Molecular Psychiatry (2000) 5, 308–315  2000 Macmillan Publishers Ltd All rights reserved 1359-4184/00 $15.00 www.nature.com/mp ORIGINAL RESEARCH ARTICLE ␮ ␣ -Opioid and 2-adrenoceptor stimulation of [35S]GTP␥S binding to G-proteins in postmortem brains of opioid addicts JJ Meana1,2, J Gonza´lez-Maeso2, JA Garcı´a-Sevilla1 and J Guimo´n1

1Department of Psychiatry, Faculty of , University of Geneva, HUG, Hoˆpital Belle-Ide´e, CH-1225 Cheˆne-Bourg, Switzerland; 2Department of , University of the Basque Country, E-48940 Leioa, Bizkaia, Spain

Repeated opioid administration has been associated in human brain with unaltered density of ␮-opioid receptors (agonist radioligand binding sites and immunodetected receptor

protein). These receptors are coupled to Gi/Go-proteins, which are increased in brain of heroin addicts. To assess the activity of G-proteins and their coupling to receptors after chronic opioid abuse, [35S]GTP␥S binding was quantified in postmortem prefrontal cortices of 15 opioid-dependent subjects and 15 matched controls. The stimulation of [35S]GTP␥S binding ␮ ␣ by the -opioid receptor agonist DAMGO or the 2-adrenoceptor agonist UK14304 was used as a functional measure of the status of the receptor-G-protein coupling. [35S]GTP␥S binding basal values were similar in opioid addicts (819 ± 83 fmol mg−1 of protein) and controls (918 ± 106 fmol mg−1 of protein). In opioid addicts, [35S]GTP␥S binding stimulation by DAMGO ± = ± ␮ showed a maximal effect (62 8%) and a (EC50 1.09 0.26 M) that did not differ ± = ± ␮ from the maximal effect (60 12%) and potency (EC50 2.01 0.58 M) in controls. In opioid addicts, [35S]GTP␥S binding stimulation by UK14304 was not different in maximal effect (28 ± 3%) from controls (32 ± 8%), but the potency of the agonist was decreased = ± ␮ = ± ␮ (EC50 4.36 1.81 M) when compared with controls (EC50 0.41 0.15 M). The results pro- vide a direct evidence of an apparent normal functional activity of brain ␮-opioid receptors

(Gi/Go-protein coupling) during the opioid dependence process in humans. The data also dem- ␣ onstrate a functional uncoupling of 2-adrenoceptors from G-proteins, which indicates a het- erologous desensitization of these receptors. This finding could represent an adaptive mech- anism against the decreased noradrenergic activity induced by the chronic presence of opioid . Molecular Psychiatry (2000) 5, 308–315. Keywords: opioid dependence; human brain; guanosine 5Ј-O-(3-thiotriphosphate); ␮-opioid receptor; ␣ 2-adrenoceptor; substance-related disorders

