Adenylyl Cyclase Supersensitivity in Opioid-Withdrawn NG Hybrid Cells Requires G s but Is Not Mediated by the G s Subunit

Transcription

1 /98/ $03.00/0 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 286, No. 2 Copyright 1998 by The American Society for Pharmacology and Experimental Therapeutics Printed in U.S.A. JPET 286: , 1998 Adenylyl Cyclase Supersensitivity in Opioid-Withdrawn NG Hybrid Cells Requires G s but Is Not Mediated by the G s Subunit HERMANN AMMER and RÜDIGER SCHULZ Institute of Pharmacology, Toxicology and Pharmacy, University of Munich, Königinstrasse, München, Germany Accepted for publication April 20, 1998 This paper is available online at ABSTRACT On the cellular level, opioid dependence is characterized by a significant elevation of adenylyl cyclase (AC) activity after drug withdrawal, a regulatory phenomenon termed AC supersensitivity or camp overshoot. The present study examines the role of the stimulatory G protein (G s ) in the expression of naloxone precipitated opioid withdrawal in chronically morphine (10 M; 3 days) treated neuroblastoma X glioma (NG108 15) hybrid cells. Determination of high-affinity [ 3 H]forskolin binding to intact cells, which provides a direct parameter for the binding of the activated -subunit of G s (G s ) to AC, revealed that the enhancement of AC activity after opioid withdrawal is not caused by an increased stimulation of effector activity by G s. Although not a direct function of G s, the expression of AC supersensitivity required G s -mediated stimulation of AC, because 1) the enhancement of AC activity after opioid withdrawal Opioid receptors belong to the family of seven-transmembrane domain G protein-coupled receptors (for review see: Reisine and Bell, 1993). Their acute activation leads to the inhibition of AC activity, an effect that is mediated by pertussis toxin-sensitive G proteins of the G i /G o class (Childers, 1991). Chronic exposure to an opioid, however, produces multiple adaptational changes within the stimulatory branch of AC resulting in an increased capacity of stimulatory AC signaling during the state of dependence (Ammer and Schulz, 1993; 1995; 1997). Up-regulation of stimulatory AC signaling usually is masked as long as the inhibitory opioid is present but manifests in a significant enhancement of camp production after removal of the agonist (Sharma et al., 1975). This regulatory phenomenon, generally referred to as camp overshoot or AC supersensitivity, represents a cellular correlate for opioid withdrawal and frequently has been used Received for publication November 7, was observed only in the presence of low, but not of high concentrations of forskolin, and 2) chemical inactivation of G s by low ph pretreatment abolished the induction of AC supersensitivity. Moreover, the regulatory mechanism underlying AC supersensitivity not only required the presence of activated G s per se, but functional intact stimulatory signal transduction pathways. Indeed, blockade of prostaglandin E 1 receptor/g s interaction in situ with a site-specific anti-g s antibody, as well as uncoupling of prostaglandin E 1 receptor signaling by cholera toxin-catalyzed ADPribosylation of G s, prevented the expression of AC supersensitivity in membranes from opioid-withdrawn cells. These results suggest that the enhancement of AC activity in opioid-dependent cells, triggered by drug withdrawal, is not a direct G s effect, but involves a secondary regulatory event that requires costimulation of AC by acutely receptor-activated G s. to define the state of dependence (Sharma et al., 1975; Collier, 1984). AC supersensitivity had been detected originally in chronically morphine-treated neuronal cell lines, such as neuroblastoma x glioma (NG108 15) hybrid cells (Sharma et al., 1975; Law et al., 1984) and human neuroblastoma SH- SY5Y cells (Yu et al., 1990). Heterologous expression of the recently cloned opioid receptor cdna (delta, kappa, mu) revealed that AC supersensitivity can be reconstituted with all three opioid receptor types (Law et al., 1994; Avidor-Reiss et al., 1995a, 1995b). Moreover, AC supersensitivity apparently represents a more common means of cellular adaptation toward chronic inhibitory drug action, because several other inhibitory receptors, such as alpha 2 adrenergic, muscarinic cholinergic and somatostatin receptors also induce this phenomenon (Thomas and Hoffman, 1987). Although the role of AC supersensitivity in the development of drug dependence is well recognized (Nestler et al., 1993), the biochemical signal mediating the increase in AC activity is still unknown. ABBREVIATIONS: AC, adenylyl cyclase; ANOVA, analysis of variance; camp, cyclic AMP; CTX, cholera toxin; DSLET, H-Tyr-d-Ser-Gly-Phe- Leu-Thr-OH; DTT, dithiothreitol; EGTA, ethylene glycol bis( -aminoethyl ether)-n, N,N,N -tetraacetic acid; HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; ;G protein, guanine nucleotide (GTP)-binding protein; G s, stimulatory G protein; G s, -subunit of G s ;G, G protein -subunit; GTP S, guanosine-5 -O-(3-thio)triphosphate; PGE 1, prostaglandin E 1 ; Ro , dl-4-(butoxy-4-methoxybenzyl)-2-imidazolidinone. 855

2 856 Ammer and Schulz Vol. 286 The activity of opioid-regulated AC is under the control of stimulatory receptor systems (Collier, 1984). Signal transduction from stimulatory receptors to AC involves the heterotrimeric stimulatory G protein (G s ) which, upon activation, dissociates into its GTP-bound G s and G subunits (Gilman, 1987). Activated G s subsequently binds to AC, thereby stimulating catalytic activity. Recent molecular cloning has permitted identification of at least nine distinct AC isoforms which show several common and disparate features in the regulation of effector activity (for review see: Sunahara et al., 1996). Whereas all AC isoforms can be stimulated by both activated G s and forskolin, several additional regulatory factors exist which may positively or negatively modulate catalytic activity in a subtype-specific manner. For instance, direct G i -mediated inhibition of enzymatic activity has been shown for AC types I, V and VI (Wong et al., 1991; Sunahara et al., 1996), whereas