- Email: [email protected]
Behavioral effects of benzodiazepine antagonists in chlordiazepoxide tolerant and non-tolerant rats
Llfe Sclences, Vol. 44, pp 289-299 Prlnted in the U.S.A.
Pergamon Press
BEHAVIORAL EFFECTSOF BENZODIAZEPINEANTAGONISTS IN CHLORDIAZEPOXIDE TOLERANT ANO NON-TOLERANTRATS K. Takada1., T. Suzuki2., T. Hagen** J.M. Cook** and J.L. Katz 3. *** ' *NIDA Addiction Research Center, Baltimore, Maryland 21224 **Department of Chemistry, University of Wisconsin, Milwaukee, Wisconsin 53201 ***Department of Pharmacology and Experimental Therapeutics University of Maryland School of Medicine Baltimore, Maryland 21201 (Received in f£nal form December 1, 1988)
Summary Rats were trained to respond under 3-min fixed-interval schedules of food presentation, and effects of the benzodiazepine-receptor ligands, flumazenil, 2-(4-methoxy-phenyl)-pyrazolo[4,3-c]quinolin3(5H)-one (CGS 9895), 3-carbo-t-butoxy-6-carboline (6-CCtB), and 6-carboline-3-carboxylic acid ethyl ester (13-CCE) were assessed before and after the induction of tolerance to chlordiazepoxide. Before daily administration of chlordiazepoxide, none of the antagonists produced appreciable effects on rates of responding up to doses of 32.0 mg/kg i.p. 6-CCE was the only antagonist studied at a higher dose (100.0 mg/kg i . p . ) , which decreased response rates. After 23 days of daily chlordiazepoxide administration (oral doses started at 10 and increased to 100 mg/kg/day by the 17th day), dose-effect curves for chlordiazepoxide were shifted to the right by about one-half log unit. Subjects were also more sensitive to the flumazenil, CGS 9895, and 6-CCtB, however, since these drugs produced only small effects in non-tolerant subjects, precise estimates of the degree of the shift in dose-effect curves could not be estimated. However, there were differences in the changes in the dose-effect curves induced by chlordiazepoxide tolerance. These results suggest differences in mechanism of action of antagonists in tolerant and non-tolerant subjects, and further that the s e n s i t i v i t y that is induced to antagonists in tolerant subjects is not conferred equally to all drugs having benzodiazepine antagonist a c t i v i t y . 1present address: CIEA Preclinical Research Laboratories, 1433 Nogawa, Miyamae, Kawasaki, Kanagawa, 213 Japan 2present address: Department of Applied Pharmacology, Hoshi University, School of Pharmacy, 2-4-41Ebara, Shinagawa-Ku, Tokyo, 142, Japan 3To whom reprint requests should be sent.
0024-3205/89 $3.00 + .00
290
Benzodlazeplne Antagonist Sensitivity
Vol. 44, No. 4, 1989
Several different benzodiazepine antagonists have been described recently, some of which exhibit d i f f e r e n t pharmacological actions. Antagonism of v i r t u a l l y all of the effects of benzodiazepine agonists has been reported with the imidazobenzodiazepinone, flumazenil (Ro15-1788). For example, Bonetti et al. (1) reported dose-related antagonism of the effects of diazepam including effects on locomotor behavior, rotarod performance, seizure a c t i v i t y , punished and avoidance responding. In addition tO antagonist a c t i v i t y , this compound also has some inverse agonist a c t i v i t y , although t y p i c a l l y less than that of other inverse agonists (2). For example, both flumazenil and the inverse agonist 3-hydroxymethyI-B-carboline stimulated cerebral blood flow and cerebral oxygen consumption (3, 4). Another antagonist, CGS 9895, a pyrazoloquinoline, has some agonist actions as well as antagonist effects. For example, at low doses CGS 9895 had anticonvulsant effects, whereas at higher doses the drug antagonized the anticonvulsant actions of diazepam, suggesting a partial agonist action (5). In further studies (6), CGS 9895 antagonized the effects of diazepam on rotarod performance in rats, and also antagonized the protection afforded by diazepam against pentylenetetrazole convulsions in mice. However, the antagonist did not block the increases in punished responding produced by diazepam. Several 6-carboline derivatives have also been studied for antagonist actions. The t - b u t y l - s u b s t i t u t e d 6-carboline, 3-carbo-t-butoxy-6-carboline (6-CCt8) antagonized the effects of diazepam on convulsions or punished responding without producing effects of its own. In contrast, the antagonist did not block the ataxia produced by diazepam as indicated by performance of mice climbing a wire screen (8). Based on these and previous (6) data, i t has been suggested that CGS 9895 and B-CCt8 antagonize benzodiazepine actions mediated at different subclasses of receptors (6, 8). One frequently studied 6-carboline has: been the ethyl ester, e t h y l - 6 carboline-3-carboxylate (6-CCE). Several actions of benzodiazepine agonists have been studied in combination with this B-carboline. For example, sedative effects of flurazepam (9) and anticonvulsant effects of diazepam (10) are antagonized by I3--CCE. In contrast to the effects of other antagonists, the antagonism of the effects of diazepam was obtained at doses of 6-CCE that had effects when given alone that were diametrically opposed to those of benzodiazepine agonists (11). In order to further characterize the actions of benzodiazepine antagonists, the present study was designed to examine the effects of a series of benzodiazepine antagonists in rats before and after the induction of tolerance to chlordiazepoxide. Chronic exposure to benzodiazepine agonists recently has been reported to a l t e r the effects of antagonists and inverse agonists (12). The effects on operant behavior of the r e l a t i v e l y pure antagonist, flumazenil, two B-carbolines, 6-CCtB and 6-CCE, and CGS 9895 were studied before and during the daily administration of chlordiazepoxide (100.0 mg/kg/day). Methods Subjects. Four experimentally-naive male Fischer F344 rats served as subjects. Each was six weeks of age at the s t a r t of the experiment and was for the f i r s t two weeks given unrestricted access to food. Subjects were then deprived of food for 24 hrs and were then trained to press the response key. Subsequently, a l l subjects were fed approximately 6 g per day of food p e l l e t s (#0173, Bioserv Inc., Frenchtown, NJ). Water was available at a l l times in the
Vol.
