Yaseen, M. A., Pedley, K. C., and Howell, S. L. (1982) Biochem. (1987) Diabetologia 30,575 (abstr.) J. 206,81-87. 18. Ullrich, S., and Wollheim, C. B. (1984) J.
Vol. 264, No. 2, Isaue of January 15, pp. 973-980,1989 Printed in U.S.A.
THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.
Suppression of Insulin Release by Galanin and Somatostatin Is Meditated by a G-protein AN EFFECT INVOLVING REPOLARIZATION AND REDUCTION IN CYTOPLASMIC FREE Ca2+ CONCENTRATION* (Received for publication, June 3, 1988)
Thomas NilssonS, Per Arkhammar$, Patrik Rorsmansll, and Per-Olof BerggrenSn ** From the $Department of Medical Cell Biology, Biomedicum, Box 571, Uppsala University, S-751 23 Uppsala, Sweden, the §Department of Medical Physics, Box 330 31, University of Gothenburg, S-400 33 Gothenburg, Sweden,and the [(Departmentof Endocrinology, Box 60 500, Karolinsku Institute, Karolinska Hospital, S-104 01 Stockholm, Sweden
The effects of galanin and somatostatin on insulin release, membrane potential, andcytoplasmic free Ca2+ concentration ([Ca2+]i)were investigated using @-cells isolated from obese hyperglycemic mice. Whereas insulinreleasewas measured in a column perifusion system, membrane potential and[Ca2+]iwere measured with the fluorescent indicators bisoxonol (bis-(1,3diethy1thiobarbiturate)trimethineoxonol)and quin 2, in cell suspensions in a cuvette. Galanin (16 nM) and somatostatin (400 nM) suppressed glucose-stimulated insulin release in parallel to promoting repolarization and a reduction in [Ca2+Ii.The reduction in [Ca2+Ii comprised an initial nadir followed by a slow rise and the establishment of a new steady state level. The slow rise in [Ca2+]iwas abolished by 60 PM D-600, a blocker of voltage-activated Ca2+channels. Both peptides suppressed insulin release even when [Ca2+]iwas raised by 25 mM K+. Under these conditions the inhibition of insulin release was partly reversed by an increase in the glucose concentration. Addition of 6 mM Ca2+to a cell suspension, incubated in the presence of 20 mM glucose and either galanin, somatostatin, or the a2adrenergic agonist clonidine (10 nM), induced oscillations in [Ca2+]i,this effect disappearing subsequent to the addition of D-600. Theeffects of galanin, somatostatin, and clonidine on [Ca2+Iiwere abolished in &cells treated with pertussis toxin. In accordance with measurements of [Ca2+Ii, treatmentwithpertussis toxin reversed the inhibitory effect of galanin on insulin release. Theinhibitory action of galaninand somatostatin on insulin release is probably accounted for by not only a repolarization-induced reduction in [Ca2+]iand a decreased sensitivity of the secretory machinery to Ca2+,but also by a direct interaction with the exocytotic process. It is proposed that these effects are mediated by a pertussis toxin-sensitiveGTP-binding protein.
* This work was supported by Grants 12x-562,19x-00034,and 12x08647 from the Swedish Medical Research Council, the Medical Faculty of Uppsala University, The Swedish Diabetes Association, Novo Industry, Swedish Hoechst, The Bank of Sweden T e r c e n t e n y Foundation, Nordic Insulin, Clas Groschinskys, Magn. Bergvalls, Ake Wibergs, Syskonen Svenssons, 0. and E. Johanssons, Amundssons and theFamily Ernfors foundations. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solelyto indicate this fact. II Recipient of a postdoctoral fellowship from the Swedish Medical Research Council. ** To whom correspondence and reprint requests should be addressed Dept. of Medical Cell Biology, Biomedicum, Box 571, Uppsala University, S-751 23 Uppsala, Sweden.