G subunits may either attenuate AC type I activity (Tang and Gilman, 1991) or synergistically activate G s -stimulated AC types II and IV (Tang and Gilman, 1991; Federman et al., 1992). More complex and indirect modes of regulation have been shown for Ca -dependent calmodulin, Ca and phosphorylation by protein kinases A and C (Choi et al., 1992; Premont et al., 1992; Kawabe et al., 1994). Because of the molecular diversity and regulatory complexity, investigation into the regulatory mechanism underlying AC supersensitivity is highly complicated. In a first step to decipher the stimulatory signal, the present study was initiated to determine the role of the stimulatory G protein in mediating the enhancement of AC activity in opioid-withdrawn cells. As a model system we used chronically morphine treated NG hybrid cells, which carry high levels of inhibitory delta opioid receptors (Hamprecht et al., 1985). Although morphine acts as a partial agonist on delta opioid receptors in this cell line (Vachon et al., 1987), this opiate proved particularly advantageous in inducing cellular dependence, most likely because it fails to desensitize delta opioid receptor signaling (Law et al., 1984; Keith et al., 1996). Our results demonstrate that the enhanced AC activity in morphine withdrawn NG hybrid cells is not a function of an increased stimulation by G s, but involves an additional regulatory event that requires coincident stimulation of AC by receptor-activated G s. Experimental Procedures Materials. Reagents were purchased from the following sources: [ 3 H]forskolin (31 Ci/mmol) from NEN DuPont (Dreieich, Germany); [ 125 I]cAMP tracer (2000 Ci/mmol) from Amersham International (Braunschweig, Germany); forskolin, Ro , and CTX from Calbiochem (Bad Soden, Germany); DSLET from Bachem (Heidelberg, Germany); rabbit anti-camp antibody from BioMakor (Rehovot, Israel). Tissue culture reagents were obtained from PAN Systems (Aidenbach, Germany) and Gibco/BRL (Eggenstein, Germany). PGE 1 and all standard laboratory reagents were from Sigma (Deisenhofen, Germany). Cell culture, chronic opioid treatment. Neuroblastoma x glioma (NG108 15) hybrid cells (Hamprecht et al., 1985) were grown as monolayers in Dulbecco s modified Eagle s medium, containing 5% heat-inactivated fetal calf serum, 100 M hypoxanthine, 1 M aminopterin and 16 M thymidine, in a humidified atmosphere of 5% CO 2 and 95% air at 37 C. Subconfluent monolayers were split in the ratio of 1:5, grown overnight and subjected to chronic inhibitory opioid treatment for another 3 days with the addition of 10 M morphine. This treatment regimen has been shown previously to produce high degrees of cellular dependence in NG hybrid cells (Ammer and Schulz, 1995). Untreated cultures of individual passages served as controls. [ 3 H]Forskolin binding. [ 3 H]Forskolin binding in whole NG hybrid cells was determined essentially as described (Alouisi et al., 1991; Kim et al., 1995). Cells from two 75 cm 2 culture flasks were harvested, washed three times (200 g; 10 min) with ice-cold 20 mm HEPES-buffered Dulbecco s modified Eagle s medium (ph 7.4), and resuspended in the same buffer at a density of cells/ml. Cells were equilibrated on ice for 30 min either in the absence (control) or presence of 10 M morphine (opioid-dependent cells). Binding reactions were at 4 C for 60 min in a total volume of 500 l HEPES-buffered Dulbecco s modified Eagle s medium containing cells, 10 nm [ 3 H]forskolin and various concentrations of PGE 1 to stimulate formation of G s /AC complexes. Nonspecific binding was defined with 10 M unlabeled forskolin. In some experiments, binding of [ 3 H]forskolin was displaced with increasing concentrations of unlabeled forskolin. Chronically morphine-treated cells were assayed in the presence of either morphine (10 M) for investigation of the state of dependence or of naloxone (100 M) for analysis of the state of opioid withdrawal. Binding reactions were terminated by rapid filtration over Whatman GF/C filters, followed by four washes with 5 ml each of ice-cold 50 mm Tris HCl buffer, ph 7.4, containing 10 mm MgCl 2. Cell associated radioactivity was determined by scintillation counting of the filters. Adenylyl cyclase assay. AC activity was determined in a particulate membrane preparation according to Vachon et al. (1987). Cells were collected, washed three times (200 g; 10 min) with phosphate-buffered saline (ph 7.4) and membranes were prepared in 5 mm Tris-HCl buffer (ph 7.4), containing 1 mm DTT and 1 mm EGTA. Membranes were resuspended in the above buffer at a concentration of 10 mg/ml and stored in aliquots at 70 C until use. AC activity was measured in 40 mm Tris-HCl buffer (ph 7.4), containing 0.2 mm EGTA, 0.2 mm DTT, 100 mm NaCl, 10 mm MgCl 2, 0.5 mm ATP, 5 M phosphocreatine, 5 units/ml creatine kinase, 10 M GTP and 30 M Ro Reactions (100 l total volume) were started by the addition of 10 g of membrane protein, maintained for 10 min at 32 C and stopped with 500 l of ice-cold 10 mm HCl. Membranes derived from opioid-dependent cells were first equilibrated at 4 C for 30 min with 10 M morphine before AC activity was determined either in the presence of morphine (state of dependence) or after addition of 100 M naloxone to uncover effector supersensitivity (precipitated withdrawal). The high concentration of naloxone (100 M) used to induce opioid withdrawal did not produce nonspecific effects on AC activity in membranes from control cells (not shown) and was essential in stably reproducing the phenomenon of AC supersensitivity (Lee et al., 1988; Law et al., 1994; Ammer and Schulz, 1995). Stimulation of effector activity was achieved with either PGE 1 or forskolin at the concentrations indicated. The amount of camp generated was quantitated by radioimmunoassay after dilution of the samples as described (Ammer and Schulz, 1995). Modification of G s activity. In some experiments the functional integrity of G s was altered before determination of AC activity. Selective inactivation of G s function was achieved by transient exposure of the membranes to low ph (Childers and LaRiviere, 1984). Membranes were recovered by centrifugation (10,000 g; 15 min), resuspended in 50 mm sodium acetate (ph 4.5) and incubated for 20 min at 4 C. Controls were kept in the presence of 50 mm Tris-HCl (ph 7.4) buffer. Reactions were stopped by dilution with 10 volumes of ice-cold NMT buffer (50 mm Tris-HCl, 150 mm NaCl, 5 mm MgCl 2 ; ph 7.4) and membranes were collected as above. Subsequently, membranes were resuspended in NMT buffer (0.5 mg/ml) and equilibrated for another 30 min at 4 C either without (control) or with 10 M morphine (dependence) before determination of AC activity. CTX-catalyzed ADP-ribosylation was used to constitutively acti-

