E.M. values are contained by the symbol. Bottom, Schild regression plots for antagonism by CGP35348 of the response rate-decreasing effects of baclofen, SKF97541, GHB, and GBL. Dose ratios are the ED50 values of SKF97541 in the presence of CGP35348 (3.2–32 mg/kg) divided by the ED50 value after pretreatment with saline. ED50 values were calculated from the regression lines shown in the top and middle. Data obtained with each agonist could be fitted with a regression line with a slope of -1. Calculated from these regression lines, the pA2 value of CGP35348 ranged from 3.91 (95% confidence limits, 2.98–4.85) for GBL to 4.63 (4.17–5.09) for SKF97541.

SKF97541 dose-dependently decreased response rate not only at 30 min but also at shorter and longer injection-test intervals (Fig. 4, left, top to bottom, and right, top and middle). None of the dose-response data used in the regression analyses deviated from linearity [replicates test: P values ranged from a minimum of 0.19 obtained for F(2,15) = 1.88 to a maximum of 0.98 for F(2,14) = 0.02]. CGP35348, injected 10 min before SKF97541, dose-dependently shifted the dose-response curves of SKF97541 at each interval to the right [common intercept: 15 min, F(3,42) = 3.53, P = 0.023; 30 min, F(3,49) = 7.80, P = 0.0002; 60 min, F(3,50) = 13.13, P < 0.0001; 120 min, F(3,46) = 12.38, P < 0.0001; 240 min, F(4.29), P = 0.009], in a parallel manner [common slope: 15 min, F(3,39) = 0.44, P = 0.73; 30 min, F(3,46) = 0.50, P = 0.69; 60 min, F(3,47) = 0.99, P = 0.41; 120 min, F(3,43) = 2.45, P = 0.077; 240 min, F(3,47) = 1.84, P = 0.15]. However, the pA2 values for CGP35348 obtained at these intervals were not the same (Fig. 4, bottom right; Table 1). The Schild regression plots, with a common slope [F(4,5) = 3.55, P = 0.10] not different from -1[F(1,13) = 2.48, P = 0.14], yielded apA2 value at 240 min after the injection of SKF9754 [i.e., 4.18 (3.93–4.43)] that differed from the values obtained at shorter intervals (Table 1). To examine similarities and differences among the pA2 values obtained at each of the intervals, the following models of increasing complexity (i.e., larger number of parameters) were compared by means of F tests: model 1) a common pA2 value for all intervals; model 2) a common pA2 value for the 15- to 120-min intervals, and a pA2 value for the 240-min interval; and model 3) an individual pA2 value for each interval. Model 2 fitted the data better than model 1 [F(1,13) = 14.87, P = 0.002], which indicates that the data could not be adequately fitted with a single pA2 value. However, adding more parameters, by assuming the pA2 values to differ for each interval, did not further increase the fit [comparison of model 3 with model 2: F(3,10) = 0.62, P = 0.62]. Thus, model 2 was the simplest model that could be fitted to the Schild regression data obtained at all five intervals and consisted of one plot for the 15- to 120-min intervals and one plot for the 240-min interval. Based on these plots, the pA2 value of CGP35348 was 4.53 (4.44–4.63) to antagonize SKF97541 at 15 to 120 min after its injection, and it was 4.18 (3.93–4.43) at 240 min. Thus, the potency of CGP35348 to antagonize SKF97541 was not different from 15 to 120 min after the injection of SKF97541 (i.e., 25–130 min after CGP35348); however, 250 min after the injection of CGP35348, its potency decreased approximately 2-fold. A similar trend was apparent when CGP35348 was used to antagonize the effects of baclofen at various time intervals (Fig. 5), but this failed to reach statistical significance [F(1,10) = 1.46, P = 0.25]. The pA2 values obtained with GHB and GBL did not differ across intervals [GHB: F(2,5) = 0.40, P = 0.69; GBL: F(1,4) = 0.63, P = 0.47] but were lower than those obtained with baclofen and SKF97541 at the same interval [30 min: GHB versus baclofen, F(1,4) = 11.40, P = 0.028; GHB versus SKF97541, F(1,4) = 17.39, P = 0.014; GBL versus baclofen, F(1,4) = 5.51, P = 0.079; GBL versus SKF97541, F(1,4) = 8.85, P = 0.041; 60 min: GHB versus baclofen, F(1,3) = 7.63, P = 0.07; GHB versus SKF97541, F(1,3) = 12.80, P = 0.037; GBL versus baclofen, F(1,4) = 20.42, P = 0.011; GBL versus SKF97541, F(1,4) = 29.60, P = 0.0055]. Thus, CGP35348 was most potent between 25 and 130 min after its administration, and 4-fold more potent to antagonize baclofen and SKF97541 than GHB and GBL (Fig. 5).