44, No. 4, 1989
Benzodtazepine Antagonist Sensitivity
291
home cages in which subjects were individually housed when experimental sessions were not in progress. These home cages were maintained in a temperature-controlled co|ony room which was on a 12-hr light-dark cycle ( l i g h t s were o~ from 07:00 to 19:00). Apparatus. Experimental chambers, with inside measures of 24.2 (w) x 24.1 ( I ) x 26.4 (h) cm, were constructed of aluminum Plexiglas®. On the front wall of each chamber, 3.8 cm from the grid floor and 7.0 cm left of the midline, was a response key (Model B, R. Gerbrands Co., Arlington MA). A press of the key, displacing it horizontally about 2 mm produced a feedback click of a relay mounted behind the front wall of the chamber and was counted as a response. The response key could be transilluminated by two 7.5 W a.c. green bulbs. Centered on the front wall, 1.9 cm from the floor, was a 3.8 (w) x 4.1 (h) cm opening through which 45 mg food pellets (#0021, Bioserv Inc.) could be presented via a pellet dispenser (Model D-l, R. Gerbrands Co.). Two stimulus lights (7.5 W, 120 V a . c . ) providing overall illumination were mounted above the front wall behind translucent panel. Each experimental chamber was contained within a ventilated outer shell (Model SEC-O02, BRS/LVE, Laurel, MD) to isolate i t from outside sound and l i g h t . The v e n t i l a t i o n fan provided masking noise. Illumination of the chamber, presentation of food, and recording of responses was accomplished by a PDP-11/73 Digital Equipment Corporation computer in an adjacent room. Procedure. Subjects were trained to respond under a f i x e d - i n t e r v a l schedule of food presentation; the f i r s t response after the lapse of each 2-min period produced food. Experimental sessions consisted of at least five components during which the chamber and response key were illuminated and the 2-min fixed-interval schedule was in effect. Each of these components lasted until five pellets had been delivered or 10.25 min had elapsed. Each was preceded by a timeout period that was at least 5.75 min in duration, during which the chamber and key lights were o f f and responses had no scheduled consequences. If the five food pellets were delivered before the lapse of 10.25 min, the duration of the timeout period was increased to maintain a constant time between successive fixed-interval 2-min components. Subjects were trained under this schedule u n t i l performances were stable, with l i t t l e or no responding during the timeout periods and with characteristic patterns of responding during the f i x e d - i n t e r v a l components. During sessions in which the effects of drugs were assessed, intraperitoneal ( i . p . ) injections were given at the s t a r t of each timeout period. A vehicle injection was administered before the f i r s t timeout period which served as a control observation. At the beginning of subsequent timeout periods, drugs were administered as cumulative doses. The absolute dose of drug administered per injection was calculated to sum with a l l previous doses to equal a dose ~ log unit greater than the previous cumulative dose. Each of the four antagonists were studied, followed by an assessment of the effects of chlordiazepoxide. Drugs w e r e administered at one-week intervals. Subsequently, a l l subjects were given daily oral doses of chlordiazepoxide by gavage at a dose of 10.0 mg/kg for seven days, 32.0 mg/kg for seven days, and f i n a l l y 100.0 mg/kg for the remainder of the experiment. Doses of chlordiazepoxide w e r e administered approximately 90 min before experimental sessions. After each seven days of daily oral chlordiazepoxide administration at each dose, the effects of cumulative i.p. doses of chlordiazepoxide were redetermined on the eighth day as described above. On those days, the presession administrations of ch[ordiazepoxide were omitted. Following the last redetermination of the effects of chlordiazepoxide, daily oral doses of 100.0 mg/kg chlordiazepoxide continued. The effects of the antagonists were redetermined, as before, at one-week intervals.
292
Benzodlazeplne Antagonist Sensitivity
Vol. 44, No. 4, 1989
Drugs. Chlordiazepoxide HCI (Hoffmann-La Roche, Nutley, NJ) and B-CCE were dissolved in 0.9~ NaCl. The 6-CCtB, flumazenil (Hoffmann-La Roche) and 2-(4-methoxy-phenyl)-pyrazolo[4,3-c]quinolin-3(5H)-one (CGS 9895, Ciba-Geigy Corp,, Summit, NJ) were suspended in water to which a few drops of Tween-80 were added for each 10 ml. Doses of chlordiazepoxide and 6-CCE were administered as the total weight of the HCl s a l t , whereas doses of the other drugs were administered as the weight of the base. Drug effects depicted in a l l figures are those obtained after i.p. injections. In order to induce tolerance, chlordiazepoxide was dissolved in d i s t i l l e d water and administered daily by gavage. O,~.¢.Bu I
O'-H I
~
c~ o
oH+
Figure I. 1
I It
2
Synthesis of 3-carbo-t-butoxy-6-carboline (B-CCtB) The synthesis of ~-CCtB (Figure 1, 2) was as follows. To a suspension of I3-carboline-3-carboxylic acid (Figure 1, 1) (1.50 g, 7.1 mmol) in dry dioxane (25 ml.) was added H2SO4 (2.5 mL, conc.). The mixture was then cooled in a dry ice bath (-78 ° ) and isobutylene was added (35 mL). The reaction was carried out in a Parr bottle (500 mL) and shaken in a Parr hydrogenation apparatus at room temperature for 3.5 h. The reaction vessel was then cooled to a s l u r r y , the stopper was c a r e f u l l y removed, and the solution poured onto a mixture of NaOH (20~, 200 mL); EtOAc (200 mL). The organic layer was separated and the aqueous layer extracted with EtOAc (2 x 200 mL). The organic layers were combined and dried (K2C03). The solvent was removed under reduced pressure at 30°C, to yield an o i l . Upon the addition of ether/hexane a precipitate formed to yield 2 (380 mg, 17~). The product was purified by repeated recrystalizations from ether. Imp 213°C (dec.), l i t . mp 219°C (13); IR (KBr) 3260, 2970, 1710 cm-l.] The pH of the alkaline aqueous layer was adjusted to 4 with HCi (2 N), at O°C. The resulting precipitate was collected by vacuum f i l t r a t i o n to yield recovered 1 (810 mg). Note that the ester 2 is extremely sensitive to heat, exposure to s i l i c a gel, and acid. Analysis of Results. Response rates were calculated by dividing total responses by elapsed time during the fixed-interval components for individual subjects. Response rates after the f i r s t component were expressed as a percentage of the rates that occurred during the i n i t i a l component. For comparisons of dose-effect curves, analysis of variance and linear regression techniques (14) were used to determine ED 50 values (the dose causing a decrease in response rates to 50 per cent of the control rate) and 95~ confidence l i m i t s . When antagonists were administered to subjects during the daily chlordiazepoxide treatment, grossly observable signs of withdrawal were noted. The signs observed were: vocalization during injection procedure, irritability (defined as attempts to escape during injection procedure or spontaneous vocalization or rearing before i n j e c t i o n ) . Sedation was noted i f ataxia or decreased general a c t i v i t y w e r e present. B o d y weights were determined d a i l y , and those obtained 24 hr after administration of each of the antagonists were compared before and during chlordiazepoxide treatment.
Vol. 44, No. 4, 1989
Benzodlazeplne Antagonist Sensitivity
293
Results Control performances were characterized by low rates of responding in the early portions of the fixed intervals followed by increasing rates as the intervals progressed ( c . f . 15), L i t t l e or no responding occurred during the timeout periods. Average rates of responding during a five-component session in which successive saline injections were administered decreased over the session by an average of less than 24~,
180 160 140 120 IO0 8O
I
60-
• DURING'rRF_ATMF_NTII
4020 0 I
1.0
I
I
I
I
10.0
I
I
I
I
100.0
I
I
I
560.0
CHLORDICff.EPOXIDE(MC~G) FIG. 2 Effects of chlordiazepoxide ( i . p . ) on schedule-controlled responding before and during the chronic oral administration of chlordiazepoxide at a dose of 100.0 mg/kg daily. Note that low to intermediate doses of chlordiazepoxide increased rates of responding and higher doses decreased response rates. During the chronic dosing, the doseeffect curve was shifted to the right. The ED 50 values for the linear portions of the dose-effect curves before and during chlordiazepoxide treatment were 95.0 mg/kg (95~ CL: 51.0 - 177.0 mg/kg) and 343.1 mg/kg (95~ CL: 1 1 4 . 8 - 1025.6 mg/kg), respectively.
294
Benzodlazeplne Antagonist Sensitivity
Flumazentl
1~o
Vol. 44, No. 4, 1989
8-CCtB
CGS-9895
8-CCE
~o
40
~o
,
I
,
o
3.2
10
32
1
3.2
10
32
!
i
i
i
1
3.2
10
32
1
3.2
10
32
DRuG DOSE (MC-#KG) FIG. 3 Effects of benzodiazepine antagonists (i.p.) on schedule-controlled responding before and during the chronic oral administration of 100 mg/kg chlordiazepoxide. Note that before chronic dosing flumazenii, CGS 9 8 9 5 and 6-CCtB were without effects up to doses of 10.0 mg/kg. During chronic dosing, each of the antagonists produced d o s e related decreases in response rates generally at a l l of the doses studied. The dose=effect curve for 6-CCE was affected minimally by the chronic administration of chlordiazepoxide. ED 50 values for 6-CCE before and during treatment were 33.1 mg/kg (95~ CL: 25.2 - 43.6 mg/kg) and 23.1 mg/kg (95~ CL: 15.0 35.7 mg/kg), respectively. TABLE I ED 50 values (95~ confidence l i m i t s ) for chlordiazepoxide (in mg/kg, i . p . ) induced decreases in response rates before and during chronic treatment with chlordiazepoxide. CONDITION Before COAP 101
ED 50 VALUE 95.0 (51.0 - 177.0) 182,8
(122.2 - 273.3) CDAP 322
175.5 (95.0 - 324.0)
CDAP 1003
343.1 (114.8 - 1025.6)
1E~ 50 value after 7 days of 10.0 mg/kg of oral chtordiazepoxide. 2EO 50 value after 7 days of 32.0 mg/kg of oral chlordiazepoxide, i r e c t l y following treatment with 10.0 mg/kg. ED 50 value after 7 days of 100.0 mg/kg of Qrai chlordiazepoxide, d i r e c t l y following treatment with 32.~ mg/kg.