3 1998 G s and Opioid Withdrawal 857 vate G s (Gill and Woolkalis, 1991). Membranes (100 g/50 l) from either naive or chronically morphine-treated cells were incubated for 30 min at 25 C in 10 mm Tris-HCl buffer (ph 7.4), containing 150 mm NaCl, 10 mm thymidine, 0.1 mm Gpp[NH]p, 5 M NAD and 10 g/ml CTX. Reactions were stopped by the addition of 1 ml ice-cold NMT buffer, membranes were washed and AC activity was determined as above. Membranes that were incubated in the absence of the toxin served as controls. Under these conditions, CTX treatment almost completely uncoupled stimulatory receptors from G s as indicated by the failure of PGE 1 to further activate AC. Receptor-mediated activation of G s was blocked with a C-terminal anti-g s (S1/3) antibody according to Ammer and Schulz (1997). Membranes from opioid-dependent cells were resuspended in NMT buffer and combined with a maximal effective amount of protein A-purified S1/3 antibody (30 g IgG/5 membranes). Reactions also included 10 M morphine to avoid spontaneous withdrawal during the incubation procedure. Membranes were kept for 60 min on ice before the ability of naloxone (100 M) to induce AC supersensitivity in the presence of 0.1 M PGE 1 was determined. Controls were processed as above with the exception that normal IgG was used. Data analysis. EC 50, E max and IC 50 values were determined by nonlinear least-squares regression curve fitting, with SigmaPlot (Jandell Scientific, Erkrath, Germany) software. [ 3 H]Forskolin binding parameters were calculated from homologous displacement curves according to the method of DeBlasi et al. (1989). Statistical differences were determined by one-way ANOVA or Student s twotailed t test where appropriate. Results Comparison of PGE 1 receptor-stimulated AC activation and G s /AC interaction. To determine whether the expression of AC supersensitivity in opioid-withdrawn NG hybrid cells is a function of enhanced activation of the catalytic component of AC by G s, we compared the effects of PGE 1 on the stimulation of AC activity with those on the promotion of high-affinity [ 3 H]forskolin binding, which provides a direct measure for the binding of activated G s to AC (Alouisi et al., 1991; Kim et al., 1995). In membranes from naive cells, activation of PGE 1 receptors dosedependently stimulated effector activity with an EC 50 value of nm (mean S.D.; n 4). Chronic morphine treatment (10 M; 3 days) per se had no effect on both the potency (EC nm; mean S.D.; n 4) as well as the maximum capacity of PGE 1 to activate AC as long as the assays were performed in the presence of the inhibitory opioid used for pretreatment (E max vs pmol/min/mg protein for control and opioid-dependent cells, respectively; means S.D.; n 4). However, precipitation of opioid withdrawal by the addition of naloxone (100 M) resulted in an approximately 40% increase in the capacity of PGE 1 receptor-stimulated AC activity (E max pmol/min/mg protein; mean S.D.; n 4; P.001), without any change in its EC 50 value ( nm; mean value S.D.; n 4). Thus, AC supersensitivity is characterized by an increased capacity rather than a change in the sensitivity of PGE 1 receptor signaling. The activated, GTP-bound form of G s represents the principle stimulator of all membrane-bound AC isoforms currently known (Sunahara et al., 1996). High-affinity [ 3 H]forskolin binding experiments were performed to evaluate if AC supersensitivity after opioid withdrawal is caused by an enhanced stimulation of AC by G s. Although high-affinity [ 3 H]forskolin binding to intact NG hybrid cells was only barely detectable in the absence of a stimulatory ligand, PGE 1 receptor-mediated activation of G s resulted in the formation of G s /AC complexes and, thus, in a strong increase in high-affinity [ 3 H]forskolin binding. However, as shown in figure 1B, the dose-response curves obtained for naive, opioid-dependent and opioid-withdrawn cells were almost superimposable, yielding identical values for EC 50 (10.9 3, , and nm; means S.D.; n 4) and E max (76.7 8, , and fmol [ 3 H]forskolin binding sites/10 6 cells; means S.D.; n 4). These results demonstrate that the enhancement of AC activity after opioid withdrawal is not associated with changes in the functional Fig. 1. Comparison of PGE 1 receptor-stimulated AC activity with highaffinity [ 3 H]forskolin binding in NG cells. NG hybrid cells were grown either in the absence (control; E) or presence of morphine (10 M; 3 days) to induce dependence. A: Cells were harvested and the dose-response relationship of PGE 1 -stimulated AC activity was determined in a particulate membrane preparation. Membranes from opioiddependent cells were assayed either in the presence of morphine ( ) or naloxone ( ) to investigate the states of dependence and opioid withdrawal, respectively. Enzymatic activity is expressed in picomoles of camp formed per min per mg of membrane protein. The data shown are mean values S.D. of n 4 experiments. B: Binding of G s to AC was assessed by high-affinity [ 3 H]forskolin binding to intact cells as described under Experimental Procedures. Binding of [ 3 H]forskolin (10 nm) was simulated dose dependently by the addition of increasing concentrations of PGE 1 (0.1 nm 10 M). Opioid-dependent cells were measured either in the presence of morphine ( ) or after naloxone-precipitated withdrawal ( ). The number of G s /AC complexes formed is given in femtomoles of [ 3 H]forskolin bound per 10 6 cells. The data presented are mean values S.D. from n 4 experiments. Vertical bars indicate calculated EC 50 values.