The stimulus-secretion coupling in the pancreatic @-cellis regulated by the interplay between nutrient secretagogues, of which glucose is the most important, and different receptor agonists (for a review, see Refs. 1-3). Whereas hormones may exert their action in a paracrine manner or by reaching the @-cellsvia circulation, neurotransmittors are released by the rich neural network of autonomic nerve fibers supplying the endocrine part of the pancreas (4, 5). Hormones and neurotransmittors can eitherbe stimulatory or inhibitory in terms of influence on insulin secretion. Thus, stimulation of insulin release by glucagon is believed to result from an increase of the @-cellcontent of cyclic AMP (CAMP) (4)and that of cholinergic agonists by the activation of phospholipase C, as well as by increasing the spiking activity of the @-cell(6, 7). However, the mechanisms whereby inhibitory hormones and neurotransmittors exert their action on insulin release are poorly understood. Paracrine inhibition of insulin release by somatostatin is a well established phenomenon, which has been suggested to be due to decreased formation of CAMP in the @-cells(4,8). In pituitary cells such an interaction with the adenylate cyclase system has also been described (9, lo), and in bothcell types somatostatin induces hyperpolarization anda decrease in cytoplasmic free Ca2+concentration ([Ca”]i) (10-14). In addition, studies on cells of pituitary origin have demonstrated that somatostatin suppresses currents through voltage-activated Ca2+channels (13, 14). The pancreatic @-cell isone of the few cells that respond to a-adrenergic stimulation with inhibition of its physiological function (15). We have recently demonstrated, using the aZadrenoceptor agonist clonidine, that inhibition of insulin release is paralleled by repolarization, a decrease in [Ca2+Ii, and suppression of the Ca2+current through the voltage-activated Ca2+channels (16, 17). Inhibition of the adenylate cyclase system has been suggested to account for the inhibitory action of a-adrenergic agonists, but such a mechanism cannot solely explain the multiple effects evoked by the agonist (18, 19). Apart from adrenergic nerve endings, nerves containing the recently discovered neuropeptide galanin have been located in close connection to theendocrine part of the pancreas (20). This 29-amino acid peptide inhibits insulin release both in vivo and in vitro (20, 21). Preliminary studies performed in vitro revealed that galanin-induced inhibition of insulin release was associated with repolarization and a decrease in [Ca2+],(21). Galanin is not restricted to the pancreas but is widespread throughout the body where it exerts a number of effects, including inhibition of transmittor release from neurons (22, 23). In the present study we were interested in performing a
973
974
Inhibition of Insulin Release Galanin by
more thorough investigation of the mechanisms whereby galanin and somatostatin affect the pancreatic 0-cell. The purpose of this investigation was to define a possible common regulatory pathway whereby neurotransmittors andhormones inhibit insulin release. Inthiscontext we focused on the possible involvement of mechanisms regulating membrane potential and [Ca"+]i. By using islet cell aggregates isolated from obese hyperglycemic mice these parameters were studied and related to insulin release. From the general cell biological point of view, results obtained from studies on pancreatic 8cells should be of significance also for the understanding of the cellular mechanisms(s) whereby galanin and somatostatin exert their effects in other tissues.
and Somatostatin
of intracellular Ca2+, which might mask small changes in [Ca2+Ii(33), the intracellular concentrationof quin 2 obtained with this loading procedure has previously been demonstrated not to interfere with glucose-stimulated insulin release (34). From these experiments, it seems, therefore, fair to conclude that both peptides are without significant effect on either resting [Ca*+]ior the voltage-activated Ca2+channels. As shown in Fig 2, traces A and B, the increase in [Ca2+]i promoted by a high concentration of glucose was reversed by the addition of 16 nM galanin, as demonstrated previously (21). After a couple of minutes [Ca"]i slowly rose to a higher level, although still lower than in the absence of peptide. In traces C and D it is demonstrated that somatostatin produces similar changes in [Ca2+Iias does galanin. In this case, howEXPERIMENTALPROCEDURES ever, the decrease in [Ca2+]iwas notas marked, and the Materials-Pertussis toxin was bought from List Biological Labo- plateau level was only slightly lower than that obtained with ratories, Campbell, CA, and D-600 was donated by Knoll AG. Galanin glucose stimulation alone. This is not likely to be explained was obtained from Peninsula Laboratories, Belmont, CA, whereas by rapid degradation of somatostatin, since a second addition quin 2/acetoxymethyl ester and somatostatin were bought from Sigma. Bis-(1,3-diethylthiobarbiturate)trimethineoxonol(bisoxonol)' only evoked a minor and transient decrease in [Ca"]i (data and Bio-Gel P4 polyacrylamide beads were from Molecular Probes, not shown). The latter increases in [Ca2+]iin the presence of Junction City, OR and Bio-Rad, respectively. Diazoxide and forskolin either galanin or somatostatin were completely reversed by were kind gifts from Schering Corp. and Farbwerke Hoechst, respec- the addition of D-600 (Fig. 2, traces B and D),suggesting the tively. involvement of Ca2+ influx