~
100
Vol. 44, No. 4, 1989
Benzodlazeplne Antagonist Sensltlvlty
295
TABLE II ED 50 values (in mg/kg, i . p . ) and slopes of dose-effect curves before and during chronic oral treatment with chlordiazepoxide (100 mg/kg/day). Values in parentheses are 95~ confidence l i m i t s . DRUG
ED 50 VALUE Before During
SLOPE Before
During
Flumazenil
NS
14.7 (6.0 - 35.9)
NS
-25.8 (-40.2 - -11.4)
CGS 9895
NS
7.4 (3.4 - 16.0)
NS
-87.7 (-119.0 - -56.3)
~-CCtB
NS
11.7 (2.9 - 46.9)
NS
-32.1 (-62.9 - - 2.0)
B-CCE 33.1 (25.2 43.6) -
23.1 (15.0 - 35.7)
-41.6 -61.8 (-50.6 - -32.6) (-74.8 - -48.8)
TABLE I I I Weight change as a percentage of body weight (± SD) determined 24 hr a f t e r i n j e c t i o n of each antagonist. Flumazenil
CGS 9895
~-CCtB
6-CCE
Before CDAP
+0.15 (0.83)
+1.05 (0.90)
+0.68 (1.35)
+3.00 (0.36)
During CDAP
-1.55 (1.14)
-1.85 (0.64)
-2.63 (1.21)
-2.10 (0.29)
Effects of chlordiazepoxide are shown in figure 2. Before i t s d a i l y administration, chlordiazepoxide (open symbols) increased response rates above control levels at doses from 1.0 to 10.0 mg/kg. Maximal increases were obtained at a dose of 10.0 mg/kg. Higher doses increased rates of responding less or decreased rates. The ED 50 value for chlordiazepoxide, calculated from the linear portion of the dose-effect curve (10.0 to 100.0 mg/kg), was 41.5 mg/kg (Table I ) . The ED 50 values generally increased with each seven days of treatment with increasing doses of chlordiazepoxide. The ED 50 value was 343.1 mg/kg a f t e r seven days of treatment with 100,0 mg/kg. Each increase in ED 50 valuewas s i g n i f i c a n t when compared to before treatment, with the exception of 32.0 mg/kg (95~ confidence i n t e r v a l s overlapped). The dose-effect curve obtained a f t e r seven days of treatment with 100.0 mg/kg is also shown in figure 2 ( f i l l e d symbols). Doses of 10.0 and 32.0 mg/kg increased response rates with higher doses decreasing those rates. The dose-effect curve was shifted by daily treatment to the right by approximately one-half log unit (approximately three f o l d ) .
296
Benzodlazeplne Antagonist Sensitivity
Vol. 44, No. 4, 1989
Before daily treatment with chlordiazepoxide, none of the antagonists had pronounced effects on rates of responding up to doses of 32.0 mg/kg (figure 3; open symbols). Decreases in rates to 77~ of control were obtained at 32 mg/kg of flumazenil. At the same dose CGS 9895 decreased response rates to 61~ of control, whereas rates of about 80~ of control were obtained at 10.0 and 32.0 mg/kg of 6-CCt8. Since effects were minimal over the dose ranges studied, EO 50 values could not be determined for these drugs. Only 6-CCE was studied at higher doses, lith this compound, dose-related decreases in rates of responding were obtained generally at doses greater than 1.0 mg/kg. The ED 50 value for 6-CCE was 33.1 mg/kg (95~ CL: 25.2 - 43.6 mg/kg). During treatment with 100.0 mg/kg of chlordiazepoxide, all of the antagonists produced significant decreases in response rate. ED 50 values for all of the drugs are shown in Table I I . Flumazenil produced dose-related decreases in response rates at doses from 1.0 to 32.0 mg/kg (figure 3; f i l l e d symbols). The 1.0 mg/kg dose of flumazenil decreased response rates comparably to 32.0 mg/kg before treatment. CGS 9895 also produced doserelated decreases in response rates, however, the slope of the dose-effect curve was steeper than that obtained with flumazenil (figure 3; Table I). The 10.0 mg/kg dose of CGS 9895 decreased response rates to a greater extent than 32.0 mg/kg before treatment. B-CCtB produced dose-related decreases in response rates at doses from 1.0 to 32.0 mg/kg, with the slope of the dose-effect curve similar to that obtained with flumazenil (Table I I ) . The 3.2 mg/kg dose decreased response rates to a greater extent than did 32.0 mg/kg before treatment. The dose-effect curve for B-CCE did not change as much as did the curves for the other antagonists during daily chlordiazepoxide treatment (Figure 3). At 32.0 mg/kg, 6-CCE had greater effects than before treatment, however, the ED 50 values before and during treatment were not appreciably different and the 95~ confidence l i m i t s overlapped (Table I I ) . The slopes of the 6-CCE dose-effect curves were most similar to those obtained with CGS 9895. During the administration of the antagonists, subjects were observed for any grossly detectable signs of withdrawal during handling and injections. With flumazenil, subjects exhibited vocalizations and i r r i t a b i l i t y at the higher two doses. None of the other antagonists produced similar signs. At the highest doses of CGS 9895, all of the subjects exhibited signs of sedation. All of the antagonists produced a loss of body weight 24 hr after their administration during chlordiazepoxide treatment. In contrast there was no weight loss associated with antagonist administration before chlordiazepoxide treatment (Table I I I ) . Discussion In the present study, tolerance developed to the effects of chlordiazepoxide on schedule-controlled behavior. Once subjects were tolerant, a l l of the benzodiazepine antagonists produced dose-related decreases in rates of responding maintained by food presentation. Previous to the development of tolerance, only 6-CCE decreased rates of responding. Since higher doses of 6-CCE were studied than of the other antagonists, it is possible that the other antagonists would also have been effective in non-tolerant subjects. Further studies with the other antagonists, assessing effects over a wider dose-range, w i l l more adequately characterize their behavioral effects in non-tolerant subjects.