4 858 Ammer and Schulz Vol. 286 interaction between G s and AC. Thus, the expression of AC supersensitivity is not a function of an enhanced stimulation of AC by activated G s but involves an additional regulatory event. Effect of opioid withdrawal on the affinity of [ 3 H]forskolin binding. Although both chronic morphine treatment and opioid withdrawal lack any effect on the maximum number of PGE 1 receptor stimulated G s /AC complexes, binding of an additional component to AC possibly could alter the conformation of the binding site for [ 3 H]forskolin and, hence, affect its affinity. Therefore, homologous displacement curves of PGE 1 (10 M)-stimulated high-affinity [ 3 H]forskolin binding were constructed. As shown in figure 2, there is no difference in the inhibition curves, regardless of whether naive, chronically morphine- or opioid-withdrawn cells were measured. Calculation of the apparent affinity constants by the method of DeBlasi et al. (1989) resulted in almost identical K d values for forskolin ( , , and nm for naive, opioid-dependent and opioid-withdrawn cells, respectively; means S.D.; n 4). Regulation of [ 3 H]forskolin binding by acute delta opioid receptor activation. The failure of opioid withdrawal to affect the affinity of [ 3 H]forskolin binding raises the question of whether binding of an additional regulator of AC activity must necessarily alter the conformation of the G s /AC complex, which represents the high-affinity binding site for forskolin (Alouisi et al., 1991). For this, the effect of acute delta opioid receptor activation on high-affinity [ 3 H]forskolin binding was determined. In NG hybrid cells, inhibition of AC activity is mediated via the inhibitory G protein -subunit G i 2 (McKenzie and Milligan, 1990). Acute inhibition of AC activity, neither with morphine (10 M) nor the full delta receptor agonist DSLET (1 M), had any effect on PGE 1 receptor-stimulated [ 3 H]forskolin binding parameters (table 1), which indicates that binding of G i 2 to AC does Fig. 2. Homologous displacement of high-affinity [ 3 H]forskolin binding in naive, opioid-dependent and opioid-withdrawn NG hybrid cells. The affinity of forskolin for the G s /AC complex was determined by homologous displacement of PGE 1 (10 M)-stimulated high-affinity [ 3 H]forskolin (10 nm) binding with increasing concentrations of competing unlabeled forskolin. Either native (E), chronically morphine (10 M; 3 days) -treated ( ), or opioid-withdrawn cells ( ) were measured. Specific high-affinity [ 3 H]forskolin binding obtained in the presence of 10 M forskolin was set at 100%. Each data point represents the mean value of n 4 experiments. Standard deviation was routinely less than 7% of the mean. Binding parameters were calculated according to the formalism of DeBlasi et al (1989). Dotted lines indicate half-maximal inhibition of [ 3 H]forskolin binding, which is almost identical in control, opioid-dependent and opioid-withdrawn cells, respectively. TABLE 1 Effect of acute delta-opioid receptor activation of [ 3 H]forskolin binding parameters in NG cells High-affinity [ 3 H]forskolin binding in intact NG hybrid cells was stimulated with increasing concentrations of PGE 1 to determine EC 50 and E max values. The affinity of [ 3 H]forskolin binding was determined from homologous displacement curves obtained in the presence of 10 M PGE 1. Reactions were done either in the absence of an inhibitory ligand or in the presence of morphine (10 M) or the high-potency delta-agonist DSLET (1 M). Data are mean values S.D. of n 3 experiments. PGE 1 -stimulated [ 3 H]forskolin binding Inhibitory ligand EC 50 (nm) E max (fmol/10 6 cells) K d (nm) None Morphine DSLET not interfere with the formation and conformation of G s /AC complexes, confirming previous data (Kim et al., 1995). Requirement of activated G s for AC supersensitivity. The finding that the generation of AC supersensitivity in opioid-withdrawn NG hybrid cells is not a function of enhanced G s activity raises the question of whether G s is involved at all in the regulatory mechanism leading to the increase in AC activity. To investigate this issue we first tested the effect of different concentrations of forskolin on the expression of AC supersensitivity. The diterpene forskolin directly binds to and stimulates AC activity by mechanisms which depend on the concentration used. In the nanomolar range, forskolin potentiates receptor-stimulated AC by increasing the affinity of G s to AC; whereas, high micromolar concentrations directly activate the effector molecule without the requirement of G s for its full action (Sutkowski et al., 1996). Precipitation of opioid withdrawal by naloxone (100 M) only resulted in AC supersensitivity when AC was stimulated with 100 nm but not 10 M forskolin (table 2). The finding that opioid withdrawal failed to induce AC supersensitivity in the presence of 10 M forskolin apparently was not caused by maximal activation of the effector molecule, because higher forskolin concentrations further enhanced AC activity (not shown). These results suggest that the expression of AC supersensitivity requires costimulation of AC by G s. This issue was investigated further in membranes depleted of G s function after transient exposure to acidic conditions (Childers and LaRiviere, 1984). Inactivation of G s largely decreased PGE 1 receptor-mediated stimulation of AC TABLE 2 Effect of forskolin on the generation of AC supersensitivity in opioidwithdrawn NG cells Chronic opioid regulation of membrane-bound AC activity was determined in the presence of 100 nm or 10 M forskolin. Membranes from chronically morphine (10 M; 3-days)-treated cells were either measured in the presence of morphine (10 M) or after precipitation of withdrawal (naloxone; 100 M). Enzymatic activity is expressed in picomoles of camp generated per mg of membrane protein per min. The data shown are mean values S.D. of n 6 separate experiments done in triplicate. Statistical differences were calculated by ANOVA followed by Bonferroni post test. Condition Adenylyl cyclase activity 0.1 M forskolin 10 M forskolin (camp; pmol/min/mg membrane protein) Control Acute morphine a c Chronic morphine Chronic morphine naloxone b a Significantly different from control at P.05. b Significantly different from chronic morphine at P.001. c Significantly different from control at P.001.