Introduction down-regulation are believed to play an important role.1,2 Thus, chronic exposure of cultured cells to ␮-Opioid receptors (OP receptor, NC-IUPHAR rec- 3 opioids results in ␮-opioid receptor down-regulation ommended nomenclature) are coupled to several trans- and desensitization (for a review, see 3–8) whereas the duction mechanisms via activation of guanine nucleo- studies of chronic opioid administration in rodents or tide-binding (G /G ) proteins. G-protein coupled i o humans reveal, in general, no change in opioid recep- receptors have been shown to adapt to the chronic tor density and/or activity (for a review, see 3,6,8–11). presence of in a process that involves desensi- Therefore, the cellular basis of opioid tolerance and tization (functional uncoupling of the receptors from dependence may involve homeostatic adaptations at signal-transducing G-proteins) and/or receptor down- the postreceptor level2 and/or regulation of other G- regulation (internalization and/or net loss of receptor protein coupled receptors sharing common activities protein). with ␮-opioid receptors.4,9,12 Repeated opioid administration leads to tolerance Morphine administration produces of and dependence phenomena but the molecular and noradrenergic cells within the locus coeruleus, cellular mechanisms governing these adaptations are whereas the withdrawal from chronic morphine treat- still uncertain, though desensitization and receptor ment results in hyperactivity of these neurons.2,13 The noradrenergic neurons of the locus coeruleus can also ␣ be inhibited by 2-adrenoceptor agonists through Correspondence: JJ Meana, Department of Pharmacology, Univer- receptor-signal transduction mechanisms shared with sity of the Basque Country, E-48940 Leioa, Bizkaia, Spain. 14 Ȱ ␮-opioid receptors. Currently, there is substantial evi- E-mail: kfpmemaj lg.ehu.es ␣ Received 28 May 1999; revised 23 November 1999; accepted 19 dence indicating that administration of 2-adrenocep- December 1999 tor agonists normalizes the central noradrenergic hyp- [35S]GTP␥S binding in brains of opioid addicts JJ Meana et al 309 eractivity associated with the opioid withdrawal ice reports following interviews with family members; 13,15 ␣ syndrome. Adaptive changes of 2-adrenoceptor and (3) the medical histories of the subjects when density in brain have been demonstrated in animal available. Most of these subjects had been used in pre- models of opioid dependence9,12 and in human heroin vious studies, which revealed marked alterations of addicts3 Such adaptations suggest that down-regu- specific brain markers related to opioid addic- ␣ 11,24,25 ± −1 lation of presynaptic 2-adrenoceptors (autoreceptors) tion. Ethanol (0.66 0.22 g L ; seven subjects) in brain cortex could represent a compensatory mech- and/or benzodiazepines (three subjects) were also anism to overcome the decrease of norepinephrine found in blood samples. Subjects who gave a positive release induced by the hypoactivity of noradrenergic toxicological presence at death of cocaine, cocaine neurons following chronic opioid administration.16 metabolites, amphetamine or amphetamine derivatives However, despite the existence of information on the were excluded from the study. The retrospective ␣ biochemical status of brain 2-adrenoceptors in heroin searching of data and the detection of opioid, metab- addicts,3 no data on the functional activity of these olites and/or other drugs in hair samples (3 cm from receptors during opioid dependence in humans has yet the root gives information on abuse within 6–12 been published. months before death) also allowed the study to be The [35S]GTP␥S binding assay has been developed restricted to those polydrug abusers who had been as a functional method to assess the receptor-mediated exposed mainly to chronic opioids (Table 1). In most activation of G-proteins in isolated membranes and his- opioid addicts (n = 9), the detection of the heroin tological sections.17 The binding of the radiolabeled metabolite 6-monoacetylmorphine (range 0.2– non-hydrolizable GTP analogue [35S]GTP␥S measures 4ngmg−1) or methadone (range 0.3–17 ng mg−1) in hair the exchange of GDP for GTP (or [35S]GTP␥S) on the G- samples indicated a long-term abuse of opioids in these protein and in this way it reflects directly the receptor subjects. The toxicological screenings were performed activation induced by different agonists.18,19 The tech- at the Unit, Institute of Forensic Medicine, nique has been successfully applied to evaluate ␮- Geneva, Switzerland, and at the Instituto Nacional de opioid receptor desensitization in cultured cells7 and Toxicologia, Madrid, Spain, using standard pro- rat brain preparations.6,10 Recently, we have demon- cedures.3,26 strated the applicability of the [35S]GTP␥S binding Control specimens were obtained at autopsy from assay to human brain membrane homogenates and sec- potential control subjects dying in the same period and tions in spite of the limitations inherent to postmortem who met the following criteria: (1) sudden and unex- tissue preparations.20,21 pected death (motor vehicle accidents, cardiac arrest); The present study has used the [35S]GTP␥S binding (2) appropriate age, sex, and postmortem delay to assay to provide direct functional information on the match each subject in the opioid addict group; (3) status of receptor-G-protein coupling in brain of absence of an apparent history of neuropsychiatric dis- chronic opioid-dependent subjects. The ability of the order or drug abuse; and (4) negative toxicological ␮ ␣ -opioid receptor agonist DAMGO and the 2-adreno- screening with the exception of ethanol (Table 2). Etha- ceptor agonist UK14304 to activate G-proteins was nol was found in blood samples from four subjects assessed in postmortem frontal cortices of opioid (1.5 ± 0.5 g L−1) of the control group, but evidence that addicts and matched controls. these subjects were not ethanol dependents was obtained. The definitive control group consisted of specimens from 15 male subjects with an age at death, Subjects and methods a postmortem delay and a tissue storage period that did Brain samples not differ between opioid addicts (Table 1) and con- Postmortem human brain samples were obtained at trols (Table 2). autopsy from the Institutes of Forensic Medicine, Geneva, Switzerland, and Bilbao, Spain. [35S]GTP␥S binding assays Specimens of prefrontal cortex (Brodmann area 9) The membrane preparation and the [35S]GTP␥S bind- were obtained from 15 male subjects who had died of ing assay were performed as reported in detail pre- an opioid overdose as determined by the medical viously.20 Briefly, cortical tissues were homogenized in − ° examiner, and they were stored at 70 C (Table 1). an ice-cold Tris buffer (50 mM Tris HCl, 3 mM MgCl2, Drug screening in blood samples confirmed the pres- 1 mM DTT, 1 mM EGTA, pH 7.4) containing 0.25 M ence of fatal concentrations of morphine or methadone sucrose and membrane P2 fractions were prepared by in all the subjects (Table 1). Generally, it is assumed sequential centrifugations and resuspensions in Tris an average serum concentration of over 0.4 ␮gml−1 buffer. Final aliquots were stored at −70°C for binding methadone is sufficient to cause death in subjects toler- assays and protein determination. ant to opioids.22 In blood specimens of heroin-related After thawing, 80 ␮gml−1 of membrane proteins were deaths, a mean concentration of 0.22 ␮gml−1 morphine incubated at 30°C for 2 h in incubation buffer (50 mM 23 has been found to be lethal. The determination of a Tris HCl, 3 mM MgCl2, 0.2 mM DTT, 1 mM EGTA, long-term history of opioid addiction was made by the 100 mM NaCl, 50 ␮M GDP, pH 7.4) containing 0.5 nM medical examiner on the basis of: (1) the presence of [35S]GTP␥S. The ␮-opioid receptor agonist DAMGO −10 −4 ␣ opioid and/or opioid metabolites on toxicological (10 –10 M) or the 2-adrenoceptor agonist UK14304 screenings in blood, urine, and scalp hair; (2) the pol- (10−9–10−3M) was added to determine receptor-stimu-