Fig. 4.

Fig. 4.

Open in a new tab

Time-dependent attenuation by CGP35348 of the response rate-decreasing effects of SKF97541 in pigeons (n = 5–6) (left, top to bottom, and right, top and middle). CGP35348 or saline was injected 10 min before SKF97541, and key peck responding was measured at different times after the agonist was injected (range, 15–240 min). For each time interval, the rate of key peck responding, expressed as a percentage of saline control, is plotted as a function of dose after pretreatment with saline (•) or different doses of CGP35348 [○, ▵, ▿, data comprising the linear portion of the dose-response curves, and fitted with log-linear regression lines; gray-filled symbols, data obtained with the same doses of CGP35348 as the corresponding open symbols, but not part of the linear portion of the dose-response curves, and not used in the regression calculations (see “Data Analysis”)]. Symbols represent mean ± S.E.M.; if not shown, S.E.M. values are contained by the symbol. Bottom right, Schild regression plots for the ability of CGP35348 to antagonize, at different times after its administration (in minutes), the response rate-decreasing effects of SKF97541. Dose ratios are the ED50 values of each agonist in the presence of CGP35348 (3.2–32 mg/kg) divided by the ED50 value of the agonist after pretreatment with saline. Data obtained at each time interval could be fitted with a regression line with a slope of -1. Calculated from these regression lines, the pA2 value of CGP35348 to antagonize SKF97541 at the different intervals ranged from 4.18 (95% confidence limits, 3.92–4.43) at 240 min after SKF97541 to 4.63 (4.17–5.09) at 30 min after SKF97541 (i.e., 250 and 40 min after CGP35348, respectively).

Fig. 5.

Fig. 5.

Open in a new tab

Antagonism by CGP35348 of the rate-decreasing effects of different agonists at different times after their administration. pA2 values are plotted for each agonist at the times it had rate-decreasing effects. CGP35348 was approximately 4-fold less potent to antagonize the effects of GHB and GBL than to antagonize the effects of baclofen and SKF97541. CGP35348 was approximately 2-fold less potent to antagonize SKF97541 at 240 min after its administration than at shorter intervals. a, P < 0.05 compared with baclofen and SKF; b, P < 0.05 and <0.10 compared with SKF97541 and baclofen, respectively; c, P < 0.05 compared with SKF97541 at all shorter intervals.

Discussion

GABAB receptors seem to play an important role in the effects of GHB; however, the effects of GHB, although similar, are not identical to those of the prototypical GABAB receptor agonist baclofen (e.g., Carter et al., 2009). The main finding of the present study is that the GABAB receptor antagonist CGP35348 was significantly more potent to antagonize behavioral effects of the GABAB receptor agonists baclofen and SKF97541 than those of GHB and its precursor GBL. Schild plots with slope values not significantly different from -1, consistent with simple competitive antagonism, yielded a pA2 value of CGP35348 to antagonize baclofen and SKF97541 [i.e., 4.5 (4.4–4.7)] that was similar to in vitro pA2 values reported previously for CGP35348 to antagonize baclofen (i.e., 4.3–4.7) (Kerr et al., 1993; Olianas and Onali, 1999), but that was approximately 4-fold higher than its pA2 value to antagonize GHB and GBL [i.e., 3.9 (3.7–4.2)]. These findings are consistent with previous observations of differential antagonism by CGP35348 of the discriminative stimulus effects of baclofen and GHB in rats (Carter et al., 2006) and of the cataleptic effects of baclofen and GHB in mice (Koek et al., 2007b), and extend them to effects on operant responding in pigeons. Together, these data are further evidence that GABAB receptors mediate many behavioral effects of GHB, but they also suggest that the underlying GABAB mechanisms differ from those mediating the effects of prototypical GABAB receptor agonists such as baclofen.