5 1998 G s and Opioid Withdrawal 859 by more than 95% in membranes from both naive and opioiddependent cells. Conversely, low ph treatment increased the efficacy of morphine to acutely inhibit AC activity in membranes from naive cells ( vs % inhibition; means S.D.; n 4), whereas no effect on delta opioid receptor desensitization was observed in membranes from chronically morphine treated cells, as assessed in the presence of 100 M morphine ( vs % inhibition; means S.D.; n 4). Thus, there are still functional delta opioid receptors present in low ph pretreated membranes derived from opioid-dependent cells that should be able to trigger AC supersensitivity after withdrawal. However, the addition of naloxone (100 M) to G s -depleted membranes failed to induce a cellular withdrawal response, which suggests that stimulation of AC activity by G s represents an essential requirement for the expression of AC supersensitivity (fig. 3). AC supersensitivity requires intact PGE 1 receptor/g s signaling. We further investigated whether either the presence of activated G s per se or the functional intact stimulatory signaling during the state of opioid withdrawal is required for the expression of AC supersensitivity. For this, stimulatory signal transduction in membranes from opioid-dependent NG cells was modulated by two different approaches. First, short-term CTX treatment (10 g/ ml; 30 min) was used to uncouple G s from its associated stimulatory receptors by constitutive activation of G s. CTX treatment of the membranes increased basal AC activity by about 3-fold and almost completely abolished further stimulation via PGE 1 receptors, whereas no effect on acute delta opioid receptor-mediated inhibition of AC was observed. Receptor-independent activation of AC by constitutively activated G s, however, prevented the expression of AC supersensitivity in opioid-withdrawn cell membranes (fig. 4). This result indicates that not only the presence of activated G s alone but also an intact stimulatory control of AC is necessary to elicit AC supersensitivity. The latter conclusion was confirmed by experiments in which receptor-mediated activation of G s was blocked in situ by the use of a C-terminal Fig. 4. Constitutive activation of G s by CTX-catalyzed ADP-ribosylation inhibits AC supersensitivity after opioid withdrawal. Membranes from naive and chronically morphine-treated NG cells were subjected to short-term CTX treatment to constitutively activate G s by a receptor-independent mechanism. Subsequently, the ability of morphine (10 M) to inhibit and of naloxone (100 M) to increase AC activity was determined in naive and opioid-dependent cells, respectively. Membranes from opioid-dependent cells were kept in the presence of morphine (10 M) throughout the incubation periods. Enzymatic activity is expressed in picomoles of camp generated per mg of membrane protein per min. The data shown are mean values S.D. of n 3 experiments each performed in triplicate. ***; statistically different from controls at P.001 as determined by ANOVA. anti-g s antibody (Ammer and Schulz, 1997). Blockade of PGE 1 receptor/g s interaction by preincubation of NG hybrid cell membranes with a maximal effective concentration of S1/3 antibody (30 g IgG/5 membranes) resulted in an approximately 90% attenuation of PGE 1 (0.1 M)-stimulated AC activity. Anti-G s antibody treatment apparently selectively interfered with stimulatory signal transduction, because acute inhibition of PGE 1 (0.1 M)-stimulated AC [by morphine (10 M)] remained unaffected ( vs pmol camp/min/mg of membrane protein; means S.D.; n 3). In contrast, disruption of PGE 1 receptor/g s coupling in membranes from opioid-dependent cells completely abolished the generation of AC supersensitivity after naloxone (100 M)-precipitated withdrawal (fig. 5). Preincubation of the membranes with control IgG (30 g/5 g) failed to affect both receptor mediated stimulation of AC and the induction of AC supersensitivity. Therefore, the expression of AC supersensitivity in opioid-withdrawn NG hybrid cells requires functional, intact stimulatory signal transduction pathways. Fig. 3. Inhibition of AC supersensitivity by low ph pretreatment. Membranes prepared from naive or opioid-dependent NG cells were exposed transiently to ph 4.5 before determination of PGE 1 (10 M)- stimulated AC activity. Inhibition of AC activity in naive NG hybrid cell membranes was assessed with 10 M morphine. The ability of naloxone (100 M) to precipitate AC supersensitivity was tested in membranes from opioid-dependent cells which were equilibrated for 30 min in the presence of 10 M of morphine to avoid spontaneous withdrawal. Enzymatic activity is expressed in picomoles of camp generated per mg of membrane protein per min. The data shown are mean values S.D. of n 4 experiments each performed in triplicate. ***; statistically different from controls at P.001 as determined by ANOVA. Discussion The results of the present study reveal that the enhancement of AC activity (AC supersensitivity) in opioid-withdrawn NG hybrid cells is not a function of an increased activation by G s, the G protein -subunit mediating stimulation of enzyme activity. Nevertheless, the stimulatory G protein was critical in the generation of cellular withdrawal, because both G s -independent activation of AC with high concentrations of forskolin as well as inactivation of G s function by low ph pretreatment abolished the expression of AC supersensitivity. In addition, interruption of stimulatory receptor/ac signaling after constitutive activation of G s by CTX-catalyzed ADP-ribosylation or by application of a C- terminal anti-g s antibody further demonstrated that the regulatory mechanism leading to the enhancement of AC

