BIOL PSYCHIATRY 1991~9:273-280

273

Hyperfunctional G Proteins in Mononuclear Leukocytes of Patients with Mania Gabriel Schreiber, Sofia Avissar, Abraham Danon, and Robert H. Belmaker

In a recent study, we found that lithium inhibits the function of guanine nucleotide-binding proteins, implicating G proteins as the common site for both the antimanic and antidepressant therapeutic effects of lithium. These findings may also suggest that an altered G protein function is ofpathophysiologicat importance in bipolar affective disorder. In the present study, the coupling of both muscarinic-cholinergic receptors and beta-adrenergic receptors to pertussis toxin-sensitive G proteins or cholera toxin-sensitive G proteins was compared among untreated manic patients, lithium-treated euthymic bipolar patients, and healthy volunteers using mononuclear leukocyte (MNL) membrane preparations. Hyperactive function of G proteins was detected in untreated manic patients. Both isoproterenol-induced and carbamylcholine-induced increases in Gpp(NH)p binding capacity were twofold to threefold higher t:~an the increases observed in healthy volunteers. On the other hand, lithium-treated euthymic bipolar patients showed G protein responses to agonist activation that were no different from the healthy volunteers. Altered G protein function may be ofpathophysiological importance in bipolar affective disorder.

Introduction The most striking aspect of lithium action in bipolar affective disorder is its ability, to stabilize both poles of the disorder. Two theories that attempted to explain this effect focus on the perturbation of two second-messenger effector systems: adenylate cyclase (Fore and Valdecasas 197 I; Ebstein et al 1976; Belmaker 198 I) and phosphatidy~ositol turnover (Hallcher and Sherman 1980; Berridge et al 1982). Recently, we reported that lithium at therapeutic concentrations can alter the function of a core feature of both these signaling systems, namely, G proteins. We showed that in rat cerebral cortex, lithium inhibited the coupling of both beta-adrenergic and m~scarinic-cholinergic receptors to cholera toxin-sensitive or pertussis toxin-sensitive G proteins, Gs and Gp, respectively (Avissar et al 1988; Avissar and Schreiber 1989). These findings raise the possibility that the therapeutic effects of lithium may be mediated in part by interfering with G protein function, and by implication may infer that altered G protein function could underly

From the Beer Sheva Mental Health Centre, Ida and Solomon Stem Psychiatry Research Unit (GS, RHB) and the Department of Clinical Pharmacology (SA, AD), Ben Gurion University, Beer Sheva, Israel. Address reprint requests to Dr. G. Schreiber, Beer Sheva Mental Health Centre, P.O. Box 4600, Beer Sbeva 84170, L~ael. Received November 28, 1989; revised July 26, 1990.

© 1991 Society of Biological Psychiatry

0006-3223/91/$03.50

274

BIOL PSYCHIATRY IWI ;29:273-280

G. Schreiber et al

manic-depressive illness. Stimulation by hormones and neurotransmitters of specific guanine nucleotide binding to G proteins is an exclusive characteristic of these proteins, which in turn leads to their activation (Gilman 1987). In lymphocyte membranes, Evans et al (1987) also used guanine nucleotide binding in the study of interleukin-2 interaction with G proteins. Mononuclear leukocyte (MNL) membranes have been shown to contain functional beta-adrenergic receptors (Williams et al 1976) as well as muscafinic-cholinergic receptors (Bering et al 1987). Therefore, we have presently used adrenergic agonistinduced and cholinergic agonist-induced increases in binding of 3H-Gpp(NH)p, the nonhydrolyzable analog of GTP which has higher affinity for G proteins, to assess the possibility that G protein function is disturbed in drug-free patients with mania. Methods

Patients The group of untreated manic patients consisted of 7 male and 3 female subjects, average age 35 (20-57) years, suffering from acute mania, who had received no drug treatment for at least 1 month. Nine of the 10 patients were hospitalized shortly before blood was taken for the present study. The lithium-treated group consisted of 7 male and 3 female patients, average age 36 (22-51) years; they were euthymic bipolar outpatients who had not received drug treatment other than lithium for at least I month. Lithium levels (0.62 _ 0.17 raM) were determined from the same sample of blood that was taken for the binding experiment. All patients were diagnosed according to DSM-m-R criteria, and consented to a 60ml blood donation for the experiment. The healthy volunteer group consisted of 6 men and 4 women, average age 34 (28-42) years, from the medical staff of the Beer-Sheva Mental Health Centre. The study was approved by the .Institutional Review Board.