Additional evidence that the GABAB receptor mechanisms underlying the effects of GHB and baclofen are not identical has recently been obtained in interaction studies with NMDA antagonists. The NMDA antagonist PCP and GHB enhance each other's discriminative stimulus effects, but PCP and baclofen do not, suggesting that the mechanisms underlying these effects of GHB and baclofen are differentially modulated by the glutamatergic system with which PCP interacts (Koek et al., 2007a). The recent finding that PCP and other antagonists at the NMDA subtype of glutamate receptors enhance the cataleptic effects of GHB but not those of baclofen (Koek and France, 2008) provides further evidence that the GABAB receptor systems mediating the effects of GHB and baclofen are differentially modulated by glutamate.

Recent electrophysiological studies offer further evidence of differing effects of GHB and baclofen. At concentrations described as clinically relevant, GHB disinhibits and baclofen inhibits ventral tegmental dopamine neurons (Cruz et al., 2004). It has been suggested that GHB is more likely to activate the dopamine system implicated in addiction, whereas baclofen, which may be useful to reduce a relapse to taking cocaine (Weerts et al., 2007), may have more pronounced anticraving effects (Cruz et al., 2004). GHB and baclofen differ also in their effects on neurotransmission at glutamate receptors (Li et al., 2007). GHB and baclofen both inhibited currents elicited by NMDA and α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid, and their effects could be reversed by the GABAB receptor antagonist CGP62349. However, GHB was more potent to inhibit NMDA-elicited currents, whereas baclofen was more potent to inhibit α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid-elicited currents (Li et al., 2007). These latter findings, together with previous in vivo findings (Koek et al., 2007a; Koek and France, 2008), suggest a more prominent role for NMDA receptors in the GABAB receptor-mediated effects of GHB. Thus, evidence is emerging that the effects of GHB and baclofen, although in many respects similar, may be mediated by different GABAB systems.

There is evidence for functional GABAB receptor subtypes (Seabrook et al., 1990; Bonanno and Raiteri, 1992; Lanza et al., 1993; Fassio et al., 1994; Yamada et al., 1999). Conceivably, differential activity of GHB and baclofen at GABAB autoreceptors and heteroreceptors could account for the differential enhancement of their effects by NMDA antagonists, with effects of GHB mediated by glutamatergic GABAB heteroreceptors and effects of baclofen by GABAB autoreceptors. Differential activity of GHB and baclofen at these receptors could also account for the differential ability of CGP35348 to antagonize their effects. Alternatively, GHB and baclofen may interact differently with the same GABAB receptor (e.g., GHB may induce conformational changes in the GABAB receptor that differ from those induced by baclofen). Further studies examining the different GABAB mechanisms that underlie the effects of GHB and baclofen may help to explain why GHB is effective in treating narcolepsy and is abused, whereas there is no evidence that baclofen is effective in any sleep disorder or that it is abused.

The procedure used in the present study yielded not only quantitative measures of antagonist potency, but also information about the time course of agonist and antagonist effects. GHB and GBL had a more rapid onset and a shorter duration of action than baclofen and SKF97541 to decrease operant response rate in pigeons, in agreement with and extending previous observations in rats (Carter et al., 2004). Because baclofen and SKF97541 had long-lasting effects, pA2 values for CGP35348 to antagonize these effects could be calculated for injection-test intervals ranging from 15 to 240 min. Changes of pA2 values over time have been used to provide a description of the duration of action of antagonists (Gerak and France, 2007). From these pA2 values obtained at different injection-test intervals, it seems that the antagonist potency of CGP35348 was maximal and remained unchanged from 25 to 130 min after its injection. At 240 min, the antagonist potency of CGP35348 decreased, perhaps resulting from it being eliminated. At least 120 min of duration of antagonist action makes CGP35348 suitable for use as a pretreatment in cumulative dosing procedures. In general, cumulative dosing is generally more rapid and economical than single dosing to obtain full dose-response curves. Thus, future studies with CGP35348 in pigeons will use cumulative dosing of agonists.