6 860 Ammer and Schulz Vol. 286 Fig. 5. Blockade of PGE 1 receptor/g s interaction abolishes the ability of naloxone to precipitate AC supersensitivity in opioid-dependent NG cells. Membranes (5 g) from opioid-dependent NG cells were either incubated together with 30 g of a protein A-purified C-terminal anti-g s antibody or an identical concentration of control IgG to block PGE 1 receptor/g s interaction as described under Experimental Procedures. All steps were performed in the presence of morphine (10 M) to avoid spontaneous withdrawal. After antibody treatment, the ability of naloxone (100 M) to precipitate AC supersensitivity was determined in the presence of 0.1 nm PGE 1. Enzymatic activity is expressed in picomoles of camp generated per mg of membrane protein per min. The data shown are mean values S.D. of n 3 experiments each performed in triplicate. ***; statistically different at P.001 according to Student s t test. activity after opioid withdrawal requires functional, intact stimulatory AC signaling pathways. The phenomenon of AC supersensitivity originally had been described in chronically morphine-treated NG hybrid cells after opioid withdrawal (Sharma et al., 1975) and subsequently was used to define the state of opioid dependence in several neuronal cell lines and tissues (Collier, 1984; Childers, 1991; Nestler et al., 1993). The enhancement of AC after drug withdrawal clearly represents a more general adaptive phenomenon toward chronic treatment with inhibitory drugs because prolonged activation of alpha 2 adrenergic, muscarinic cholinergic or somatostatin receptors was found to induce a similar cellular response (for review see: Thomas and Hoffman, 1987). Thus, the main criterion for the choice of a cell system to investigate the biochemical mechanisms underlying AC supersensitivity would be the development of a high degree of cellular dependence; whereas the nature of the inhibitory receptor system used to persistently inhibit AC is probably of secondary significance. Although the recent cloning of a mu opioid receptor cdna (Chen et al., 1993), the opioid binding site associated with classical withdrawal in vivo (Nestler et al., 1993), has permitted the establishment of cell lines stably expressing high levels of a single population of mu opioid receptors (Avidor-Reiss et al., 1995b; Ammer and Schulz, 1997), the present study was performed with chronically morphine treated NG hybrid cells because this cell system is well characterized and has a long history in the investigation of cellular aspects of opioid dependence and withdrawal (Sharma et al., 1975; Collier, 1984; Childers, 1991). NG hybrid cells carry ample amounts of endogenous delta opioid receptors (Hamprecht et al., 1985) which are particularly advantageous in inducing cellular correlates of opioid dependence because morphine, unlike more selective delta agonists, fails to desensitize and down-regulate the delta opioid receptor (Law et al., 1984; Keith et al., 1996). Regulation of membrane-bound AC is highly complex and largely depends on the type of AC present in a particular tissue (Sunahara et al., 1996). Whereas all AC isoforms are stimulated by the GTP-bound form of G s and the diterpene forskolin (Sunahara et al., 1996), certain cyclases are also the target of regulation by G i subunits (Wong et al., 1991), G subunits (Tang and Gilman, 1991; Federman et al., 1992), Ca /calmodulin (Choi et al., 1992) and protein kinases A and C (Premont et al., 1992; Kawabe et al., 1994). Thus, because of this regulatory complexity, the overall level of AC activity measured in a certain tissue reflects the sum of multiple stimulatory and inhibitory inputs. To gain insight into the regulatory mechanism underlying the phenomenon of AC supersensitivity, we first determined whether the enhancement of AC activity is caused by an increased stimulation by G s, the principle activator of AC (Gilman, 1987), or by secondary regulation of G s -stimulated AC activity. Discrimination between the effects of G s on AC activity from other stimulatory signals was achieved by comparing PGE 1 receptor-mediated stimulation of AC activity with the promotion of high-affinity [ 3 H]forskolin binding, which provides a direct measure for the physical interaction of activated G s and AC (Alouisi et al., 1991). Although there is currently little information about the expression pattern of individual AC isoforms in NG cells (these cells contain AC type VI but not AC type I and II, others have not been searched for (MacEwan et al., 1996). The property of both forskolin and activated G s to bind to all AC isoforms (Sutkowski et al., 1996) overcomes this limitation and renders high-affinity [ 3 H]forskolin binding a reliable measure for the number of G s /AC complexes formed, regardless of the cellular expression profile of different AC isoforms (Kim et al., 1995). With this approach we could demonstrate clearly that neither the states of opioid dependence nor withdrawal are associated with changes in the maximum number of G s /AC complexes formed and the affinity of [ 3 H]forskolin binding. These findings confirm that chronic morphine treatment per se has no effect on the steady-state levels of G s (Ammer and Schulz, 1995) and is not likely to affect overall AC abundance, because changes in the expression levels of AC should be detected by this method (MacEwan et al., 1996). Moreover, the findings also indicate that the enhancement of AC activity after opioid withdrawal is not caused by an increased stimulation of AC by activated G s. Thus, the phenomenon of AC supersensitivity seems to represent a regulatory phenomenon that is mediated by a stimulatory signal different of activated G s. This notion is supported by our previous work in which we demonstrated that opioid withdrawal is not associated with an increased activation of G s molecules by PGE 1 receptors (Ammer and Schulz, 1995). In addition, the observation that the expression of AC supersensitivity is confined specifically to only some (AC types I, V, VI, VIII) but not all AC isoforms and that tissue specific differences apparently exist in the ability of AC type I to exhibit AC supersensitivity (Thomas and Hoffman, 1996; Avidor-Reiss et al., 1997) further argues against a direct G s effect as the underlying mechanism. In addition to stimulation by G s, there are several other potential regulatory mechanisms that could account for the enhancement of AC activity after opioid withdrawal, such as direct activation of AC type I by Ca -dependent calmodulin (Choi et al., 1992) and stimulation of AC types II and IV by