MNL Isolation and 3H-Gpp(NH)p Binding Assay Mononuclear leukocytes were isolated from 60 ml heparinized fresh blood of adult donors, using Ficoll-Paque gradient according to Boyum (1968). Cells were homogenized in 25 mM Tris-HCl, pH 7.4, and 1 mM dithiotreitol (DTr). The homogenate was passed through two layers of cheesecloth to remove debris, and the membranes were collected by further centrifugation at 8,000 g for 10 mira Gpp(NH)p was chosen for the binding experiments carried out in MNL for two Masons: (1) It has a higher affinity to G proteins compared with GTP. Because MNL membranes bind GTP with an affinity that is 10-fold lower than brain cortical membranes, this low affinity makes it technically difficult to achieve reliable specific binding. The at least one order of magnitude higher affinity of Gpp(NH)p to G proteins bypasses this problem; (2) It is nonhydrolyzable. In contrast to isoproterenol, carbamylcholine does not induce any detectable increases in GTP binding capacity in peripheral tissues predominantly populated by muscarinic Me-receptors, i.e., cardiac membranes (a,vissar and Schreiber 1989). Such increases, however, can be detected using the nonhydrolyzable analog of GTP, Gpp(NH)p (Avissar and Schreiber 1990). The nonhydrolyzability oi Gpp(NH)p enabled us to use the more convenient method of filtration instead of the centrifugation me'hod used previously for evaluating GTP binding (Avissar et al 1988).

Hyperfunctional G Proteins in Patients with Mania

t~IoLPSVC'mAT~V 1991~.273-280

275

Binding reactions were carried out for l0 min at room temperature in a final volt,me of 200 td. The reaction buffer consisted of 25 mM Tris-HCl, pH 7.4, I mM ATP, 1 mM Mg 2+ , 1 mM EGTA, and 1 mM DTT, with varying concentrations of 3H-Gpp(~!~p (0.05-5 ttm). Reactions were started by adding 50 ttg of membrane protein and terminate! with 5 ml of ice-cold buffer containing 10 mM Tris-HCl, pH 7.4, and 100 taM NaC~. and filtering through GF/C Whatman filters. Filters were subsequently washed twice ~ t h 3 ml of cold buffer and taken for scintillation counting. Agonists effects on 3H43pp(NH)p binding were assessed by adding isoproterenol (25 ttm) or carbamylcholine (50 ttM) to the reaction mixture. These represent the minimal concentration resulting in maximal effect of the agonist.

ADP-Ribosylation MNL membranes (3--4 mg) were suspended in I ml buffer containing 25 mM Tris-HC1, pH 7.4, 10 mM NAD, I mM ATP, 10 mM thymidine, and 100 ~M GTP. ADP ribosylation was carried out for 15 rain at 30°C by adding cholera toxin (20 ~.g/ml-~) preactivated for I0 rain at 37°C with 20 mM DTI" or pertussis toxin ( l0 tLg/ml-z) preactivated for !O rain at 30°C with 20 mM DTr. The reaction was stopped by adding 25 ml ice-cold 25 mM Tris-HCl, pH 7.4, immediately followed by centrifugation at 8000 g for 10 rain. 3HGpp(NH)p binding was then carded out as described above.

Statistical Analysis The significance of the differences in agonist-induced increases in 3H-Gpp(NH)p binding capacity between untreated manic patients, lithium-treated euthymic bipolar patients, and healthy volunteers was tested statistically by analysis of variance using the Bonferroni (Dunn) t-test for variables. Results Saturation curves for basal-specific 3H-Gpp(NH)p binding to ~ membranes from healthy volunteers, together with the nonspecific binding curve are presented in Figure 1. Specific binding reached equilibrium within 5 rain and remained constant for at least 30 min. As shown in Figure 2, the binding capacity of Gpp(NH)p to MNL membranes from healthy volunteers increased in the presence of either isoproterenol or carbamylcholine. Moreover, the increased Gpp(NH)p binding by isoproterenol and by carbamylcholine was selectively abolished in the presence of their respective antagonists, propranolol and atropine (not shown). Tools that have been particularly useful in s.*udying G proteins include a pair of bacterial-derived toxins, cholera and pertussis. Cholera toxin ribosylates Gs, and pertussis toxin ribosylates G proteins other than Gs, i.e., Gi and Go. Table 1 shows that the isoproterenol-induced increase in Gpp(NH)p binding was cholera toxin-sensitive but unaffected by pretreatment with pertussis toxin, indicating selective action of isoproterenol through Gs. By contrast, pertussis toxin blocked the carbamylcholine-induced increase in Gpp(NH)p binding, which was insensitive to cholera toxin (Table I), suggesting G proteins other than Gs as the sites of action for the muscarinic agonist. Hyperactive function of G proteins was detected in MNL membranes from untreated manic patients. Both isoproterenol and carbaraylcholine induced increases in Gpp(NH)p

276

O. Schreiber et al

BIOL PSYCHIATRY 199 ! ;29:273-280

.c 6,e ,L_ C1.