through voltage-activated Ca2+ Animals and Preparation of Cells-Adult obese hyperglycemicmice (oblob)of both sexes were taken from a local non-inbred colony (24) channels. In the presence of either of the peptides, 25 mM K' to the and starved overnight. The islets were isolated with collagenase, and increased [Caz+]i and D-600lowered[Ca*+]idown cell suspensions comprising more than 90% &cells (25) were prepared resting level (Fig. 2, traces A and C ) . These results further and cultured overnight as described previously (26, 27). support the notion that galanin andsomatostatininhibit Medium-The medium used was a Hepes buffer, pH 7.4,containing insulin release, at least partly, by lowering [Ca*+]i,utilizing a 1.28 mM Ca2+ with Cl- as the sole anion (28), unless otherwise mechanism(s) other than direct blockage of the voltage-actiindicated. In all experiments, except for estimation of membrane potential, the medium was supplemented with 1 mg/ml bovine serum vated Ca2+ channels. As observed also in the absence of agonist (Fig. 2, trace E), the K+-induced increase in [Ca2+]i albumin. Measurements of [Ca"]i and Membrane Potential-Cell suspen- was more transient in the presence than in theabsence of 20 sions were incubated for 45 min with 5 PM quin 2/acetoxymethyl mM glucose (cf. Fig. l),probably accounted for by increased ester resulting in a loading of about 1.3 nmol of quin 2/106 cells (16). Ca2+buffering induced by the sugar. [Ca"]i was measured and calibrated as described previously (29). When studying the effects of galanin and somatostatin on Qualitative changes in membrane potential were estimated with the fluorescent dye bisoxonal (30). The dye was used a t a concentration membrane potential,as judged by bisoxonol fluorescence, of 150 nM, and fluorescence was measured according to a previous under similar conditions as in Figs. 1 and 2, none of the description (31).All traces shown are typical for experiments repeated peptides had any effect under nonstimulatory conditions, i.e. with at least three different cell preparations. in theabsence of glucose (Fig.3, traces A and B). As expected, Measurements of Insulin Release-The dynamics of insulin release increasing the extracellular concentration ofK' with 25 mM were studied by perifusing 0.5-1 X lo6 cells mixed with Bio-Gel P4 depolarized the cells. Trace C demonstrates the increase in polyacrylamide beads in a 0.5-ml column at 37 "C (32). The flow rate was 0.3 ml/min, and 1-2 min fractions were collected and analyzed membrane potential evoked by20 mM glucose and shows that for insulin radioimmunologically, using crystalline rat insulin as the galanin repolarizes the cells. This response pattern was simstandard. The datashown are typical for experiments performed with ilar using somatostatin, with the exception that themembrane three different cell preparations. potential tended to increase slightly during the observation period (trace D). Under these conditions, addition of 25 mM RESULTS K+ evoked depolarization (truces C and D). Although the In Fig. 1, trace A , 16 nM galanin was added to quin 2-loaded actual method is less suitable for measuring small shifts in islet cell aggregates suspended in a medium lacking glucose. membrane potential, it is of interest to note that theobserved Addition of galanin neither changed basal [Ca"+li nor pre- changes subsequent to the addition of either galanin or SOvented K+-induced opening of voltage-activated Ca2+chan- matostatin seem to reflect those described previously for nels and the subsequent increase in [Ca2+]i.The increase was [Ca2+Ii.To investigate the possibility that the peptides lowreversed by the addition of 50 p~ D-600, a blocker of voltage- ered membrane potential by promoting C1- influx, we tested activated Ca2+channels. Since the experimental protocol used the effects of galanin and somatostatin in C1"deficient mein trace A does not exclude that galanin exerts minor effects dium (Cl- replaced by glutamate, final concentration ofC1on the Ca2+channels, the peptide was supplemented after the about 3 mM). However, also under these conditions the pepaddition of K+ (trace C). Neither, in this case, was there a tides repolarized the cells in the presence of 20 mM glucose change in [Ca2+]i.In the next set of experiments we tested (data not shown). It should be noted that using bisoxonol on the effect of 400 nM somatostatin under similar conditions. a population of unsynchronized cells does not permit resoluAs indicated in traces B and D,somatostatin did not affect tion of the intricate electrical response pattern obtained subbasal [Ca2+Iior interfere with the opening of voltage-activated sequent to stimulation of pancreatic B-cells with glucose. Ca2+channels. Although quin 2 imposes some extra buffering Furthermore, the response time is slow, since the changes in fluorescence are dependent on the rateof diffusion of the dye. The abbreviations used are: bisoxonol, bis-(1,3-ðylthiobarbiturate)trimethineoxonol); Hepes, 4-(2-hydroxyethyl)-l-piperazine-Nevertheless, despite these drawbacks such measurements ethanesulfonic acid; GTP-yS, guanosine 5'-04thiotriphosphate); G- still provide an estimation of qualitative changes in membrane potential. protein, GTP-binding protein.
975
Inhibition of Insulin Release by Galanin and Somatostatin
[Ca*+] (nM)
270 120 -
480
FIG.1. Effects of galanin and so-
-
25 mM
i
50 UM
1
16 nM
Galanin
480 250 140
matostatin on [CaB+Iiunder basal conditions (traces A and B ) and when [Cas+], is raised by the addition of 25 mM K+ (traces C and D).