7 1998 G s and Opioid Withdrawal 861 G subunits released from inhibitory G proteins (Tang and Gilman, 1991; Federman et al., 1992). Both components act only at defined AC isoforms and their effects are highly synergistic or conditional to the presence of activated G s (Sunahara et al., 1996). Although there is currently no information about the presence of one or more susceptible AC isoforms in NG hybrid cells, the results presented herein strongly suggest that the regulatory mechanism underlying AC supersensitivity must involve an additional component that requires costimulation of AC by activated G s to exert its effect. This notion is based on the observation that opioid withdrawal is able to elicit AC supersensitivity only in the presence of nanomolar concentrations of forskolin, which act to potentiate receptor-mediated stimulation of AC (Sutkowski et al., 1996). Direct activation of catalytic activity by high micromolar forskolin concentrations prevents the development of a cellular withdrawal sign. In addition, transient exposure of membranes from opioid-dependent cells to low ph also was found to abolish the expression of AC supersensitivity. This effect is most likely caused by selective inactivation of G s function (Childers and LaRiviere, 1984), which results in a dramatic decrease of G s -mediated stimulation of AC (Ammer and Schulz, 1993). Both findings indicate that the expression of AC supersensitivity depends on the presence of functional active G s. Such a cooperatively acting mechanism would explain the phenomenon that the percentage increase in AC activity after opioid withdrawal is identical regardless of whether AC activity is measured under basal or stimulated conditions (Yu et al., 1990; Ammer and Schulz, 1997; Thomas and Hoffman, 1996). The involvement of a synergistically acting component in the regulatory mechanism underlying AC supersensitivity does not conflict with the finding that opioid withdrawal fails to affect the K d value of high-affinity [ 3 H]forskolin binding. Because the complex between activated G s and AC constitutes the high-affinity binding site for forskolin (Alouisi et al., 1991; Sutkowski et al., 1996), a change in the high-affinity binding state would be observed only when the binding of an additional regulator of AC alters the conformation of the G s /AC complex. In this respect, Taussig et al. (1994) demonstrated that G i and G s subunits bind to different domains at the effector molecule. To investigate whether the interaction of G i subunits with AC complex would alter the high-affinity binding state for forskolin, the effect of acute delta opioid receptor mediated inhibition of AC on highaffinity [ 3 H]forskolin binding was tested. Our results demonstrate that inhibition of AC activity, which in naive NG hybrid cells is transduced via the inhibitory G protein G i 2 (McKenzie and Milligan, 1990), lacks any effect on both the maximum capacity and the affinity of PGE 1 receptor-stimulated high-affinity [ 3 H]forskolin binding. This finding suggests that the interaction of G i subunits with AC does not influence both the formation and conformation of G s /AC complexes. The present study further revealed that the expression of AC supersensitivity not only requires the presence of activated G s per se but requires functional intact stimulatory receptor signaling. This notion is based on the observation that receptor-independent activation of AC by CTX treatment blocks the ability of naloxone to precipitate AC supersensitivity. In contrast to receptor-mediated activation of G s, CTX-catalyzed ADP-ribosylation of G s results in constitutive activation of G s by inhibiting its intrinsic GTPase activity (Cassel and Selinger, 1977). Consequently, CTX treatment leads to the uncoupling of stimulatory receptors from subsequent signal transduction pathways as indicated by the failure of PGE 1 to further stimulate effector activity. From this we concluded that intact stimulatory receptor/g s interaction may represent a critical step in the expression of AC supersensitivity. To test this notion experiments were performed in which the interaction between PGE 1 receptors and G s was blocked with a C-terminal anti-g s antibody (Ammer and Schulz, 1997). Abrogation of stimulatory AC signaling completely abolished the expression of AC supersensitivity, which demonstrates that the enhancement of AC activity requires acute receptor mediated activation and dissociation of G s into G s and G subunits. Because AC supersensitivity apparently does not reflect a direct G s effect, this finding may suggest that G subunits, acutely released from receptor-activated G s, are critical in the regulatory mechanism underlying AC supersensitivity. In this context, Thomas and Hoffman (1996) and Avidor-Reiss et al. (1996) recently have proposed that G subunits contribute to the development of AC supersensitivity by a yet unidentified indirect mechanism, because heterologous expression of G scavengers was found to prevent the induction of a cellular withdrawal sign. In conclusion, the present study demonstrates that the stimulatory G protein is involved in the regulatory mechanisms underlying AC supersensitivity. Although the enhancement of AC activity during the state of opioid withdrawal is not a direct function of G s, it was found to require stimulation of AC by receptor-activated G s. These results indicate the involvement of an additional stimulatory regulator of AC in the development of AC supersensitivity. Acknowledgment We thank Th. Christ for expert technical assistance. References Alouisi AA, Jasper JR, Insel PA and Motulsky HJ (1991) Stoichiometry of receptor- G s -adenylate cyclase interactions. FASEB J 5: Ammer H and Schulz R (1993) Coupling of prostaglandin E 1 receptors to the stimulatory GTP-binding protein G s is enhanced in neuroblastoma x glioma (NG108 15) hybrid cells chronically exposed to an opioid. Mol Pharmacol 43: Ammer H and Schulz R (1995) Chronic activation of inhibitory -opioid receptors cross-regulates the stimulatory adenylate cyclase-coupled prostaglandin E 1 receptor system in neuroblastoma x glioma (NG108 15) hybrid cells. J Neurochem 64: Ammer H and Schulz R (1997) Chronic morphine treatment increases stimulatory beta-2 adrenoceptor signaling in A431 