Vol. 44, No. 4, 1989
Benzodtazepine Antasonist S e n s i t i v i t y
297
Regardless~of whether flumazenil, CGS 9895, and ~-CCtB have etfects in non-tolerant subjects, each of these drugs e f f e c t i v e l y decreased response rates in tolerant subjects at doses that had l i t t l e effect in non-tolerant subjects. Thus, exposure to the agonist chlordiazepoxide produced large changes in the effects of these drugs. Those changes in effects were different from effects obtained with the inverse agonist, i~-CCE. Once subjects were rendered tolerant to chlordiazepoxide, the effects of this drug on schedule-controlled responding were not appreciably d i f f e r e n t from those obtained prior to the induction of tolerance. The present findings can be compared to those obtained previously in a study of the proconvulsant effects determined by assessing threshold for pentylenetetrazol-induced siezures (12). In that study, chronic treatment for 14 days with the benzodiazepine agonist, Iorazepam, increased the proconvulsant effectiveness of flumazenil and the antagonist ZK 93426, but did not a,lter the effects of the inverse agonists, FG 7142 and DMCM, on pentylenetetrazol siezure threshold. Thus, as in the present study, the smallest changes in the effects of these drugs induced with agonist treatment, occurred with the drugs having the greatest inverse agonist i n t r i n s i c a c t i v i t y . The present results are similar in many respects to those obtained with opioid antagonists. In several studies, including studies on operant behavior, s e n s i t i v i t y to opioid-antagonists in naive subjects is appreciably less than that observed in opioid-dependent subjects, and the antagonist dose-effect curve is shifted to the l e f t (e.9. 16, 17, 18). In the present study, the s h i f t s to the left of the antagonist dose-effect curves could not be estimated since the highest doses studied produced only small decreases in response rates. However, s h i f t s to the l e f t of the dose-effect curves were clearly evident since the antagonists had effects in the subjects, once tolerant, that did not occur in the subjects prior to the induction of tolerance. Since each of the antagonists had l i t t l e or no i n t r i n s i c a c t i v i t y in non-tolerant subjects, their actions in tolerant subjects is l i k e l y due to the displacement of agonist from receptor sites and suggests the development of physiological dependence on chlordiazepoxide. Several other results suggest that dependence accompanied the tolerance to effects of chlordiazepoxide in the present study. The loss in body weight produced by each of the antagonists of the present study is similar to, though of lesser magnitude than, that observed with cessation of chlordiazepoxide (19, 20) and opioid treatment (21). Several of the signs observed following antagonist administration have b e e n characteristic of withdrawal from chlordiazepoxide or diazepam as described by others (20, 22). In addition, Boisse et al. (23) have indicated development of dependence to chlordiazepoxide in rats at lower doses than those administered in the present study. However, in that study chlordiazepoxide was administered more frequently. Regardless, the present dosage regimen and duration is within a range that would be expected to produce physiological dependence (for a review see 24). Several papers have suggested that the repeated administration of agonist changes the efficacy of benzodiazepine-receptor ligands towards inverseagonist a c t i v i t y . For example, in the study by Petersen and Jensen (12), the maximal effect of Iorazepam was decreased when subjects were rendered tolerant to Iorazepam. Further, antagonists with l i t t l e efficacy in non-tolerant subjects had inverse-agonist a c t i v i t y in tolerant subjects. In many cases, however, increased inverse agonist efficacy may be d i f f i c u l t to d i s t i n g u i s h from precipitated withdrawal due to displacement of agonist from sites of action. For example, the present decreases in operant behavior may be due to either of these actions. S i m i l a r l y , certain grossly observable behaviors may also be due to either of these effects.
298
BenzodlazepineAntagonist Sensitivity
Vol. 44, No. 4, 1989
In the present study, none of the antagonists, including the inverse agonist B-CCE, produced weight loss in non-tolerant subjects, whereas each of the drugs produced s i g n i f i c a n t weight loss in the subjects when chlordiazepoxide tolerant. The weight loss suggests that physiological dependence on chlordiazepoxide had developed in these subjects. That weight loss did not follow the administration of B-CCE in non-tolerant subjects suggests that not all of the effects of antagonists in tolerant subjects are due to a change in inverse agonist efficacy occuring with repeated administration of agonist. Although flumazenil, CGS 9895 and 6-CCtB each precipitated some signs of withdrawal in tolerant subjects, there were differences in the effects of the antagonists. For example, the dose-effect curves for flumazenil and ~-CCtB were similar, whereas the dose-effect curve for CGS 9895 had a r e l a t i v e l y greater slope. In addition, only flumazenil produced grossly observable behavioral signs of withdrawal. The effects of flumazenil also differed from those of the other antagonists in that injection of this compound in tolerant subjects produced vocalization and i r r i t a b i l i t y on handling. None of the other antagonists produced these effects. Since not a l l of the antagonists produced these effects, i t is possible that the actions of these antagonists are mediated by different mechanisms. ACKNOWLEDGEMENTS This research was supported in part by NIH Grant NS-22287. We thank Jose A. Prada and Richard Rodriguez for their expert assistance in the conduct of these studies, Dr. D. B. Vaupel for s t a t i s t i c a l advice, and Dr. R. R. G r i f f i t h s for the loan of CGS 9895. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
E.P. BONETTI, L. PIERI, R. CUMIN, R. SCHAFFNER, M. PIERI, E.R. GAMZU, R.K.M. MULLER, W. HAEFELY, Psychopharmacology 78:8-18 (1982). S.E. FILE, S. PELLOW, Psychopharmacology 88:1-11 (1986). R.F. ALBRECHT, J. COOK, W.E. HOFFMAN, P. LARSCHEID, D.J. MILETICH, N. NAUGHTON, Neuropharmacology 24:957-963 (1985). W.E. HOFFMAN, R.F. ALBRECHT, D.J. MILETICH, T.J. HAGEN, J.M. COOK, Anesth. Analg. 65:639-644 (1986). N. YOKOYAMA, B. RITTER, A.D. NEUBERT, J. Med. Chem. 25:337-339 (1982). N.J. KATZMAN, H.E. SHANNON, J. Pharmacol. Exp. Ther. 235:589-595 (1985). P.N. VAN GORDER, W.E. HOFFMAN, V. BAUGHMAN, R.F. ALBRECHT, D.J. MILETICH, F. GUZMAN, J.M. COOK, Anesth. Analg. 64:129-135 (1985). H.E. SHANNON, F. GUZMAN, J.M. COOK, Life Sci. 35:2227-2236 (1984). P.J. COWEN, A.R. GREEN, D.J. NUTT, I.L. MARTIN, Nature 290:54-55 (1981). N.R. OAKLEY, B.J. JONES, Eur. J. Pharmacol. 68:381-382 (1980). P. POLC, E.P. BONETTI, R. SCHAFFNER, W. HAEFELY, Naunyn Schmeideberg's Arch. Pharmacol. 321:260-264 (1982). E.N. PETERSEN, L.H. JENSEN, Eur. J. Pharmacol. 133:309-317 (1987). T.J. HAGEN, F. GUZMAN, C. SCHULTZ, J.M. COOK, P. SKOLNICK, H.E. SHANNON, Heterocycles 24:2845-2855 (1986). G.W. SNEDECOR, W.G. COCHRAN, S t a t i s t i c a l Methods. 6th ed. pp. 135-171, Iowa State University Press, Ames Iowa (1967).
Vol. 44, No. 4, 1989
Benzodlazeplne Antagonist Sensitivity
299
15. C.B. FERSTER, B.F. SKINNER, Schecl~les of Reinforcement. AppletonCentury-Crofts, New York (~ST) 16. R.J. VALENTINO, S. HERLING, J.H. WOODS, J. Pharmacol. Exp. Ther. 224: 307-313 (1983). 17. J.H. WOODS, J. CARNEY, In Behavioral Tolerance: Research and Treatment Implications, N.A. Krasnegor (ed) NIDA Research Monograph Series, Vol. 19, U.S. Gov't Printing Office, Washington, D.C., pp. 54-66, (1978) 18 J.L. KATZ, Psychopharmacology, 89: 323-326, 1986. 19. S. YANAURA, E. TAGASHIRA, Folia Pharmacol. Japon 71:157-168 (1975). 20. G.P. RYAN, N.R. BOISSE, J. Pharmacol. Exp. Ther. 226:100-107 (1983). 21. E. WEI, H.H. LOH, E.L. WAY, J. Pharmacol. Exp. Ther. 184:398-403 (1973). 22. W.R. MARTIN, L.F. McNICHOLAS, S. CHERIAN, Life Sci. 31:721-730 (1982). 23. N.R. BOISSE, G.P. RYAN, J.J. GUARINO, Problems of Drug Dependence, 1981. L.S. ,Harris (ed) NIDA Monograph 41, U.S. Government Printing Office, Washington OC (1982). 24. J.H. WOODS, J.L. KATZ, G. WINGER, Pharmacol. Rev. 39: 251-413, 1987.