-
2 5 rnM
50 UM
K'
0-600
B
400 nM Somatostatin
-
16 nM Galanln
4 0 0 nM Somatostaton
C
D-600
D-600
D
I
;::I 130
-
120
-
LA
2 MIN
2 5 mM K+
I
16 nM Galanin
230 -
450
110
50 U M D-600
2 0 mM Glucose
A 350 230
-
150
16 nM Galanin
-
2 0 mM Glucose
50 U M
6
D-600
1
-
2 5 mM K+
C
400 nM Somatostatin
Somatostatin
560" 260
-
110
2 9 mM Glucose
2 5 mM K+
370 -
400 120
200
-
100
-
310
-
2 0 mM Glucose
400 nM
E
50 UM D-600
D
2 0 mM Glucose
2 MIN
1
-
FIG.2. Effects of galanin, somatostatin, K+, and D-600 on [Ca2'Ii during stimulation with 20 mM glucose. Studies on pituitary cells and neuroblastoma-glioma cells have revealed that somatostatin reduces Ca2+currents through voltage-activated Ca2+channels (13, 14, 35). Similar effects of a-adrenergic agonists have been described for sensory neurons and pancreatic P-cells (17, 36). To our knowledge, such studies have not been performed with galanin. Although no support for a direct interaction of galanin or somatostatin with the voltage-activated Ca2+channels in the P-cells wasobtained in the previously described experiments, we decided to study this problem further by applying another experimental protocol. After exposing the cells to either galanin,somatostatin, or the a2-adrenergic agonist clonidine (Fig. 4), we investigated the effects of raising the medium concentration of Ca2+. Addition of 5 mM Ca2+ to glucosestimulated cells, in the presence of galanin, promoted a marked and fast increase in [Ca2+]i followed by a decrease,
which turned into an oscillatory pattern of [Ca2+];changes decaying with time (truce A ) . Similar results were obtained with both somatostatin (truce B ) and 10 nM clonidine (truce C). It is noteworthy that in the absence of the agonists no oscillations emerged and only thetransient increase in [Ca2+];was observed (trace D).In all situations, [Ca2+];was decreased by the addition of D-600, implying that the oscillations in[Ca2+Iiinvolve the participation of voltage-activated Ca2+channels. The fact that an increase in the extracellular Ca2+ concentration indeed can increase [Ca2+Iidespite the presence of either galanin, somatostatin, or clonidine suggests that theagonists may have a certain influence directly on the voltage-activated Ca2+ channels. However, direct measurements of the Ca2+current have to be performed before such an effect can be clearly established. Attempts to demonstrate oscillatory changes in membrane potential, using bisoxonol
Inhibition of Insulin Release by Galanin and Somatostatin
976
25 mM
A
B
Somatostatin
Z
d
-
0 -
FIG. 3. Effects of galanin, somatostatin, ar.d K+ onmembrane potential under basal conditions ( A and B ) andduringstimulationwith 20 mM glucose (C and D )
.
t4
16 nM Galanin
4
2w
C 25 mM
~
Glucose
0
Glucose
. . J
16 nM Galanon
5 mM 50 uM D-BOO
I
I
A
5 mM
400 nM Somatostatin
450 330 100
-
ca2' I
50 uM D-800
I
20 mM
B
I
Glucose
-
* 5 mM
1 0 nM Clonidine
3oo 220
-
-
120-
2 0 mM Glucose
I
4
Ca2+
50 uM 0-800
I
1
C
5 mM ca2+
380 280
-
130-
subsequent addition of 20 mM glucose evoked a pronounced stimulatory effect. In the lower trace, closed symbols, 16 nM galanin was present throughout the experiment. Under this condition high concentrations of K+ elicited only a minor release of insulin, despite the rise in [Ca2+]i(cf. Fig. 1, trace A). Moreover, as expected from panel A, the combination of 20 mM glucose and high K+ promoted insulin release even in the presence of galanin. Thus, thispeptide suppresses insulin release despite increased [Ca2+]i,an effect counteracted by glucose. For a more detailed investigation of the influence of glucose, the experiments shown in panel B were repeated using 4 mM glucose in the basal medium. As shown in panel C, open symbols, stimulation with 25 mM K+ produced a pronounced release of insulin, which was further augmented by the addition of 16 mM glucose. In the continuous presence of galanin, high K+ stimulated insulin release to a greater extent when the basal medium contained 4 mM glucose compared to 0 mM glucose (cf. panelB),but still less than in the absence of the peptide. Interestingly, there was an additional stimulation of insulin release when the concentration of the sugar was raised to 20 mM. In Fig. 6 results obtained from similar experiments using somatostatin are presented. As evident from panel A , 400 nM somatostatin suppressed insulin release promoted by glucose, although not asefficiently as 16 nM galanin. Hormone release in thepresence of somatostatin remained slightly above basal levels, despite the rise in [Ca2+]i(cf. Fig. 2, trace B).Supplementation of the medium with 25 mM Kf promoted a rapid and large release of insulin. With somatostatin being present throughout the perifusion (panel B, closed symbols), the addition of 25 mM K+ stimulated secretion, albeit not as much as in the absence of peptide (open symbols). Again, the secretory response was much improved when 20 mM glucose was supplemented to the medium. However, as was the case for galanin, somatostatin still suppressed secretion compared to the control under these conditions. Since the effects of galanin and somatostatin showed several similarities with those obtained with a2-adrenergic receptor activation (16), compounds interfering with the action of a-adrenergic agonists might also influence the response to the peptides. Previous studies have shown that pertussin toxin increases the levels of CAMPin the &cell and protectsinsulin secretion against the inhibitory action of epinephrine as well as somatostatin (37, 38). In the present study we used cell suspensions incubated with 50 ng/ml pertussis toxin for 8 h. In Fig. 7, traces A, C, and E , it is shown that treatment with pertussis toxin prevented the reduction in [Ca2+]ipromoted
1
5 0 uM 0-600
1 /J--" 20 mM Glucose
D
.