cells stably expressing the mu opioid receptor. J Pharmacol Exp Ther 280: Avidor-Reiss T, Zippel R, Levy R, Saya R, Ezra V, Barg J, Matus-Leibovich N and Vogel Z (1995a) Kappa-opioid receptor-transfected cell lines: modulation of adenylyl cyclase activity following acute and chronic opioid treatments. FEBS Lett 361: Avidor-Reiss T, Bayewitch M, Levy R, Matus-Leibovich N, Nevo I and Vogel Z (1995b) Adenylylcyclase supersensitization in -opioid receptor-transfected Chinese hamster ovary cells following chronic opioid treatment. J Biol Chem 270: Avidor-Reiss T, Nevo I, Levy R, Pfeuffer T and Vogel Z (1996) Chronic opioid treatment induces adenylyl cyclase V superactivation. J Biol Chem 271: Avidor-Reiss T, Nevo I, Saya D, Bayewitch M and Vogel Z (1997) Opiate-induced adenylyl cyclase superactivation is isozyme-specific. J Biol Chem 272: Cassel P and Selinger Z (1977) Mechanism of adenylate cyclase activation by cholera toxin: inhibition of GTP hydrolysis at the regulatory site. Proc Natl Acad Sci USA 74: Chen Y, Mestek A, Liu J, Hurley JA and Yu L (1993) Molecular cloning and functional expression of a -opioid receptor from rat brain. Mol Pharmacol 44:8 12. Childers SR (1991) Opioid Receptor-Coupled Second Messenger Systems. Life Sci 48: Childers SR and LaRiviere G (1984) Modification of guanine nucleotide-regulatory components in brain membranes - II. relationship of guanosine 5 -triphosphate

8 862 Ammer and Schulz Vol. 286 effects on opiate receptor binding and coupling receptors with adenylate cyclase. J Neurosci 4: Choi EJ, Wong ST, Hinds TR and Storm DR (1992) Calcium and muscarinic agonist stimulation of type-i adenylyl cyclase in whole cells. J Biol Chem 267: Collier HOJ (1984) Cellular aspects of opioid tolerance and dependence, in Opioids Past, Present and Future (Hughes J, Collier HOJ, Rance MJ and Tyres MB eds) pp , Taylor & Frances Ltd., London. DeBlasi A, O Reilly K and Motulsky HJ (1989) Calculating receptor number from binding experiments using same compound as radioligand and competitor. Trends Pharmacol Sci 10: Federman AD, Conklin BR, Schrader KA, Reed RR and Bourne HR (1992) Hormonal stimulation of adenylyl cyclase through G i -protein subunits. Nature 256: Gill DM and Woolkalis MJ (1991) Cholera toxin-catalyzed [ 32 P]ADP-ribosylation of proteins. Methods Enzymol 195: Gilman AG (1987) G Proteins: transducers of receptor-generated signals. Annu Rev Biochem 56: Hamprecht B, Glaser T, Reiser G, Bayer E and Probst F (1985) Culture and characteristics of hormone-responsive neuroblastoma x glioma hybrid cells. Methods Enzymol 109: Kawabe J, Iwami G, Ebina T, Ohno S, Katada T, Ueda Y, Homcy CJ and Ishikawa Y (1994) Differential activation of adenylyl cyclase by protein kinase C isoenzymes. J Biol Chem 269: Keith DE, Murray SR, Zaki PA, Chu PC, Lissin DV, Kang L, Evans CJ and von Zastrow M (1996) Morphine activates opioid receptors without causing their rapid internalization. J Biol Chem 271: Kim G-D, Carr C and Milligan G (1995) Detection and analysis of agonist-induced formation of the complex of the stimulatory guanine nucleotide-binding protein with adenylate cyclase in intact wild-type and 2 -adrenoceptor-expressing NG cells. Biochem J 308: Law P-Y, Hom DS and Loh HH (1984) Down-regulation of opiate receptor in neuroblastoma x glioma NG hybrid cells. J Biol Chem 259: Law P-Y, McGinn TM, Wick MJ, Erikson LJ, Evans C and Loh HH (1994) Analysis of delta-opioid receptor activities stably expressed in CHO cell lines: function of receptor density? J Pharmacol Exp Ther 271: Lee S, Rosenberg CR and Musacchio JM (1988) Cross-dependence to opioid and 2 -adrenergic receptor agonists in NG cells. FASEB J 2: MacEwan DJ, Kim G-D and Milligan G (1996) Agonist regulation of adenylate cyclase activity in neuroblastoma x glioma hybrid NG cells transfected to co-express adenylate cyclase type II and the 2 -adrenoceptor. Biochem J 318: McKenzie FR and Milligan G (1990) -Opioid receptor mediated inhibition of adenylate cyclase is transduced specifically by the guanine nucleotide-binding protein G i 2. Biochem J 367: Nestler EJ, Hope BT and Widnell KL (1993) Drug addiction: a model for the molecular basis of neural plasticity. Neuron 11: Premont RT, Jacobowitz O and Iyengar R (1992) Lowered responsiveness of the catalyst of adenylyl cyclase to stimulation by G s in heterologous desensitization: a role for camp dependent phosphorylation. Endocrinology 131: Reisine T and Bell GI (1993) Molecular biology of opioid receptors. Trends Neurosci 16: Sharma SK, Klee WA and Nirenberg M (1975) Dual regulation of adenylate cyclase accounts for narcotic dependence and tolerance. Proc Natl Acad Sci USA 72: Sunahara RK, Dessauer CW and Gilman AG (1996) Complexity and diversity of mammalian adenylyl cyclases. Annu Rev Pharmacol Toxicol 36: Sutkowski EM, Robbins JD, Tang W-J and Seamon KB (1996) Irreversible inhibition of forskolin interactions with type I adenylyl cyclase by a 6-isothiocyanate derivative of forskolin. Mol Pharmacol 50: Tang WJ and Gilman AG (1991) Type-specific regulation of adenylyl cyclase by G protein subunits. Science 254: Taussig R, Tang W-J, Hepler JR and Gilman AG (1994) Distinct patterns of bidirectional regulation of mammalian adenylyl cyclases. J Biol Chem 269: Thomas JM and Hoffman BB (1987) Adenylate cyclase supersensitivity: a general means of cellular adaptation to inhibitory agonists? Trends Pharmacol Sci 8: Thomas JM and Hoffman BB (1996) Isoform-specific sensitization of adenylyl cyclase activity by prior activation of inhibitory receptors: role of subunits in transducing enhanced activity of the type VI isoform. Mol Pharmacol 49: Vachon L, Costa T and Herz A (1987) GTPase and adenylate cyclase desensitize at different rates in NG cells. Mol Pharmacol 31: Wong YH, Federman A, Pace AM, Zachary I, Evans T, Pouysségur J and Bourne HR (1991) Mutant subunits of G i2 inhibit cyclic AMP accumulation. Nature 351: Yu VC, Eiger S, Duan D-S, Lameh J and Sadée W (1990) Regulation of cyclic AMP by the mu-opioid receptor in human neuroblastoma SH-SY5Y cells. J Neurochem 55: Send reprint requests to: Dr. Hermann Ammer, Institute of Pharmacology, Toxicology and Pharmacy, University of Munich, Königinstrasse 16, D München, Germany.