Molecular Psychiatry [35S]GTP␥S binding in brains of opioid addicts JJ Meana et al 310 Table 1 Characteristics of individual opioid addicts

Addict Age at death Postmortem delay Brain storage Opioids in blood Other drugs (␮gml−1) (years) (h) (months) (␮gml−1) Ethanol (g L−1)

1 42 28 5 Met (0.4) None 2 44 19 13 Mor (0.6) None 3 32 29 12 Mor (1.6), Cod (0.5) Ethanol (0.15) 4 40 16 23 Met (0.8) Quinine (1.2) 5 29 25 25 Met (0.5) Citalopram (3.4) 6 26 73 34 Mor (1.7) None 7 26 20 22 Met (0.3) Ethanol (0.12) 8 26 13 21 Mor (2.0) Ethanol (0.28) 9 33 16 21 Mor (1.5) Ethanol (0.87) 10 23 52 34 Met (1.3) None 11 34 26 12 Mor (+) Ethanol (1.24) 12 32 50 96 Mor (0.7) Ethanol (1.3) 13 29 49 56 Mor (0.3) None 14 19 60 40 Mor (0.5) None 15 26 25 84 Mor (+) Ethanol (+) Group values 31 ± 233± 533± 7

Group values are mean ± SEM of 15 men. Methadone (Met), morphine (Mor), codeine (Cod), ethanol and other drugs were quantified in blood samples. In two cases the presence of morphine in blood was made by qualitative methods. Therapeutic concentrations of benzodiazepines were also detected in some subjects. The postmortem delay represents the time period from death to the storage of the brain specimens.

Table 2 Characteristics of individual control subjects

Control Age at death Postmortem delay Brain storage Cause of death Drugs in blood (␮gml−1) (years) (h) (months) Ethanol (g L−1)

1 50 33 23 Cardiac arrest None 2 52 48 21 Cardiac arrest None 3 28 15 46 Accident Ethanol (2.71) 4 49 24 20 Cardiac arrest None 5 18 16 26 Accident None 6 31 62 30 Accident None 7 27 22 20 Accident None 8 29 45 20 Accident None 9 28 26 25 Accident Ethanol (2.02) 10 20 35 34 Accident None 11 30 24 14 Accident None 12 35 40 60 Accident Ethanol (0.34) 13 25 51 76 Accident Ethanol (0.93) 14 21 56 67 Accident None 15 24 20 72 Accident None Group values 31 ± 335± 437± 6

Group values are mean ± SEM of 15 men. The number (first column) indicates the control subject that was matched with each opioid addict identified with the same number in the addict group (first column in Table 1). Toxicological screening was similar to that performed in opioid addicts and included ethanol, opioids and other drugs. The postmortem delay represents the time period from death to the storage of the brain specimens.

lated [35S]GTP␥S binding. Basal binding was assumed Data analysis to be the specific [35S]GTP␥S binding in the absence of Agonist-stimulated [35S]GTP␥S binding data are agonist. Nonspecific [35S]GTP␥S binding was deter- expressed either as a percentage of basal binding mined in the presence of 10 ␮M unlabeled GTP␥S and ([stimulated-basal] × 100/basal), or as the maximal net

represented around 2% of total binding. The reaction stimulation (Emax) (stimulated-basal) expressed in was terminated by rapid vacuum filtration through fmol mg−1 of protein. Nonlinear regression analysis was Whatman GF/C glass fiber filters and the remaining performed on individual concentration-response bound radioactivity was measured by liquid scintil- curves of the agonist-induced stimulations. The con- lation spectrophotometry. centration of the agonists that elicited half-maximal

Molecular Psychiatry [35S]GTP␥S binding in brains of opioid addicts JJ Meana et al 311 Table 3 Basal and agonist-stimulated [35S]GTP␥S binding parameters in postmortem frontal cortex of opioid addicts and controls

−1 ␮ ␣ Group Basal (fmol mg -opioid receptors 2-adrenoceptors protein) DAMGO stimulation UK14304 stimulation

␮ ␮ EC50 ( M) Emax EC50 ( M) Emax (fmol mg−1 protein) (fmol mg−1 protein)

Opioid addicts (n = 15) 819 ± 83 1.09 ± 0.26 493 ± 79 4.36 ± 1.81** 212 ± 31 Controls (n = 15) 918 ± 106 2.01 ± 0.58 502 ± 101 0.41 ± 0.15 219 ± 61