,°t

E 0

20 E 13. 0 Z

0

03

1.0

2.0

3.0

3H GpplNH)p pM Figure 1. Specific and nonspecific binding of 3H-Gpp(NH)p to mononuclear leukocyte membranes. Mononuclear leukocytes were isolated from blood samples taken from healthy volunteers. 31-1Gpp(NH)p binding was carried out in triplicate, together with triplicate control samples containing 100 I~m unlabeled GT_i~ S for determination of nonspecific binding. (e e) specific binding; (o--o) nonspecific binding. Specific binding was calculated by substracting the nonspeeific binding (in the presence of 100 I~m unlabeled GTP~ S) and the "filter binding" in the absence of membranes from total binding.

binding capacity which were twofold to threefold higher than the increases observed in healthy volunteers (Figures 2 and 3). Membranes of MNL from lithium-treated euthymic bipolar patients showed agonist-induced increases in Gpp(NH)p binding capacity similar to those found in healthy volunteers (Figures 2 and 3). Basal 3H-Gpp(NH)p binding characteristics were indistinguishable among the three groups. The affinity of the agonistactivated Gpp(NH)p binding was decreased in the drug-free or lithium-treated patients (Ka = 1.59 ± 0.4 I~M in the drug-free mania versus Ka = 0.86 ± 0.1 ILm in controls, versus Ka = 1.35 ± 0.5 I~m in lithium-treated bipolars, F = 13.34, p < 0.01).

Discussion In the present study we show that both isoproterenol and carbamylcho!ine induce increases in 3H-Gpp(NH)p bi~ding to G proteins in MNL membranes. Previously, we did not find any effect of carbamylcholine on GTP binding coupled to peripheral M2-muscarinic receptors (Avissar and Schreiber 1989). This difference may stem from the nonhydrolyzability of Gpp(NH)p (Avissar and Schreiber 1990). These findings are in accord with the results of Fleming et al (1987) who showed that in peripheral cardiac tissue, muscarinic agonists inhibit GTP-activated, but not Gpp(NH)p-activated, adenylate cyclase activity. Our results show higher G protein responses to activation by muscadnic and betaadrenergic agonists in untreated manic patients, as compared with healthy volunteers. Both muscarinic agonist-induced and beta-adrenergic agom~t-induced increases in Gpp(NH)p

alOL Psvcma~v | 99 | 29:273- 280

Hyperfunctional G Proteins in Patients with Mania

277

40 e

Figure 2. Representative comparison of agonist-induced increases in 3H-Gpp(NH)p binding to MNL membranes h'a the following three groups: Upper panel: Untreated manic patient. Maximal binding capacity ( B ~ ) in the presence of carbamylcholine was 67 pmol/mg pro~.-~ and in the presence of isoprotereno! was 73 pmo~mg prot, ~, Middle panel: Healthy volunteer. B ~ in the presence of each of the agonists was 46 pmo~mg tarot.-~. Lower panel: Lithium-treated euthymic bipolar patient. Bm~ m ~ presence of each of the agonists was 4~ pmoUmg prot. -~. (A) B ~ specific 3H-Gpp(NH)p binding; (e) in the presence of the beta-adrenergic agonist, isoproterenol, 25 v.M; fro) in the presence of the muscarinic cho|~e, gic agonist, ca~arnylcholine, 50 p.M. Concentrations of agomsts were the ~ n hnal necessary to obtain maximal agonist effect. Basal binding characteristics for 3H-Gpp(NH}p" healthy volunteers, Kd = l.l _ 0.3 p.M;-?,-~., = 39 ~ 6 pmoI x mg prot.-t; untreated manic patier~s, Ka = 0.86 _ 0.2 p.M, Bm~ = 41 - 7 pmol x mg prot.-r; lithium-treated bipolar patients, Ea = 0.8q +_ 0.2 l=l~l, B ~ = 40 ~ 7 p ~ l x mg prot. -~. No stattstically si~ficant dependence of 3H Gpp(NH)p I:,asal and agonist.fffected binding characteristics on age or gender of the subjects particiw'ing in study were c'Jsc.~ed.

20

:E =IL C

@

i o O,

E "6 E

20

¢1

t.M

J.r

I.L

0

g

20

20

4O

BqUND (prnol rng protein "1 }

binding capacity in MNL from untreated manic patients were twofold to threefold higher than those obtained in MNL from healthy volunteers. Lithium-treated euthymic bipolar patients ~how G protein responses to agonist activation that are no different from normal volunteers. Lithium was previously shown to inhibit muscarinic and beta-adrenergic receptorcoupled G protein function (Avissar et al 1988; Avissar and Schreiber 1989). Preliminary, results (S. Avissar, D.L. Murphy, G. Schreiber, unpublished results) show that lithium Table 1. Effects of Cholera Toxin-Catalyzed and PerPdssis Toxin-Catalyzed ADP-Ribosylation on Agonist-lnduced Increase in 3H-Gpp(NH)p Bending to MNL Membranes % Increase in 3H-Gpp(NH)p binding capacity

None

Cholera toxin

Pertussis toxin

Isoproterenol (25 p,M)

25.1 __- 3.1

Hyperfunctional G proteins in mononuclear leukocytes of patients with mania.

In a recent study, we found that lithium inhibits the function of guanine nucleotide-binding proteins, implicating G proteins as the common site for b...
831KB Sizes 0 Downloads 0 Views