In summary, the GABAB receptor antagonist CGP35348 was significantly more potent to antagonize behavioral effects of the GABAB receptor agonists baclofen and SKF97541 than those of GHB and its precursor GBL. This is further evidence that the effects of GHB and prototypical GABAB agonists are mediated by GABAB receptor mechanisms that are not identical. A better understanding of these mechanisms may help to explain why GHB is effective for treating narcolepsy and is abused, whereas baclofen is not, and this knowledge could lead to more effective medications with fewer adverse effects.

Acknowledgments

We thank Debbie Rodriguez, Adela Garza, Jason Persyn, and Christopher Limas for technical assistance.

This work was supported by the National Institutes of Health National Institute on Drug Abuse [Grants DA15692, DA17918, DA19634].

Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.

doi:10.1124/jpet.109.151845.

ABBREVIATIONS: GHB, γ-hydroxybutyrate; GBL, γ-butyrolactone; SKF97541, 3-aminopropyl(methyl)phosphinic acid hydrochloride; CGP35348, 3-aminopropyl(diethoxymethyl)phosphinic acid; CGP62349, [3-[1-(R)-[[(2S)-2-hydroxy-3-[hydroxyl[4-methoxyphenyl)methyl]phosphinyl]propyl]-amino]ethyl]-benzoic acid; NMDA, N-methyl-d-aspartate; PCP, phencyclidine; MK-801, dizocilpine.