EC50 represents the concentration of the agonist that induces half-maximal effect. Emax is the maximal net stimulation elicited by the agonist. Ͻ **P 0.01 when comparing log EC50 in opioid addicts vs log EC50 in controls (Student’s t-test).

effects (EC50) were obtained and normalized as log EC50 ance (ANCOVA) to control the possible influence of values to be compared.27 Data are expressed as mean interassay variations. The ANCOVA analysis of basal values ± standard error of the mean (SEM). The Stud- [35S]GTP␥S binding did not reveal significant differ- ent’s t-test was used for the statistical evaluations with ences between opioid addicts and control subjects a level of significance of P = 0.05. (F[1,27] = 0.60; P = 0.45). The ␮-opioid receptor agonist DAMGO (10−10– Drugs 10−4 M) stimulated [35S]GTP␥S binding to cortical [35S]GTP␥S (guanosine-5Ј-O-(3-[35S]thiotriphosphate, membranes in a concentration-dependent and satu- − specific activity 1250 Ci mmol 1) was purchased from rable manner with maximal stimulations of 62 ± 8% in DuPont NEN (Brussels, Belgium). DAMGO ([D-Ala2, opioid addicts and 60 ± 12% in controls (Figure 2a). 4 5 −1 MePhe , Gly-OH ]enkephalin) was obtained from The maximal net stimulation (Emax in fmol mg of Sigma Chemical Co (St Louis, MO, USA) and UK14304 protein) induced by DAMGO was also similar in both

(bromoxidine, brimonidine) was from Tocris Cookson groups of subjects (Table 3). The EC50 values of the con- Inc (Bristol, UK). centration-response curves were in the micromolar range (Table 3) and they did not differ between opioid = = Results addicts and controls (log EC50 values: t 0.94; df 28; P = 0.35). The mean concentration-response curve to 35 ␥ The basal binding of [ S]GTP S to cortical brain mem- DAMGO in opioid addicts was almost superimposed branes in opioid addicts was not significantly different to that in controls (Figure 2a) indicating the absence of = = from that in matched controls (t 0.74; df 28; desensitization of this ␮-opioid receptor-mediated = P 0.46) (Table 3). Because of the expected variability functional response (receptor/G-protein coupling) in in basal values among individual assays (Figure 1), the chronic opioid abusers. data were further evaluated by an analysis of covari- ␣ The functional status of 2-adrenoceptors was also assessed in the same membrane preparations from opioid addicts and matched controls. The [35S]GTP␥S ␣ binding stimulation through 2-adrenoceptors was con- ducted with UK14304, a selective full agonist drug. UK14304 (10−9–10−3 M) stimulated the [35S]GTP␥S binding to cortical membranes in a concentration- dependent and saturable manner with similar maximal stimulations in opioid addicts (28 ± 3%) and control subjects (32 ± 8%) (Figure 2b). Similarly, there were no differences between both groups of subjects in the −1 maximal net stimulation (Emax in fmol mg of protein) of the [35S]GTP␥S binding induced by UK14304

(Table 3). In contrast, the EC50 values of the concen- tration-response curves were clearly increased (10- fold) in opioid addicts when compared with controls 35 ␥ Figure 1 Individual basal values of [ S]GTP S binding in (Table 3) (log EC values: t = 2.91; df = 28; P = 0.007). the postmortem frontal cortex (Brodmann area 9) of opioid 50 The mean concentration-response curve to UK14304 in addicts (᭹) and control subjects (᭺). Data are expressed in fmol mg−1 of protein of [35S]GTP␥S specifically bound in the opioid addicts was rightward shifted with respect to absence of any agonist drug. Each point represents the mean that in controls (Figure 2b), indicating a desensitization of two independent measurements performed in triplicate. of this functional response (coupling to G-protein) ␣ The line represents the mean value of each group. mediated by 2-adrenoceptors in opioid addicts.

Molecular Psychiatry [35S]GTP␥S binding in brains of opioid addicts JJ Meana et al 312 with the subgroup of control subjects in whom the absence of ethanol in blood samples was established (n = 11; Table 2). No significant differences between groups could be observed either in basal values of [35S]GTP␥S binding or in DAMGO stimulated [35S]GTP␥S binding parameters. Again, the stimulation of [35S]GTP␥S binding by UK14304 displayed higher

EC50 values in the drug-free opioid addict group = ± ␮ (EC50 6.61 3.32 M) than in the drug/ethanol-free = ± ␮ control group (EC50 0.50 0.19 M) (log EC50 values: t = 1.90; df = 15; P = 0.07).