2 MIN
FIG. 4. Effects of 5 mM Ca2'on[Caa+]t in cells incubated both in the presenceof 20 mM glucose alone andin the presence of 20 mM glucose plus either galanin, somatostatin, or clonidine.
fluorescence, failed (data notshown). However, such afailure could be explained by the slow response time inherent with this method. In order to relate changes in [Ca2+]iand membrane potential to changes in insulin release, the latter parameter was investigated in a column perifusion system. As demonstrated previously (21) and confirmed in Fig. 5, panel A, 16 nM galanin rapidly inhibited glucose-stimulated insulin release. While the effect of galanin on [Ca"]i reversed slightly with time (cf. Fig. 2, trace A), insulin release remained blocked. However, subsequent addition of 25 mM K+ increased secretion, despite the presence of galanin. In panel B,open symbols, 25 mM K+ was added to a medium lacking glucose. This resulted in a peak of insulin release followed by a slow decline, and in this case
Inhibition of Insulin Release by Galanin Somatostatin and
I
I
I
20 mM Glucose 16 nM Galanin
A1 "
977
I
25mM K + 120 mM Glucose]
I
25mM K ' 116 mM Glucose
3000
2000
1000
100
0
,
1
I
20
60
40
I
t
I
20
40
,
1
20
I
1
40
TIME (MINI
FIG.5. Effects of galanin on insulin release from perifused islet cell aggregates. The horizontal bars indicate respective addition of test substances. In panel B a perifusion performed in the continuous presence of 16 nM galanin (04)is compared with a control performed in the absence of peptide (0-0).In panel C, the medium contained 4 mM glucose until the rise of the glucose concentration to 20 mM. A perifusion performed in the continuous presence of 16 nM galanin ( 0 - 0 ) is compared with a control performed in the absence of peptide (00).The traces shown in panels B and C are from paired experiments. The results are expressed as percentages of the average insulin release during the 5-min period preceding the introduction of test substances. contrast, galanin had no effect on pertussis toxin-treatedcells (open symbols),which responded almost like cells treated with the toxin and incubated in the absence of peptide (x). 2
2
3000
DISCUSSION
4
rn U
0
P
2000
W v)
4
w A w
=
1000
E
100
0 20
40
60
20
40
TIME WIN)
FIG. 6. Effects of somatostatin (Set) on insulin release from perifused islet cell aggregates. In panel E a perifusion performed in the continuous presence of 400 nM somatostatin (0-0)is compared with a control performed in the absence of peptide ( G O ) . The traces in panels A and B are from paired experiments. The results are expressed as percentages of the average insulin release during the 5min period preceding the introduction of test substances.
by either galanin, somatostatin, or clonidine. The cells remained responsive to diazoxide, a sulfonamide inhibiting insulin release (39,40). The effect of pertussis toxin treatment was not explained by the raised levels of CAMPin the8-cells, since 5 p~ of the adenylate cyclase activator forskolin, added 6-10 min prior to glucose stimulation, did not prevent the response to the agonists (truces B, D, and F). In Fig. 8, the effect of galanin on insulin release, stimulated by high K+ alone or together with 20 mM glucose, was investigated in @-cells pretreated with pertussis toxin. Insulin release from control cells (closed symbols) was markedly suppressed in the continuous presence of 16 nM galanin. In
Insulin release from pancreatic @-cellsis an example of a secretory process regulated by [Ca2+];(1-3). Glucose stimulation resultsin an increased [Ca2+];and therebyan activation of the secretory machinery. However, recentstudies have demonstrated that insulin release can also be stimulated simply by increasing the sensitivity of the secretory apparatus to the prevailing Ca2+concentration (41,42). Moreover, in permeabilized RINm5F cells the stable GTP analog GTPyS has been found to stimulate insulin release even at Ca2+ concentrations below lo-" M, suggesting a Caz+-independent pathway to be involved in the regulation of exocytosis (43). Insulin release can also be inhibited despite a high [Ca2+]i,a situation which is found in both intact and permeabilized pancreatic 8-cells subsequent to a-adrenoceptor activation (16, 44). Thus, available data indicate that neither [Ca2+Iias such nor the sensitivity of the secretory machinery to Ca2+is the ultimate regulator of the stimulus-secretion coupling in pancreatic ,&cells. In the present study it was found that both galanin and somatostatin suppressed insulin release by a mechanism involving a repolarization-induced reduction in [Ca2+]i.The fact that the slow increase in [Ca"];, observed after exposure to the peptides for a couple of minutes, was not paralleled by a stimulation of insulin release may indicate that both galanin and somatostatin also reduce the sensitivity of the secretory machinery to Ca2+.Moreover, galanin and somatostatin suppressed insulin release even when [Ca2+]iwas raised by stimulation with high concentrations of K+. Hence, the increased insulin release induced by the combination of high [Ca2+]iand glucose implies that Caz+and glucose act in concert to stim-
978
Inhibition of Insulin Release by Galanin and Somatostatin
[C a 2 j MI)
16 nM 4 0 0 UM Galanin Diaroxide
i
20 mM
:ttz 190
I
I
1 I _I/"--2
A
2 0 mM
Glucose
Glucose
170 -
4 25 90
-
400 nM
4 0 0 nM
FIG. 7. Effects ofgalanin, either somatostatin, or clonidine and diazoxide on [Cas'];. In the experiments represented by traces A, C, and E, the cell suspensions were preincubated with 50 ng/ml pertussis toxin for 8 h. The experiments represented by traces B, D, and F were performed in the presence of 5 p M forskolin.