References

  1. Arunlakshana O and Schild HO (1959) Some quantitative uses of drug antagonists. Br J Pharmacol Chemother 14 48-58. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Benavides J, Rumigny JF, Bourguignon JJ, Cash C, Wermuth CG, Mandel P, Vincendon G, and Maitre M (1982) High affinity binding sites for gamma-hydroxybutyric acid in rat brain. Life Sci 30 953-961. [DOI] [PubMed] [Google Scholar]
  3. Bonanno G and Raiteri M (1992) Functional evidence for multiple γ-aminobutyric acidB receptor subtypes in the rat cerebral cortex. J Pharmacol Exp Ther 262 114-118. [PubMed] [Google Scholar]
  4. Carai MA, Colombo G, Brunetti G, Melis S, Serra S, Vacca G, Mastinu S, Pistuddi AM, Solinas C, Cignarella G, et al. (2001) Role of GABA(B) receptors in the sedative/hypnotic effect of gamma-hydroxybutyric acid. Eur J Pharmacol 428 315-321. [DOI] [PubMed] [Google Scholar]
  5. Carter LP, Chen W, Coop A, Koek W, and France CP (2006) Discriminative stimulus effects of GHB and GABA(B) agonists are differentially attenuated by CGP35348. Eur J Pharmacol 538 85-93. [DOI] [PubMed] [Google Scholar]
  6. Carter LP, Flores LR, Wu H, Chen W, Unzeitig AW, Coop A, and France CP (2003) The role of GABAB receptors in the discriminative stimulus effects of γ-hydroxybutyrate in rats: time course and antagonism studies. J Pharmacol Exp Ther 305 668-674. [DOI] [PubMed] [Google Scholar]
  7. Carter LP, Koek W, and France CP (2009) Behavioral analyses of GHB: receptor mechanisms. Pharmacol Ther 121 100-114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Carter LP, Wu H, Chen W, Cruz CM, Lamb RJ, Koek W, Coop A, and France CP (2004) Effects of gamma-hydroxybutyrate (GHB) on schedule-controlled responding in rats: Role of GHB and GABAB receptors. J Pharmacol Exp Ther 308 182-188. [DOI] [PubMed] [Google Scholar]
  9. Carter LP, Wu H, Chen W, Matthews MM, Mehta AK, Hernandez RJ, Thomson JA, Ticku MK, Coop A, Koek W, et al. (2005) Novel gamma-hydroxybutyric acid (GHB) analogs share some, but not all, of the behavioral effects of GHB and GABAB receptor agonists. J Pharmacol Exp Ther 313 1314-1323. [DOI] [PubMed] [Google Scholar]
  10. Colombo G, Agabio R, Lobina C, Reali R, and Gessa GL (1998) Involvement of GABA(A) and GABA(B) receptors in the mediation of discriminative stimulus effects of gamma-hydroxybutyric acid. Physiol Behav 64 293-302. [DOI] [PubMed] [Google Scholar]
  11. Cruz HG, Ivanova T, Lunn ML, Stoffel M, Slesinger PA, and Lüscher C (2004) Bi-directional effects of GABA(B) receptor agonists on the mesolimbic dopamine system. Nat Neurosci 7 153-159. [DOI] [PubMed] [Google Scholar]
  12. Draper N and Smith H (1998) Checking the straight line fit, in Applied Regression Analysis, 3rd ed, pp 47-56, Wiley Interscience, Malden, MA.
  13. Fassio A, Bonanno G, Cavazzani P, and Raiteri M (1994) Characterization of the GABA autoreceptor in human neocortex as a pharmacological subtype of the GABAB receptor. Eur J Pharmacol 263 311-314. [DOI] [PubMed] [Google Scholar]
  14. Fuller DE and Hornfeldt CS (2003) From club drug to orphan drug: sodium oxybate (xyrem) for the treatment of cataplexy. Pharmacotherapy 23 1205-1209. [DOI] [PubMed] [Google Scholar]
  15. Gerak LR and France CP (2007) Time-dependent decreases in apparent pA2 values for naltrexone studied in combination with morphine in rhesus monkeys. Psychopharmacology (Berl) 193 315-321. [DOI] [PubMed] [Google Scholar]
  16. Gonzalez A and Nutt DJ (2005) Gamma hydroxy butyrate abuse and dependency. J Psychopharmacol 19 195-204. [DOI] [PubMed] [Google Scholar]
  17. Goodwin AK, Froestl W, and Weerts EM (2005) Involvement of gamma-hydroxybutyrate (GHB) and GABA-B receptors in the acute behavioral effects of GHB in baboons. Psychopharmacology (Berl) 180 342-351. [DOI] [PubMed] [Google Scholar]
  18. Institute of Laboratory Animal Resources (1996) Guide for the Care and Use of Laboratory Animals 7th ed. Institute of Laboratory Animal Resources, Commission on Life Sciences, National Research Council, Washington DC.
  19. Kaupmann K, Cryan JF, Wellendorph P, Mombereau C, Sansig G, Klebs K, Schmutz M, Froestl W, van der Putten H, Mosbacher J, et al. (2003) Specific gamma-hydroxybutyrate-binding sites but loss of pharmacological effects of gamma-hydroxybutyrate in GABA(B)(1)-deficient mice. Eur J Neurosci 18 2722-2730. [DOI] [PubMed] [Google Scholar]
  20. Kenakin T (1997) Competitive antagonism, in Pharmacologic Analysis of Drug-Receptor Interaction, 3rd ed, pp 331-373, Lippincott-Raven, Philadelphia, PA.
  21. Kerr DI, Ong J, Doolette DJ, Abbenante J, and Prager RH (1993) 3-Amino-2-(4-chlorophenyl)-nitropropane is a new GABAB receptor agonist, more active peripherally. Eur J Pharmacol 236 239-245. [DOI] [PubMed] [Google Scholar]