Discussion The present findings provide direct information on the ␮ ␣ functional status of -opioid receptors and 2-adreno- ceptors in postmortem brain of opioid addicts. The results of the study indicate a lack of desensitization in the [35S]GTP␥S binding stimulation by the ␮-opioid receptor selective agonist DAMGO in cortical brain membranes from chronic abusers of opioids. In con- ␣ trast, the decreased potency of the selective 2-adreno- ceptor agonist UK14304 to stimulate [35S]GTP␥S bind- ing in the same membrane preparations indicates the ␣ evidence of a marked desensitization of 2-adrenocep- tors during opioid dependence in humans. The data reported herein represent a functional confirmation of previous findings obtained in brain of opioid addicts by receptor radioligand binding studies (ie, unaltered density of ␮-opioid receptors and decreased density of ␣ 3 2-adrenoceptors). The quantification of [35S]GTP␥S binding in post- 35 ␥ Figure 2 Specific [ S]GTP S binding to cortical membranes mortem human brain and its stimulation by selective of opioid addicts (᭹) and control subjects (᭺) as a function of increasing concentrations of (a) the ␮-opioid receptor agonist receptor agonists represents a powerful technique to ␣ evaluate the activity of G-proteins (basal values) and DAMGO or (b) the 2-adrenoceptor agonist UK14304. Each point represents the mean ± SEM of the individual percentage the receptor ability to activate the nucleotide exchange stimulation over basal values for each subject shown in on these G-proteins (mainly Gi/Go-proteins in brain, see Tables 1 and 2. The curves represent the mean concentration- below).20 Basal [35S]GTP␥S binding in cortical brain response isotherms obtained from the simultaneous analysis membranes was not different between opioid addicts of all the experiments together in each group. Absolute basal and controls (Table 3; Figure 1) suggesting normal values were similar in both groups and are expressed in functioning in G-protein basal activity. This unaltered ᭹ 35 ␥ Table 3. Note that in opioid addicts ( ), the [ S]GTP S bind- basal activity in brains of opioid addicts is at variance ing stimulation by DAMGO was not altered whereas that of with the increased immunoreactivities of G␣ -, G␣ -, UK14304 was markedly rightward shifted with respect to con- 1/2 o and G␣s-proteins observed in postmortem frontal cortex trols. 26 of heroin addicts and the decrease of G␣1/2-protein density obtained in platelets of methadone-maintained To evaluate a possible influence of the type of opioid subjects.28 However, it should be noted that measure- agonist present at death in blood samples, [35S]GTP␥S ments of G-protein levels may not provide accurate binding parameters were compared between subjects estimates of functional activity of these proteins. In with blood presence of morphine and subjects with fact, a similar dissociation between basal [35S]GTP␥S

blood presence of methadone (Table 1). The basal binding activity and Gi/Go-protein immunoreactivities [35S]GTP␥S binding was not different between groups can be observed in the rat brain after chronic morphine (morphine group: 782 ± 114 fmol mg−1 of protein; treatment.6,29,30 Despite the fact that [35S]GTP␥S bind- methadone group: 891 ± 99 fmol mg−1 of protein). In a ing signal does not discriminate among the different G- similar manner, the maximal net stimulations induced proteins, the existence of compensatory changes

by the agonists DAMGO and UK14304 and their EC50 between Gi/Go-proteins and other G-proteins seems values were also found unchanged between both unlikely to explain the absence of changes in basal opioid addict subgroups (data not shown). [35S]GTP␥S binding values when opioid addicts are The subgroup of opioid addicts in whom the pres- compared with control subjects. Thus, the high abun-

ence in blood of other drugs than opioids including dance of pertussis toxin-sensitive G-proteins (Gi/Go)in ethanol was negative (n = 6; Table 1) was compared brain makes the overall measurement of [35S]GTP␥S