Pergamon Press
BEHAVIORAL EFFECTSOF BENZODIAZEPINEANTAGONISTS IN CHLORDIAZEPOXIDE TOLERANT ANO NON-TOLERANTRATS K. Takada1., T. Suzuki2., T. Hagen** J.M. Cook** and J.L. Katz 3. *** ' *NIDA Addiction Research Center, Baltimore, Maryland 21224 **Department of Chemistry, University of Wisconsin, Milwaukee, Wisconsin 53201 ***Department of Pharmacology and Experimental Therapeutics University of Maryland School of Medicine Baltimore, Maryland 21201 (Received in f£nal form December 1, 1988)
Summary Rats were trained to respond under 3-min fixed-interval schedules of food presentation, and effects of the benzodiazepine-receptor ligands, flumazenil, 2-(4-methoxy-phenyl)-pyrazolo[4,3-c]quinolin3(5H)-one (CGS 9895), 3-carbo-t-butoxy-6-carboline (6-CCtB), and 6-carboline-3-carboxylic acid ethyl ester (13-CCE) were assessed before and after the induction of tolerance to chlordiazepoxide. Before daily administration of chlordiazepoxide, none of the antagonists produced appreciable effects on rates of responding up to doses of 32.0 mg/kg i.p. 6-CCE was the only antagonist studied at a higher dose (100.0 mg/kg i . p . ) , which decreased response rates. After 23 days of daily chlordiazepoxide administration (oral doses started at 10 and increased to 100 mg/kg/day by the 17th day), dose-effect curves for chlordiazepoxide were shifted to the right by about one-half log unit. Subjects were also more sensitive to the flumazenil, CGS 9895, and 6-CCtB, however, since these drugs produced only small effects in non-tolerant subjects, precise estimates of the degree of the shift in dose-effect curves could not be estimated. However, there were differences in the changes in the dose-effect curves induced by chlordiazepoxide tolerance. These results suggest differences in mechanism of action of antagonists in tolerant and non-tolerant subjects, and further that the s e n s i t i v i t y that is induced to antagonists in tolerant subjects is not conferred equally to all drugs having benzodiazepine antagonist a c t i v i t y . 1present address: CIEA Preclinical Research Laboratories, 1433 Nogawa, Miyamae, Kawasaki, Kanagawa, 213 Japan 2present address: Department of Applied Pharmacology, Hoshi University, School of Pharmacy, 2-4-41Ebara, Shinagawa-Ku, Tokyo, 142, Japan 3To whom reprint requests should be sent.
0024-3205/89 $3.00 + .00
290
Benzodlazeplne Antagonist Sensitivity
Vol. 44, No. 4, 1989
Several different benzodiazepine antagonists have been described recently, some of which exhibit d i f f e r e n t pharmacological actions. Antagonism of v i r t u a l l y all of the effects of benzodiazepine agonists has been reported with the imidazobenzodiazepinone, flumazenil (Ro15-1788). For example, Bonetti et al. (1) reported dose-related antagonism of the effects of diazepam including effects on locomotor behavior, rotarod performance, seizure a c t i v i t y , punished and avoidance responding. In addition tO antagonist a c t i v i t y , this compound also has some inverse agonist a c t i v i t y , although t y p i c a l l y less than that of other inverse agonists (2). For example, both flumazenil and the inverse agonist 3-hydroxymethyI-B-carboline stimulated cerebral blood flow and cerebral oxygen consumption (3, 4). Another antagonist, CGS 9895, a pyrazoloquinoline, has some agonist actions as well as antagonist effects. For example, at low doses CGS 9895 had anticonvulsant effects, whereas at higher doses the drug antagonized the anticonvulsant actions of diazepam, suggesting a partial agonist action (5). In further studies (6), CGS 9895 antagonized the effects of diazepam on rotarod performance in rats, and also antagonized the protection afforded by diazepam against pentylenetetrazole convulsions in mice. However, the antagonist did not block the increases in punished responding produced by diazepam. Several 6-carboline derivatives have also been studied for antagonist actions. The t - b u t y l - s u b s t i t u t e d 6-carboline, 3-carbo-t-butoxy-6-carboline (6-CCt8) antagonized the effects of diazepam on convulsions or punished responding without producing effects of its own. In contrast, the antagonist did not block the ataxia produced by diazepam as indicated by performance of mice climbing a wire screen (8). Based on these and previous (6) data, i t has been suggested that CGS 9895 and B-CCt8 antagonize benzodiazepine actions mediated at different subclasses of receptors (6, 8). One frequently studied 6-carboline has: been the ethyl ester, e t h y l - 6 carboline-3-carboxylate (6-CCE). Several actions of benzodiazepine agonists have been studied in combination with this B-carboline. For example, sedative effects of flurazepam (9) and anticonvulsant effects of diazepam (10) are antagonized by I3--CCE. In contrast to the effects of other antagonists, the antagonism of the effects of diazepam was obtained at doses of 6-CCE that had effects when given alone that were diametrically opposed to those of benzodiazepine agonists (11). In order to further characterize the actions of benzodiazepine antagonists, the present study was designed to examine the effects of a series of benzodiazepine antagonists in rats before and after the induction of tolerance to chlordiazepoxide. Chronic exposure to benzodiazepine agonists recently has been reported to a l t e r the effects of antagonists and inverse agonists (12). The effects on operant behavior of the r e l a t i v e l y pure antagonist, flumazenil, two B-carbolines, 6-CCtB and 6-CCE, and CGS 9895 were studied before and during the daily administration of chlordiazepoxide (100.0 mg/kg/day). Methods Subjects. Four experimentally-naive male Fischer F344 rats served as subjects. Each was six weeks of age at the s t a r t of the experiment and was for the f i r s t two weeks given unrestricted access to food. Subjects were then deprived of food for 24 hrs and were then trained to press the response key. Subsequently, a l l subjects were fed approximately 6 g per day of food p e l l e t s (#0173, Bioserv Inc., Frenchtown, NJ). Water was available at a l l times in the
Vol.