150
\
4 0 0 JIM Diaroxlde
1
Glucose
230 150
-
2Ofie
20 mM Glucose
170
1 0 nM Clonldine
340
c
-
2o Glucose mM
-
\
j
Somatostatin
Somatostatin 20 mM Glucose
16 nM 400 uM Galanm Diazoxide
-
400 ,M
2o mM OiaTxide
-
E
D
160 -
420 260-
1 0 nM 4 0 0 JIM Cfonidine Dlaroxide
1
F
I
2 MIN
ulate exocytosis. It might be speculated that the role of the sugar is to increase the sensitivity of the secretory process to Caz+(41, 42,45). A possible explanation to such a sensitizing action of glucose might be activation of phospholipase C, with subsequent formation of diacylglycerol (6). The latter compound activates protein kinase C, resulting in an increased sensitivity of the secretory machinery to Ca2+(45). Phorbol esters, which directly activateprotein kinase C, stimulate insulin release in permeabilized islets as well as in RINm5F cells, in the latter case despite a reduction in [Ca2+];(31,44). However, the stimulatory effects of glucose on diacylglycerol formation is mainly secondary to increased Ca2+influx (6), and it is not likely that an increase in the glucose concentration from 0 to 4 mM activates phospholipase c to any major extent. Since studies on permeabilized cells have revealed the requirement for ATP in the secretory process (45, 46), the stimulatory effect of glucose on insulin release might also reflect increased production of this nucleotide (47). In this context it should be noted that clonidine has recently been reported to decrease glucose utilization by 40% in pancreatic islets (48). On the contrary, other investigators have failed to detect any effect of the agonist on oxidation of glucose (49). Although galanin and somatostatin might interfere with the glucose metabolism, such an effect is not likely to explain the pronounced inhibition of glucose-stimulated insulin release by the peptides. Since secretion was not completely restored in the presence of high [Ca2+]iand glucose, galanin and somatostatin also seem to interfere directly with the exocytotic machinery, by a mechanism not regulated by Ca" (43, 50). Although not studying temporal changes in [CaZ+];,it was recently found that somatostatin lowered [Ca"]; in islet cells stimulated with glucose but not with high K+ (12). These data taken together with previous data on galanin (21) as well as the present findings suggest thatthe decrease in [Ca2+]; mainly reflects a repolarization-induced closure of the voltage-activated Ca2+channels rather than a major direct inter-
ference of the peptides with these channels. Although we cannot exclude that the initial nadir in [Ca2+]iis due to a transient inhibition of the voltage-activated Ca" channels, the slow rise can be explained by some of these channels being re-opened since it was blocked by D-600. Worthy of note is that in pituitary cells exposed to the Ca2+channel blocker nifedipine, there was an initial decrease in [Ca2+Ii followed by a slow recovery (13). The fact that somatostatin evoked similar effects in these cells (13, 14) was interpreted as reflecting a direct interaction of the peptide with the voltage-activated Ca" channels, an effect observed also in other cells (35). Although there are some indications that somatostatin directly inhibits the voltage-activated Ca2+ channels, the initial nadir and the subsequent slow increase in [Ca2+];can be explained in terms of changes in membrane potential. If the peptide-induced repolarization is large enough most Ca2+channels will close. A later, although not complete, recovery of membrane potential wouldallow a limited number of Ca2+ channels to re-open and thereby induce a partial increase in [Ca2+];.Such changes in membrane potential has indeed been demonstrated in glucosestimulated 8-cells subsequent to a-adrenoceptor activation (51). In the presence of either galanin, somatostatin, or clonidine, addition of a high concentration of Ca2+evoked oscillatory changes in [Ca2+];.Similar oscillations in [Ca2+];have been observed recently in single ,&cells (52) and resemble those obtained in membrane potential subsequent to stimulation with intermediate glucose concentrations (51). This finding could actually be predicted since changes in membrane potential directly regulate the entrance of Caz+into the 8-cell (13). The reason for being able to detect these oscillations in [Ca2+];under the present experimental conditions is probably accounted for by a synchronization of the 8-cells. Although the exact mechanism is unknown, such a synchronization seemed to be initiated by the increase in [Ca2+];.It is noteworthy that the oscillations in [Ca2+]ilasted for a longer