Molecular Psychiatry [35S]GTP␥S binding in brains of opioid addicts JJ Meana et al 313 binding as representative of the activity of inhibitory should be taken into account as a possible in vivo G-proteins.21 mechanism underlying tolerance and dependence.4 Chronic treatment of rats with morphine has been In contrast to responses mediated by ␮-opioid recep- 35 ␥ ␣ shown to result in decreased G-protein activity, as tors, the [ S]GTP S binding stimulation by the 2- measured by basal [35S]GTP␥S binding, restricted to adrenoceptor agonist UK14304 showed a clear desensi- the locus coeruleus whereas other brain areas (cortex, tization in the postmortem frontal cortex of the same basal ganglia, hypothalamus, amygdala) did not show opioid addicts (Table 3; Figure 2b). This result rep- apparent changes.6,10 The present findings in the fron- resents a functional demonstration that the down-regu- ␣ tal cortex of opioid addicts are similar to those lation of 2-adrenoceptor agonist binding sites obtained in the cerebral cortex of morphine-dependent ([3H]clonidine binding) previously observed3 consti- rats.10 Whether or not changes in G-protein basal tutes a relevant biochemical adaptation during the activity also occur in the locus coeruleus of opioid development of opioid dependence in humans. ␣ addicts remains unknown. However, the changes in G- Noradrenergic neurons in the CNS possess both 2- proteins induced by chronic morphine in rodents and adrenoceptors and ␮-opioid receptors sharing effector 11,26,30–32 14 ␣ by the long-term abuse of opioid in humans do mechanisms of transduction. In brain cortex, 2-adre- not seem to be restricted to those G-proteins within the noceptors and ␮-opioid receptors are colocalized as locus coeruleus as initially found.29 inhibitory receptors on noradrenergic terminals arising 33 The ability of the ␮-opioid receptor agonist DAMGO from the locus coeruleus. Cross tolerance between ␣ to stimulate the binding of [35S]GTP␥S to cortical mem- opioids and 2-adrenoceptor agonists has been clearly 12,34–37 branes was similar in opioid addicts and control sub- demonstrated in several tests. It has also been jects (Table 3; Figure 2a). This finding indicates a lack shown that chronic morphine administration in rats ␣ 9,12 of desensitization or uncoupling of ␮-opioid receptors decreases the density of brain 2-adrenoceptors. from G-proteins in brain of chronic opioid abusers. Therefore, the down-regulation and functional desensi- ␣ This agrees well with previous data reporting that ␮- tization of 2-adrenoceptors appears to be one of the opioid receptor agonist binding ([3H]DAMGO binding) adaptive mechanisms that would tend to overcome the 3 decreased function of noradrenergic pathways during is not down-regulated in brains of heroin addicts and 16,38 that ␮-opioid receptor immunoreactivity is unaltered chronic administration of opioids. This type of het- in brains of heroin or methadone-dependent subjects.11 erologous desensitization may play the functional role of a tolerance mechanism in brain areas such as the Consistent with these findings in humans, the frontal cortex where the ␮-opioid receptor population [35S]GTP␥S binding stimulation by DAMGO is not remains resistant to down-regulation during chronic modified in several brain areas of morphine-dependent opioid administration. rats.6,10 However, in contrast to data in mammals, rel- In the human frontal cortex, the ␣ -adrenoceptor evant decreases in potency (EC ) and/or in maximal 2 50 population appears to be predominantly of the ␣ -sub- effect (E )of␮-opioid receptor-stimulated 2A max type.39,40 The ␣ -adrenoceptor also appears to be the [35S]GTP␥S binding have been reported in cultured 2A autoreceptor that regulates the release of norepi- cells chronically treated with opioids.5,7 nephrine in the human frontal cortex.41 Therefore, the The issue of ␮-opioid receptor regulation in experi- ␣ results of this study indicate that subsensitivity of 2A- mental models of chronic opioid administration has adrenoceptors plays a major regulatory role during produced a voluminous literature with apparent con- opioid dependence in humans. tradictory data reporting down-regulation, up-regu- Taking all the previous points together makes it lation, or no changes of biochemical and functional tempting to speculate that ␣ -adrenoceptor (and/or 3–9 2 receptor parameters (see and references therein). other brain receptors) down-regulation and desensitiz- Differences in exposure time, dose, type of opioid drug, ation may prevent in vivo the down-regulation of ␮- route of administration, tissue, and assay conditions opioid receptors and their functional uncoupling from could account for such discrepancies. Furthermore, the G-proteins. In brain, one can assume the existence of ␮ absence of conclusive changes in -opioid receptor an equilibrium state between different receptors shar- density and functional coupling to G-proteins could be ing similar functional activities on the same neural explained by the development of compensatory pathway. In this context, the cellular adaptations to the responses, at the postreceptor level, to the sustained chronic presence of agonists could represent a coordi- effect exerted by the chronic presence of the opioid nate reaction between receptor-effector systems, in ␣ drug. Thus, up-regulation of the cAMP system is a well which 2A-adrenoceptors display in frontal cortex established cellular adaptation that overcomes in selec- higher sensitivity than ␮-opioid receptors to the intra- ted brain areas the inhibitory effect of opioids on this cellular mechanisms promoting down-regulation. transduction system.1 Alternatively, homeostatic Obviously, this must mean that in cell cultures, an in changes in other transduction systems of ␮-opioid vitro system expressing a limited number of different receptors (K+ and Ca2+ channel conductance, mitogen- receptor structures, the equilibrium state between activated protein kinase pathway) might also occur.2 receptors changes with respect to that in brain. Thus, ␣ On the other hand, the heterologous desensitization of in the absence of 2A-adrenoceptor expression, the other brain receptors sharing common coupling sys- chronic exposure to opioid agonists might result in a tems and cellular locations with ␮-opioid receptors selective loss of ␮-opioid receptor binding sites and/or