44, No. 4, 1989
Benzodtazepine Antagonist Sensitivity
291
home cages in which subjects were individually housed when experimental sessions were not in progress. These home cages were maintained in a temperature-controlled co|ony room which was on a 12-hr light-dark cycle ( l i g h t s were o~ from 07:00 to 19:00). Apparatus. Experimental chambers, with inside measures of 24.2 (w) x 24.1 ( I ) x 26.4 (h) cm, were constructed of aluminum Plexiglas®. On the front wall of each chamber, 3.8 cm from the grid floor and 7.0 cm left of the midline, was a response key (Model B, R. Gerbrands Co., Arlington MA). A press of the key, displacing it horizontally about 2 mm produced a feedback click of a relay mounted behind the front wall of the chamber and was counted as a response. The response key could be transilluminated by two 7.5 W a.c. green bulbs. Centered on the front wall, 1.9 cm from the floor, was a 3.8 (w) x 4.1 (h) cm opening through which 45 mg food pellets (#0021, Bioserv Inc.) could be presented via a pellet dispenser (Model D-l, R. Gerbrands Co.). Two stimulus lights (7.5 W, 120 V a . c . ) providing overall illumination were mounted above the front wall behind translucent panel. Each experimental chamber was contained within a ventilated outer shell (Model SEC-O02, BRS/LVE, Laurel, MD) to isolate i t from outside sound and l i g h t . The v e n t i l a t i o n fan provided masking noise. Illumination of the chamber, presentation of food, and recording of responses was accomplished by a PDP-11/73 Digital Equipment Corporation computer in an adjacent room. Procedure. Subjects were trained to respond under a f i x e d - i n t e r v a l schedule of food presentation; the f i r s t response after the lapse of each 2-min period produced food. Experimental sessions consisted of at least five components during which the chamber and response key were illuminated and the 2-min fixed-interval schedule was in effect. Each of these components lasted until five pellets had been delivered or 10.25 min had elapsed. Each was preceded by a timeout period that was at least 5.75 min in duration, during which the chamber and key lights were o f f and responses had no scheduled consequences. If the five food pellets were delivered before the lapse of 10.25 min, the duration of the timeout period was increased to maintain a constant time between successive fixed-interval 2-min components. Subjects were trained under this schedule u n t i l performances were stable, with l i t t l e or no responding during the timeout periods and with characteristic patterns of responding during the f i x e d - i n t e r v a l components. During sessions in which the effects of drugs were assessed, intraperitoneal ( i . p . ) injections were given at the s t a r t of each timeout period. A vehicle injection was administered before the f i r s t timeout period which served as a control observation. At the beginning of subsequent timeout periods, drugs were administered as cumulative doses. The absolute dose of drug administered per injection was calculated to sum with a l l previous doses to equal a dose ~ log unit greater than the previous cumulative dose. Each of the four antagonists were studied, followed by an assessment of the effects of chlordiazepoxide. Drugs w e r e administered at one-week intervals. Subsequently, a l l subjects were given daily oral doses of chlordiazepoxide by gavage at a dose of 10.0 mg/kg for seven days, 32.0 mg/kg for seven days, and f i n a l l y 100.0 mg/kg for the remainder of the experiment. Doses of chlordiazepoxide w e r e administered approximately 90 min before experimental sessions. After each seven days of daily oral chlordiazepoxide administration at each dose, the effects of cumulative i.p. doses of chlordiazepoxide were redetermined on the eighth day as described above. On those days, the presession administrations of ch[ordiazepoxide were omitted. Following the last redetermination of the effects of chlordiazepoxide, daily oral doses of 100.0 mg/kg chlordiazepoxide continued. The effects of the antagonists were redetermined, as before, at one-week intervals.
292
Benzodlazeplne Antagonist Sensitivity
Vol. 44, No. 4, 1989
Drugs. Chlordiazepoxide HCI (Hoffmann-La Roche, Nutley, NJ) and B-CCE were dissolved in 0.9~ NaCl. The 6-CCtB, flumazenil (Hoffmann-La Roche) and 2-(4-methoxy-phenyl)-pyrazolo[4,3-c]quinolin-3(5H)-one (CGS 9895, Ciba-Geigy Corp,, Summit, NJ) were suspended in water to which a few drops of Tween-80 were added for each 10 ml. Doses of chlordiazepoxide and 6-CCE were administered as the total weight of the HCl s a l t , whereas doses of the other drugs were administered as the weight of the base. Drug effects depicted in a l l figures are those obtained after i.p. injections. In order to induce tolerance, chlordiazepoxide was dissolved in d i s t i l l e d water and administered daily by gavage. O,~.¢.Bu I
O'-H I
~
c~ o
oH+
Figure I. 1
I It
2
Synthesis of 3-carbo-t-butoxy-6-carboline (B-CCtB) The synthesis of ~-CCtB (Figure 1, 2) was as follows. To a suspension of I3-carboline-3-carboxylic acid (Figure 1, 1) (1.50 g, 7.1 mmol) in dry dioxane (25 ml.) was added H2SO4 (2.5 mL, conc.). The mixture was then cooled in a dry ice bath (-78 ° ) and isobutylene was added (35 mL). The reaction was carried out in a Parr bottle (500 mL) and shaken in a Parr hydrogenation apparatus at room temperature for 3.5 h. The reaction vessel was then cooled to a s l u r r y , the stopper was c a r e f u l l y removed, and the solution poured onto a mixture of NaOH (20~, 200 mL); EtOAc (200 mL). The organic layer was separated and the aqueous layer extracted with EtOAc (2 x 200 mL). The organic layers were combined and dried (K2C03). The solvent was removed under reduced pressure at 30°C, to yield an o i l . Upon the addition of ether/hexane a precipitate formed to yield 2 (380 mg, 17~). The product was purified by repeated recrystalizations from ether. Imp 213°C (dec.), l i t . mp 219°C (13); IR (KBr) 3260, 2970, 1710 cm-l.] The pH of the alkaline aqueous layer was adjusted to 4 with HCi (2 N), at O°C. The resulting precipitate was collected by vacuum f i l t r a t i o n to yield recovered 1 (810 mg). Note that the ester 2 is extremely sensitive to heat, exposure to s i l i c a gel, and acid. Analysis of Results. Response rates were calculated by dividing total responses by elapsed time during the fixed-interval components for individual subjects. Response rates after the f i r s t component were expressed as a percentage of the rates that occurred during the i n i t i a l component. For comparisons of dose-effect curves, analysis of variance and linear regression techniques (14) were used to determine ED 50 values (the dose causing a decrease in response rates to 50 per cent of the control rate) and 95~ confidence l i m i t s . When antagonists were administered to subjects during the daily chlordiazepoxide treatment, grossly observable signs of withdrawal were noted. The signs observed were: vocalization during injection procedure, irritability (defined as attempts to escape during injection procedure or spontaneous vocalization or rearing before i n j e c t i o n ) . Sedation was noted i f ataxia or decreased general a c t i v i t y w e r e present. B o d y weights were determined d a i l y , and those obtained 24 hr after administration of each of the antagonists were compared before and during chlordiazepoxide treatment.
Vol. 44, No. 4, 1989
Benzodlazeplne Antagonist Sensitivity
293
Results Control performances were characterized by low rates of responding in the early portions of the fixed intervals followed by increasing rates as the intervals progressed ( c . f . 15), L i t t l e or no responding occurred during the timeout periods. Average rates of responding during a five-component session in which successive saline injections were administered decreased over the session by an average of less than 24~,
180 160 140 120 IO0 8O
I
60-
• DURING'rRF_ATMF_NTII
4020 0 I
1.0
I
I
I
I
10.0
I
I
I
I
100.0
I
I
I
560.0
CHLORDICff.EPOXIDE(MC~G) FIG. 2 Effects of chlordiazepoxide ( i . p . ) on schedule-controlled responding before and during the chronic oral administration of chlordiazepoxide at a dose of 100.0 mg/kg daily. Note that low to intermediate doses of chlordiazepoxide increased rates of responding and higher doses decreased response rates. During the chronic dosing, the doseeffect curve was shifted to the right. The ED 50 values for the linear portions of the dose-effect curves before and during chlordiazepoxide treatment were 95.0 mg/kg (95~ CL: 51.0 - 177.0 mg/kg) and 343.1 mg/kg (95~ CL: 1 1 4 . 8 - 1025.6 mg/kg), respectively.
294
Benzodlazeplne Antagonist Sensitivity
Flumazentl
1~o
Vol. 44, No. 4, 1989
8-CCtB
CGS-9895
8-CCE
~o
40
~o
,
I
,
o
3.2
10
32
1
3.2
10
32
!
i
i
i
1
3.2
10
32
1
3.2
10
32
DRuG DOSE (MC-#KG) FIG. 3 Effects of benzodiazepine antagonists (i.p.) on schedule-controlled responding before and during the chronic oral administration of 100 mg/kg chlordiazepoxide. Note that before chronic dosing flumazenii, CGS 9 8 9 5 and 6-CCtB were without effects up to doses of 10.0 mg/kg. During chronic dosing, each of the antagonists produced d o s e related decreases in response rates generally at a l l of the doses studied. The dose=effect curve for 6-CCE was affected minimally by the chronic administration of chlordiazepoxide. ED 50 values for 6-CCE before and during treatment were 33.1 mg/kg (95~ CL: 25.2 - 43.6 mg/kg) and 23.1 mg/kg (95~ CL: 15.0 35.7 mg/kg), respectively. TABLE I ED 50 values (95~ confidence l i m i t s ) for chlordiazepoxide (in mg/kg, i . p . ) induced decreases in response rates before and during chronic treatment with chlordiazepoxide. CONDITION Before COAP 101
ED 50 VALUE 95.0 (51.0 - 177.0) 182,8
(122.2 - 273.3) CDAP 322
175.5 (95.0 - 324.0)
CDAP 1003
343.1 (114.8 - 1025.6)
1E~ 50 value after 7 days of 10.0 mg/kg of oral chtordiazepoxide. 2EO 50 value after 7 days of 32.0 mg/kg of oral chlordiazepoxide, i r e c t l y following treatment with 10.0 mg/kg. ED 50 value after 7 days of 100.0 mg/kg of Qrai chlordiazepoxide, d i r e c t l y following treatment with 32.~ mg/kg.