Inhibition of Insulin Release by Galanin Somatostatin and
979
ity, an effect similar to thatobtained with diazoxide in normal @-cells(61). In addition, preliminary data applying the patchclamp technique indicate that galanin also activates the ATPregulated K+ channels in normal P-cells (62). One should be aware of the fact that theeffects of galanin and somatostatin on both membrane potential and [Ca2+]iare similar to those evoked by a-adrenoceptor activation (161, but so far it has 6000 not been possible to reveal any consistent effect of such an activation on the actual K+ channels, either with the patchclamp technique' or with %Rb+ efflux studies (63). Hence, more thorough electrophysiological investigations are needed before a possible interaction of galanin and somatostatinwith the ATP-regulated K+ channels can be finally established. Since peptides exert their effects by activating cell surface receptors, the signal initiated by receptor binding has to be transferred to theeffector site by a mechanism in or beneath the plasma membrane. In many cells, GTP-binding proteins (G-proteins) are part of such a signal-transduction pathway (64). Pertussis toxin is known to modify G-proteins by the addition of an ADP-ribose group to the a-subunit of the protein (64). In the pancreatic @-cells,treatment with the toxin blocked the reduction in [Ca'+]i and the inhibition of insulin release promoted by either galanin, somatostatin (37), or a-adrenergic agonists (37, 38). It is therefore likely that these effects are mediated by a G-protein. So far the identity of this putative G-protein is unknown, but since the agonists induced a lowering in [Ca2+Iidespite the presence of forskolin, it is not likely that pertussis toxin exerts its effects solely by 100 increasing the cellular content of CAMP. A pertussis toxin0 20 40 sensitive G-protein purified from red blood cells has been reported to directly activate K+ channels in atrial muscle cells (65). It is possible that the repolarizing effect of galanin, TIME (MIN) somatostatin, and clonidine is mediated by a similar mechaFIG.8. Effects of galanin on insulin release from control nism. A direct effect of the toxin on the ATP-regulated K+ cells (0-0) or cells pretreated with 60 ng/ml pertussis toxin for 8 h (0-0).Cells pretreated with pertussis toxin but perifused in channels was excluded, since addition of diazoxide still prothe absence of galanin (X-X). The traces arefrom paired experiments. moted a repolarization-induced reduction in [Ca2+Ii.Recent The results areexpressed as percentages of the average insulin release studies on RINm5F tumor cells failed to demonstrate any during the 5-min period preceding the introduction of high K+. effect of pretreatment with pertussis toxin on the ability of galanin to activate ATP-regulated K+ channels (60). This period of time in the presence of either galanin, somatostatin, might reflect an intrinsic difference in the K+ channel protein or clonidine than would beexpected from membrane potential and/orits regulation between normal @-cellsand insulinmeasurements at 11 mM glucose. Interestingly, a similar pro- producing tumor cells. Nevertheless, in normal @-cellsit seems longation of the time between bursts has been observed when that the suppressive effects of galanin and somatostatin on measuring changes in membrane potential subsequent to the insulin release are mediated by a G-protein. Since the repoaddition of epinephrine to glucose-stimulated @-cells(51). larization-induced decrease in [Ca2+]iwas not solely responThe resting membrane potential of the pancreatic @-cell is sible for the inhibitory effects on insulin release, it is likely determined mainly by the membrane permeability to K+ (53). that there exist several agonist-associated G-proteins with This permeability is regulated by K+ channels sensitive to sensitivity to the toxin. Alternatively, the putative G-protein ATP, and an increase in intracellular ATP and/or the ATP/ may have more than one effector system (65). ADP ratio will close these channels (54-57). Because of the presence of other ion permeabilities, the reduction in K+ Acknowledgment-We are indebted to Dr. Rolf Sammann, Swedish conductance will result in depolarization of the @-cell(58).It Hoechst, for the generous gift of '2SI-labeledinsulin. is of interest to note that theATP-regulated K' channels are REFERENCES directly opened by diazoxide, resulting in repolarization of the @-cell(40, 59). More than 99% of these channels have to be 