Molecular Psychiatry [35S]GTP␥S binding in brains of opioid addicts JJ Meana et al 314 in a receptor uncoupling from the transduction path- 5Ј-[␥-[35S]thio]-triphosphate binding. Proc Natl Acad Sci USA ways activated by this receptor type. 1995; 92: 7242–7246. 18 Selley DE, Sim LJ, Xiao R, Liu Q, Childers SR. ␮-Opioid receptor- stimulated guanosine-5Ј-O-(␥-thio)-triphosphate binding in rat thalamus and cultured cell lines: signal transduction mechanisms Acknowledgements underlying agonist efficacy. Mol Pharmacol 1997; 51: 87–96. 19 Alt A, Mansour A, Akil H, Medzihradsky F, Traynor JR, Woods JH. This study was supported by grants from the FNSRS Stimulation of guanosine-5Ј-O-(3-[35S]thio)triphosphate binding by (32-57066.99), Switzerland to JAG-S, the Basque endogenous opioids acting at a cloned mu receptor. J Pharmacol Government (PI-98/8) and the University of the Basque Exp Ther 1998; 286: 282–288. Country (G13/98) Spain to JJM. JG-M was supported by 20 Gonza´lez-Maeso J, Rodrı´guez-Puertas R, Gabilondo AM, Meana JJ. Characterization of receptor-mediated [35S]GTP␥S binding to a fellowship from the MEC, Madrid, Spain. The authors cortical membranes from postmortem human brain. Eur J Pharma- wish to thank the Institutes of Forensic Medicine of col (in press). Geneva (Dr R La Harpe) and Bilbao (Dr R Alcaraz) for 21 Rodrı´guez-Puertas R, Gonza´lez-Maeso J, Meana JJ, Pazos A. Auto- their cooperation in the study. JAG-S is a member of radiography of receptor-activated G-proteins in postmortem human the Institut d’Estudis Catalans (Barcelona, Spain). brain. 2000; 96: 169–180. 22 Merrill J, Garvey T, Rosson C. Methadone concentrations taken as indicating deaths due to overdose need to be reviewed. Br Med J 1996; 313: 1481. References 23 Goldberg BA, Cone EJ, Grant TM, Caplan YH, Levine BS, Smialek JE. Disposition of heroin and its metabolites in heroin-related 1 Nestler EJ. Molecular mechanisms of drug addiction. J Neurosci deaths. J Anal Toxicol 1994; 18: 22–28. 1992; 12: 2439–2450. 24 Garcı´a-Sevilla JA, Ventayol P, Busquets X, La Harpe R, Walzer C, 2 Nestler EJ, Aghajanian GK. Molecular and cellular basis of addic- Guimo´n J. Marked decrease of immunolabeled 68 kDa neurofila- tion. Science 1997; 278: 58–63. ment (NF-L) proteins in brains of opiate addicts. Neuroreport 1997; 3 Gabilondo AM, Meana JJ, Barturen F, Sastre M, Garcı´a-Sevilla JA. 8: 1561–1570. ␮-Opioid receptor and ␣ -adrenoceptor agonist binding sites in the 2 25 Ozaita A, Escriba´ PV, Ventayol P, Murga C, Mayor Jr F, Garcı´a- postmortem brain of heroin addicts. 1994; Sevilla JA. Regulation of G protein-coupled receptor kinase 2 in 115: 135–140. brains of opiate-treated rats and human opiate addicts. J Neuro- 4 Noble F, Cox BM. Differential desensitization of ␮- and ␦-opioid chem 1998; 70: 1249–1257. receptors in selected neural pathways following chronic morphine 26 Escriba´ PV, Sastre M, Garcı´a-Sevilla JA. Increased density of guan- treatment. Br J Pharmacol 1996; 117: 161–169. ine nucleotide-binding proteins in the postmortem brains of heroin 5 Breivogel CS, Selley DE, Childers SR. Acute and chronic effects of addicts. Arch Gen Psychiatry 1994; 51: 494–501. opioids on ␦ and ␮ receptor activation of G proteins in NG108–15 27 Hancock AA, Bush EN, Stanisic D, Kyncl JJ, Lin CT. Data normaliz- and SK-N-SH cell membranes. J Neurochem 1997; 68: 1462–1472. ation before statistical analysis: keeping the horse before the cart. 6 Selley DE, Nestler EJ, Breivogel CS, Childers SR. Opioid receptor- Trends Pharmacol Sci 1988; 9: 29–32. coupled G-proteins in rat locus coeruleus membranes: decrease in activity after chronic morphine treatment. Brain Res 1997; 746: 28 Manji HK, Chen G, Potter W, Kosten TR. Guanine nucleotide bind- 10–18. ing proteins in opioid-dependent patients. Biol Psychiatry 1997; 41: 7 Elliot J, Guo L, Traynor JR. Tolerance to ␮-opioid agonists in 130–134. human neuroblastoma SH-SY5Y cells as determined by changes in 29 Nestler EJ, Erdos JJ, Terwilliger R, Duman RS, Tallman JF. Regu- guanosine-5Ј-O-(3- [35S]-thio)triphosphate binding. Br J Pharmacol lation of G-proteins by chronic morphine treatment in the rat locus 1997; 121: 1422–1428. coeruleus. Brain Res 1989; 476: 230–239. 