  22. Koek W, Chen W, Mercer SL, Coop A, and France CP (2006) Discriminative stimulus effects of γ-hydroxybutyrate: role of training dose. J Pharmacol Exp Ther 317 409-417. [DOI] [PubMed] [Google Scholar]
  23. Koek W, Flores LR, Carter LP, Lamb RJ, Chen W, Wu H, Coop A, and France CP (2004) Discriminative stimulus effects of γ-hydroxybutyrate in pigeons: role of diazepam-sensitive and -insensitive GABAA and GABAB receptors. J Pharmacol Exp Ther 308 904-911. [DOI] [PubMed] [Google Scholar]
  24. Koek W and France CP (2008) Cataleptic effects of gamma-hydroxybutyrate (GHB) and baclofen in mice: Mediation by GABA(B) receptors, but differential enhancement by N-methyl-d-aspartate (NMDA) receptor antagonists. Psychopharmacology (Berl) 199 191-198. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Koek W, Khanal M, and France CP (2007a) Synergistic interactions between `club drugs': gamma-hydroxybutyrate and phencyclidine enhance each other's discriminative stimulus effects. Behav Pharmacol 18 807-810. [DOI] [PubMed] [Google Scholar]
  26. Koek W, Mercer SL, and Coop A (2007b) Cataleptic effects of gamma-hydroxybutyrate (GHB), its precursor gamma-butyrolactone (GBL), and GABAB receptor agonists in mice: differential antagonism by the GABAB receptor antagonist CGP35348. Psychopharmacology (Berl) 192 407-414. [DOI] [PubMed] [Google Scholar]
  27. Lanza M, Fassio A, Gemignani A, Bonanno G, and Raiteri M (1993) CGP 52432: A novel potent and selective GABAB autoreceptor antagonist in rat cerebral cortex. Eur J Pharmacol 237 191-195. [DOI] [PubMed] [Google Scholar]
  28. Li Q, Kuhn CM, Wilson WA, and Lewis DV (2007) Effects of gamma hydroxybutyric acid on inhibition and excitation in rat neocortex. Neuroscience 150 82-92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Maitre M (1997) The gamma-hydroxybutyrate signalling system in brain: organization and functional implications. Prog Neurobiol 51 337-361. [DOI] [PubMed] [Google Scholar]
  30. Mathivet P, Bernasconi R, De Barry J, Marescaux C, and Bittiger H (1997) Binding characteristics of gamma-hydroxybutyric acid as a weak but selective GABAB receptor agonist. Eur J Pharmacol 321 67-75. [DOI] [PubMed] [Google Scholar]
  31. Olianas MC and Onali P (1999) GABA(B) receptor-mediated stimulation of adenylyl cyclase activity in membranes of rat olfactory bulb. Br J Pharmacol 126 657-664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Poldrugo F and Addolorato G (1999) The role of gamma-hydroxybutyric acid in the treatment of alcoholism: from animal to clinical studies. Alcohol Alcohol 34 15-24. [DOI] [PubMed] [Google Scholar]
  33. Schlinger H and Poling A (1988) Evaluation of a procedure to measure the time course of a drug's behavioral action. J Pharmacol Methods 20 169-174. [DOI] [PubMed] [Google Scholar]
  34. Seabrook GR, Howson W, and Lacey MG (1990) Electrophysiological characterization of potent agonists and antagonists at pre- and postsynaptic GABAB receptors on neurones in rat brain slices. Br J Pharmacol 101 949-957. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Sevak RJ, France CP, and Koek W (2004) Neuroleptic-like effects of gamma-hydroxybutyrate: Interactions with haloperidol and dizocilpine. Eur J Pharmacol 483 289-293. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sevak RJ, Koek W, and France CP (2005) Streptozotocin-induced diabetes differentially modifies haloperidol- and gamma-hydroxybutyric acid (GHB)-induced catalepsy. Eur J Pharmacol 517 64-67. [DOI] [PubMed] [Google Scholar]
  37. Tallarida RJ (2000) Dose-response analysis, in Drug Synergism and Dose-Effect Data Analysis, pp 44-50, Chapman & Hall/CRC, Boca Raton, FL.
  38. Weerts EM, Froestl W, Kaminski BJ, and Griffiths RR (2007) Attenuation of cocaine-seeking by GABA B receptor agonists baclofen and CGP44532 but not the GABA reuptake inhibitor tiagabine in baboons. Drug Alcohol Depend 89 206-213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Winter JC (1981) The stimulus properties of gamma-hydroxybutyrate. Psychopharmacology (Berl) 73 372-375. [DOI] [PubMed] [Google Scholar]
  40. Wong CG, Gibson KM, and Snead OC 3rd (2004) From the street to the brain: Neurobiology of the recreational drug gamma-hydroxybutyric acid. Trends Pharmacol Sci 25 29-34. [DOI] [PubMed] [Google Scholar]
  41. Yamada K, Yu B, and Gallagher JP (1999) Different subtypes of GABAB receptors are present at pre- and postsynaptic sites within the rat dorsolateral septal nucleus. J Neurophysiol 81 2875-2883. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Pharmacology and Experimental Therapeutics are provided here courtesy of American Society for Pharmacology and Experimental Therapeutics

RESOURCES