~
100
Vol. 44, No. 4, 1989
Benzodlazeplne Antagonist Sensltlvlty
295
TABLE II ED 50 values (in mg/kg, i . p . ) and slopes of dose-effect curves before and during chronic oral treatment with chlordiazepoxide (100 mg/kg/day). Values in parentheses are 95~ confidence l i m i t s . DRUG
ED 50 VALUE Before During
SLOPE Before
During
Flumazenil
NS
14.7 (6.0 - 35.9)
NS
-25.8 (-40.2 - -11.4)
CGS 9895
NS
7.4 (3.4 - 16.0)
NS
-87.7 (-119.0 - -56.3)
~-CCtB
NS
11.7 (2.9 - 46.9)
NS
-32.1 (-62.9 - - 2.0)
B-CCE 33.1 (25.2 43.6) -
23.1 (15.0 - 35.7)
-41.6 -61.8 (-50.6 - -32.6) (-74.8 - -48.8)
TABLE I I I Weight change as a percentage of body weight (± SD) determined 24 hr a f t e r i n j e c t i o n of each antagonist. Flumazenil
CGS 9895
~-CCtB
6-CCE
Before CDAP
+0.15 (0.83)
+1.05 (0.90)
+0.68 (1.35)
+3.00 (0.36)
During CDAP
-1.55 (1.14)
-1.85 (0.64)
-2.63 (1.21)
-2.10 (0.29)
Effects of chlordiazepoxide are shown in figure 2. Before i t s d a i l y administration, chlordiazepoxide (open symbols) increased response rates above control levels at doses from 1.0 to 10.0 mg/kg. Maximal increases were obtained at a dose of 10.0 mg/kg. Higher doses increased rates of responding less or decreased rates. The ED 50 value for chlordiazepoxide, calculated from the linear portion of the dose-effect curve (10.0 to 100.0 mg/kg), was 41.5 mg/kg (Table I ) . The ED 50 values generally increased with each seven days of treatment with increasing doses of chlordiazepoxide. The ED 50 value was 343.1 mg/kg a f t e r seven days of treatment with 100,0 mg/kg. Each increase in ED 50 valuewas s i g n i f i c a n t when compared to before treatment, with the exception of 32.0 mg/kg (95~ confidence i n t e r v a l s overlapped). The dose-effect curve obtained a f t e r seven days of treatment with 100.0 mg/kg is also shown in figure 2 ( f i l l e d symbols). Doses of 10.0 and 32.0 mg/kg increased response rates with higher doses decreasing those rates. The dose-effect curve was shifted by daily treatment to the right by approximately one-half log unit (approximately three f o l d ) .
296
Benzodlazeplne Antagonist Sensitivity
Vol. 44, No. 4, 1989
Before daily treatment with chlordiazepoxide, none of the antagonists had pronounced effects on rates of responding up to doses of 32.0 mg/kg (figure 3; open symbols). Decreases in rates to 77~ of control were obtained at 32 mg/kg of flumazenil. At the same dose CGS 9895 decreased response rates to 61~ of control, whereas rates of about 80~ of control were obtained at 10.0 and 32.0 mg/kg of 6-CCt8. Since effects were minimal over the dose ranges studied, EO 50 values could not be determined for these drugs. Only 6-CCE was studied at higher doses, lith this compound, dose-related decreases in rates of responding were obtained generally at doses greater than 1.0 mg/kg. The ED 50 value for 6-CCE was 33.1 mg/kg (95~ CL: 25.2 - 43.6 mg/kg). During treatment with 100.0 mg/kg of chlordiazepoxide, all of the antagonists produced significant decreases in response rate. ED 50 values for all of the drugs are shown in Table I I . Flumazenil produced dose-related decreases in response rates at doses from 1.0 to 32.0 mg/kg (figure 3; f i l l e d symbols). The 1.0 mg/kg dose of flumazenil decreased response rates comparably to 32.0 mg/kg before treatment. CGS 9895 also produced doserelated decreases in response rates, however, the slope of the dose-effect curve was steeper than that obtained with flumazenil (figure 3; Table I). The 10.0 mg/kg dose of CGS 9895 decreased response rates to a greater extent than 32.0 mg/kg before treatment. B-CCtB produced dose-related decreases in response rates at doses from 1.0 to 32.0 mg/kg, with the slope of the dose-effect curve similar to that obtained with flumazenil (Table I I ) . The 3.2 mg/kg dose decreased response rates to a greater extent than did 32.0 mg/kg before treatment. The dose-effect curve for B-CCE did not change as much as did the curves for the other antagonists during daily chlordiazepoxide treatment (Figure 3). At 32.0 mg/kg, 6-CCE had greater effects than before treatment, however, the ED 50 values before and during treatment were not appreciably different and the 95~ confidence l i m i t s overlapped (Table I I ) . The slopes of the 6-CCE dose-effect curves were most similar to those obtained with CGS 9895. During the administration of the antagonists, subjects were observed for any grossly detectable signs of withdrawal during handling and injections. With flumazenil, subjects exhibited vocalizations and i r r i t a b i l i t y at the higher two doses. None of the other antagonists produced similar signs. At the highest doses of CGS 9895, all of the subjects exhibited signs of sedation. All of the antagonists produced a loss of body weight 24 hr after their administration during chlordiazepoxide treatment. In contrast there was no weight loss associated with antagonist administration before chlordiazepoxide treatment (Table I I I ) . Discussion In the present study, tolerance developed to the effects of chlordiazepoxide on schedule-controlled behavior. Once subjects were tolerant, a l l of the benzodiazepine antagonists produced dose-related decreases in rates of responding maintained by food presentation. Previous to the development of tolerance, only 6-CCE decreased rates of responding. Since higher doses of 6-CCE were studied than of the other antagonists, it is possible that the other antagonists would also have been effective in non-tolerant subjects. Further studies with the other antagonists, assessing effects over a wider dose-range, w i l l more adequately characterize their behavioral effects in non-tolerant subjects.