8 Kaplan GB, Leite-Morris KA, Sethi RK, McClelland EG. Regulation 30 Ventayol P, Busquets X, Garcı´a-Sevilla JA. Modulation of immuno- ␣ ␤ of G protein-mediated adenylyl cyclase in striatum and cortex of reactive protein kinase C- and isoforms and G proteins by acute opiate-dependent and opiate withdrawing mice. Brain Res 1998; and chronic treatments with morphine and other opiate drugs in rat 788: 104–110. brain. Naunyn-Schmieberg’s Arch Pharmacol 1997; 355: 491–500. ␣ 31 Parolaro D, Rubino T, Gori E, Massi P, Bendotti C, Patrini G et al. 9 Ulibarri I, Garcı´a-Sevilla JA, Ugedo L. Modulation of brain 2-adre- noceptor and ␮-opioid receptor densities during morphine depen- In situ hybridization reveals specific increases in G␣s and G␣O dence and spontaneous withdrawal in rats. Naunyn-Schmiede- mRNA in discrete brain regions of morphine-tolerant rats. Eur J berg’s Arch Pharmacol 1987; 336: 530–537. Pharmacol 1993; 244: 211–222. 10 Sim LJ, Selley DE, Dworkin SI, Childers SR. Effects of chronic mor- 32 Van Vliet BJ, Van Rijswijk ALCT, Wardeh G, Mulder AH, Schoffel- phine administration on ␮ opioid receptor-stimulated [35S]GTP␥S meer ANM. Adaptive changes in the number of Gs- and Gi-proteins autoradiography in rat brain. J Neurosci 1996; 16: 2684–2692. underlie adenylyl cyclase sensitization in morphine-treated rat stri- 11 Garcı´a-Sevilla JA, Ventayol P, Busquets X, La Harpe R, Walzer C, atal neurons. Eur J Pharmacol 1993; 245: 23–29. Guimo´n J. Regulation of immunolabelled ␮-opioid receptors and 33 Simonato M. The neurochemistry of morphine addiction in the protein kinase C-␣ and ␰ isoforms in the frontal cortex of human neocortex. Trends Pharmacol Sci 1996; 17: 410–415. opiate addicts. Neurosci Lett 1997; 226: 29–32. 34 Solomon RE, Gebhart GF. Intrathecal morphine and clonidine: anti- 12 Smith CB, Moises HC, Spengler RN, Hollingsworth PJ. Changes in nociceptive tolerance and cross-tolerance and effects on blood ␣ 2-adrenoceptor number and function in brains of morphine- pressure. J Pharmacol Exp Ther 1988; 245: 444–454. dependent rats. Eur J Pharmacol 1989; 161: 111–119. 35 Stevens CW, Monasky MS, Yaksh TL. Spinal infusion of opiate and ␣ 13 Aghajanian GK. Tolerance of locus coeruleus neurones to mor- 2-agonists in rats: tolerance and cross-tolerance studies. J Pharma- phine and suppression of withdrawal response by clonidine. Nat- col Exp Ther 1988; 244: 63–70. ure 1978; 276: 186–188. 36 Bartoletti M, Gaiardi M, Gubellini C, Bacchi A, Babbini M. Cross- ␣ 14 Aghajanian GK, Wang YY. Common 2- and opiate effector mech- tolerance between morphine and clonidine: a study on motility in anisms in the locus coeruleus: intracellular studies in brain slices. rats. 1989; 28: 1159–1162. Neuropharmacology 1987; 26: 793–799. 37 Kalso EA, Sullivan AF, McQuay HJ, Dickenson AH, Roques BP. 15 Gold MS, Redmond Jr DE, Kleber HD. Clonidine blocks acute opi- Cross-tolerance between mu opioid and alpha-2 adrenergic recep- ate-withdrawal symptoms. Lancet 1978; ii: 599–602. tors, but not between mu and delta opioid receptors in the spinal 16 Done C, Silverstone P, Sharp T. Effect of naloxone-precipitated cord of the rat. J Pharmacol Exp Ther 1993; 265: 551–558. morphine withdrawal on noradrenaline release in rat hippocampus 38 El-Kadi AOS, Sharif SI. The influence of chronic treatment with in vivo. Eur J Pharmacol 1992; 215: 333–336. clonidine, yohimbine and idazoxan on morphine withdrawal. Psy- 17 Sim LJ, Selley DA, Childers SR. In vitro autoradiography of recep- chopharmacology 1997; 132: 67–73. ␣ tor-activated G proteins in rat brain by agonist-stimulated guanylyl 39 Sastre M, Garcı´a-Sevilla JA. 2-Adrenoceptor subtypes identified

Molecular Psychiatry [35S]GTP␥S binding in brains of opioid addicts JJ Meana et al 315 by [3H]RX821002 binding in the human brain: the agonist guanox- delineation of [3H]RX821002 binding to membranes and tissue sec- abenz does not discriminate different forms of the predominant tions. Eur J Pharmacol 1996; 310: 83–93. ␣ 2A-subtype. J Neurochem 1994; 63: 1077–1085. 41 Raitieri M, Bonanno G, Maura G, Pende M, Andrioli GC, Ruelle A. ␣ ␣ 40 Grijalba B, Callado LF, Meana JJ, Garcı´a-Sevilla JA, Pazos A. 2- Subclassification of release-regulating 2-autoreceptors in human Adrenoceptor subtypes in the human brain: a pharmacological brain cortex. Br J Pharmacol 1992; 107: 1146–1151.

Molecular Psychiatry