Vol. 44, No. 4, 1989
Benzodtazepine Antasonist S e n s i t i v i t y
297
Regardless~of whether flumazenil, CGS 9895, and ~-CCtB have etfects in non-tolerant subjects, each of these drugs e f f e c t i v e l y decreased response rates in tolerant subjects at doses that had l i t t l e effect in non-tolerant subjects. Thus, exposure to the agonist chlordiazepoxide produced large changes in the effects of these drugs. Those changes in effects were different from effects obtained with the inverse agonist, i~-CCE. Once subjects were rendered tolerant to chlordiazepoxide, the effects of this drug on schedule-controlled responding were not appreciably d i f f e r e n t from those obtained prior to the induction of tolerance. The present findings can be compared to those obtained previously in a study of the proconvulsant effects determined by assessing threshold for pentylenetetrazol-induced siezures (12). In that study, chronic treatment for 14 days with the benzodiazepine agonist, Iorazepam, increased the proconvulsant effectiveness of flumazenil and the antagonist ZK 93426, but did not a,lter the effects of the inverse agonists, FG 7142 and DMCM, on pentylenetetrazol siezure threshold. Thus, as in the present study, the smallest changes in the effects of these drugs induced with agonist treatment, occurred with the drugs having the greatest inverse agonist i n t r i n s i c a c t i v i t y . The present results are similar in many respects to those obtained with opioid antagonists. In several studies, including studies on operant behavior, s e n s i t i v i t y to opioid-antagonists in naive subjects is appreciably less than that observed in opioid-dependent subjects, and the antagonist dose-effect curve is shifted to the l e f t (e.9. 16, 17, 18). In the present study, the s h i f t s to the left of the antagonist dose-effect curves could not be estimated since the highest doses studied produced only small decreases in response rates. However, s h i f t s to the l e f t of the dose-effect curves were clearly evident since the antagonists had effects in the subjects, once tolerant, that did not occur in the subjects prior to the induction of tolerance. Since each of the antagonists had l i t t l e or no i n t r i n s i c a c t i v i t y in non-tolerant subjects, their actions in tolerant subjects is l i k e l y due to the displacement of agonist from receptor sites and suggests the development of physiological dependence on chlordiazepoxide. Several other results suggest that dependence accompanied the tolerance to effects of chlordiazepoxide in the present study. The loss in body weight produced by each of the antagonists of the present study is similar to, though of lesser magnitude than, that observed with cessation of chlordiazepoxide (19, 20) and opioid treatment (21). Several of the signs observed following antagonist administration have b e e n characteristic of withdrawal from chlordiazepoxide or diazepam as described by others (20, 22). In addition, Boisse et al. (23) have indicated development of dependence to chlordiazepoxide in rats at lower doses than those administered in the present study. However, in that study chlordiazepoxide was administered more frequently. Regardless, the present dosage regimen and duration is within a range that would be expected to produce physiological dependence (for a review see 24). Several papers have suggested that the repeated administration of agonist changes the efficacy of benzodiazepine-receptor ligands towards inverseagonist a c t i v i t y . For example, in the study by Petersen and Jensen (12), the maximal effect of Iorazepam was decreased when subjects were rendered tolerant to Iorazepam. Further, antagonists with l i t t l e efficacy in non-tolerant subjects had inverse-agonist a c t i v i t y in tolerant subjects. In many cases, however, increased inverse agonist efficacy may be d i f f i c u l t to d i s t i n g u i s h from precipitated withdrawal due to displacement of agonist from sites of action. For example, the present decreases in operant behavior may be due to either of these actions. S i m i l a r l y , certain grossly observable behaviors may also be due to either of these effects.
298
BenzodlazepineAntagonist Sensitivity
Vol. 44, No. 4, 1989
In the present study, none of the antagonists, including the inverse agonist B-CCE, produced weight loss in non-tolerant subjects, whereas each of the drugs produced s i g n i f i c a n t weight loss in the subjects when chlordiazepoxide tolerant. The weight loss suggests that physiological dependence on chlordiazepoxide had developed in these subjects. That weight loss did not follow the administration of B-CCE in non-tolerant subjects suggests that not all of the effects of antagonists in tolerant subjects are due to a change in inverse agonist efficacy occuring with repeated administration of agonist. Although flumazenil, CGS 9895 and 6-CCtB each precipitated some signs of withdrawal in tolerant subjects, there were differences in the effects of the antagonists. For example, the dose-effect curves for flumazenil and ~-CCtB were similar, whereas the dose-effect curve for CGS 9895 had a r e l a t i v e l y greater slope. In addition, only flumazenil produced grossly observable behavioral signs of withdrawal. The effects of flumazenil also differed from those of the other antagonists in that injection of this compound in tolerant subjects produced vocalization and i r r i t a b i l i t y on handling. None of the other antagonists produced these effects. Since not a l l of the antagonists produced these effects, i t is possible that the actions of these antagonists are mediated by different mechanisms. ACKNOWLEDGEMENTS This research was supported in part by NIH Grant NS-22287. We thank Jose A. Prada and Richard Rodriguez for their expert assistance in the conduct of these studies, Dr. D. B. Vaupel for s t a t i s t i c a l advice, and Dr. R. R. G r i f f i t h s for the loan of CGS 9895. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
E.P. BONETTI, L. PIERI, R. CUMIN, R. SCHAFFNER, M. PIERI, E.R. GAMZU, R.K.M. MULLER, W. HAEFELY, Psychopharmacology 78:8-18 (1982). S.E. FILE, S. PELLOW, Psychopharmacology 88:1-11 (1986). R.F. ALBRECHT, J. COOK, W.E. HOFFMAN, P. LARSCHEID, D.J. MILETICH, N. NAUGHTON, Neuropharmacology 24:957-963 (1985). W.E. HOFFMAN, R.F. ALBRECHT, D.J. MILETICH, T.J. HAGEN, J.M. COOK, Anesth. Analg. 65:639-644 (1986). N. YOKOYAMA, B. RITTER, A.D. NEUBERT, J. Med. Chem. 25:337-339 (1982). N.J. KATZMAN, H.E. SHANNON, J. Pharmacol. Exp. Ther. 235:589-595 (1985). P.N. VAN GORDER, W.E. HOFFMAN, V. BAUGHMAN, R.F. ALBRECHT, D.J. MILETICH, F. GUZMAN, J.M. COOK, Anesth. Analg. 64:129-135 (1985). H.E. SHANNON, F. GUZMAN, J.M. COOK, Life Sci. 35:2227-2236 (1984). P.J. COWEN, A.R. GREEN, D.J. NUTT, I.L. MARTIN, Nature 290:54-55 (1981). N.R. OAKLEY, B.J. JONES, Eur. J. Pharmacol. 68:381-382 (1980). P. POLC, E.P. BONETTI, R. SCHAFFNER, W. HAEFELY, Naunyn Schmeideberg's Arch. Pharmacol. 321:260-264 (1982). E.N. PETERSEN, L.H. JENSEN, Eur. J. Pharmacol. 133:309-317 (1987). T.J. HAGEN, F. GUZMAN, C. SCHULTZ, J.M. COOK, P. SKOLNICK, H.E. SHANNON, Heterocycles 24:2845-2855 (1986). G.W. SNEDECOR, W.G. COCHRAN, S t a t i s t i c a l Methods. 6th ed. pp. 135-171, Iowa State University Press, Ames Iowa (1967).
Vol. 44, No. 4, 1989
Benzodlazeplne Antagonist Sensitivity
299
15. C.B. FERSTER, B.F. SKINNER, Schecl~les of Reinforcement. AppletonCentury-Crofts, New York (~ST) 16. R.J. VALENTINO, S. HERLING, J.H. WOODS, J. Pharmacol. Exp. Ther. 224: 307-313 (1983). 17. J.H. WOODS, J. CARNEY, In Behavioral Tolerance: Research and Treatment Implications, N.A. Krasnegor (ed) NIDA Research Monograph Series, Vol. 19, U.S. Gov't Printing Office, Washington, D.C., pp. 54-66, (1978) 18 J.L. KATZ, Psychopharmacology, 89: 323-326, 1986. 19. S. YANAURA, E. TAGASHIRA, Folia Pharmacol. Japon 71:157-168 (1975). 20. G.P. RYAN, N.R. BOISSE, J. Pharmacol. Exp. Ther. 226:100-107 (1983). 21. E. WEI, H.H. LOH, E.L. WAY, J. Pharmacol. Exp. Ther. 184:398-403 (1973). 22. W.R. MARTIN, L.F. McNICHOLAS, S. CHERIAN, Life Sci. 31:721-730 (1982). 23. N.R. BOISSE, G.P. RYAN, J.J. GUARINO, Problems of Drug Dependence, 1981. L.S. ,Harris (ed) NIDA Monograph 41, U.S. Government Printing Office, Washington OC (1982). 24. J.H. WOODS, J.L. KATZ, G. WINGER, Pharmacol. Rev. 39: 251-413, 1987.