1. Wollheim, C. B., and Sharp, G. W.G. (1981) Physiol. Reu. 61, 914-973 closed in order to obtain the necessary depolarization to open 2. Hellman, B., and Gylfe, E. (1986) in Calcium and Cell Function the voltage-activated Ca2+channels (40, 58). Thus, repolari(Cheung, W. Y., ed) Vol. 6, pp. 253-326, Academic Press, New zation can be induced by a relatively small increase in the York activity of the K+ channels, making them suitable targets for 3. Prentki, M., and Matschinsky, F. M. (1987) Physiol. Reu. 6 7 , galanin and somatostatin. Since the effects of both peptides 1185-1248 4. Pipeleers, D. G . (1987) Diabetologia 30, 277-291 have a certain resemblance to those obtained with diazoxide, 5. AhrBn, B., Taborsky, G. J., Jr., and Porte, D., Jr. (1986) Diabeit is tempting to speculate that galanin andsomatostatin tologia 29, 827-836 interact with the ATP-regulated K+ channels. Indeed, data 6. Biden, T. J., Peter-Riesch, B., Schlegel, W., and Wollheim, C. B. obtained with the patch-clamp technique support the notion (1987) J.Biol. Chem. 262,3567-3571 that galanin activates the actual K+ channels in RINm5F cells (60). When using 86Rb+asa marker for K+, it was *P. Rorsman, T. Nilsson, P. Arkhammar, and P.-0. Berggren, demonstrated that galanin stimulated the efflux of radioactiv- unpublished observations. 2 5 mM K +
20 mM Glucose
980
Inhibition of Insulin Release by Galanin Somatostatin and
7. Gagerman, E., Idahl, L.-A., Meissner, H. P., and Taljedal, I.-B. Acad. Sci. U. S. A. 83,9832-9836 36. Dunlap, K., and Fischbach, D. G. (1981) J . Physiol. (Lond.) 317, (1978) Am. J. Physiol. 235, E493-E500 8. Efendic, S., Grill, V., and Luft, R. (1975) FEBS Lett. 5 5 , 131519-535 133 37. Katada, T., and Ui, M. (1979) J. Biol. Chem. 254,469-479 9. Heindel, J. J., Williams, E., Robison, G.A., and Strada, S. J. 38. Wollheim, C.B., and Biden, T. J. (1986) Ann. N. Y. Acad. Sci. (1978) J. Cyclic Nucleotide Res. 4, 453-462 488,317-333 10. Schlegel, W., Wuarin, F., Wollheim, C. B., and Zahnd, G. R.39. Henquin, J. C., Charles, S., Nenquin, M., Mathot, F.,and Tam(1984) Calcium Cell 4, 223-236 agawa, T. (1982) Diabetes 31, 776-783 11. Pace, C. S., Murphy, M., Conant, S., and Lacy, P. E. (1977) Am.40. Arkhammar, P., Nilsson, T., Rorsman, P., and Berggren, P.-0. J. Physiol. 233, C164-Cl71 (1987) J. Biol. Chem. 2 6 2 , 5448-5454 12. Sussman, K. E., Leitner, J. W., and Draznin, B. (1987) Diabetes 41. Wollheim, C . B., Ullrich, S., and Pozzan, T. (1984) FEBS Lett. 36,571-577 177,17-22 13. Luini, A., Lewis, D., Guild, S., Schofield, G., and Weight, F.42. Rorsman, P., and Abrahamsson, H. (1985) Acta Physiol. Scand. (1986) J . Neurosci. 6 , 3128-3132 125,639-647 Vallar, L.,Biden, T. J., and Wollheim, C. B. (1987)J. Biol. Chem. 14. Lewis, D.L., Weight, F. F., and Luini, A. (1986) Proc. Natl. Acud. 43. Sci. U. S. A. 262,5049-5056 83,9035-9039 44. Tamagawa, T., Niki, I., Niki, H., and Niki, A. (1985) Biomed. 15. Exton, J. H. (1980) Am. J. Physiol. 238, E3-El2 16. Nilsson, T., Arkhammar, P., Rorsman, P., and Berggren, P.-0. Res. 6,429-432 (1988) J. Biol. Chem. 2 6 3 , 1855-1860 45. Knight, D. E. (1987) Biosci. Rep. 7, 355-367 17. Rorsman, P., Arkhammar, P., Berggren, P.-O., and Nilsson, T. 46. Yaseen, M. A., Pedley, K. C., and Howell, S. L. (1982) Biochem. (1987) Diabetologia 30,575 (abstr.) J. 206,81-87 18. Ullrich, S., and Wollheim, C. B. (1984)J . Bioi. Chem. 259,411147. Ashcroft, S. J. H., Weerasinghe, L. C. C., and Randle, P. J. (1973) J. 132,223-231 4115 19. Malaisse, W. J., and Malaise-Lagae, F. (1984) Experientia 48. Laychock, S. G., and Bilgin, S. (1987) FEBS Lett. 2 1 8 , 7-10 (Basel) 40,1068-1075 49. Leclercq-Meyer, V., Herchuelz, A., Valverde, I., Couturier, E., 20. Dunning, B. E.,Ah&,B., Veith, R. C., Bottcher, G., Sundler, Marchand, J., and Malaisse, W. J. (1980) Diabetes 2 9 , 193200 F., and Taborsky, G. J., Jr. (1986) Am. J. Physiol. 251, E127E133 50. Barrowman, M.M., Cockcroft, S., and Gomperts, B. D. (1986) 21. Ahrkn, B., Arkhammar, P., Berggren, P.-O., and Nilsson, T. Nature 319,504-507 (1986) Bjochem.Biophys.Res.Commun. 140, 1059-106351.Cook,D. L.,andPerara, E. (1982) Diabetes 31,985-990 52. Grapengiesser, E., Gylfe,E., and Hellman, B. (1988) Biochem. 22. Rokaeus, A. (1987) Trends Neurosci. 10, 158-164. Biophys. Res. Commun. 151, 1299-1304 23. Fisone, G., Wu, C.F., Consolo, S., Nordstrom, O., Brynne, N., Bartfai, T., Melander, T., and Hokfelt, T. (1987) Proc. Natl. 53. Meissner, H. P., Henquin, J. C., and Preissler, M. (1978) FEBS Lett. 94,87-89 Acad. Sci. U. S. A. 84, 7339-7343 54. Cook, D. L., and Hales, C. N. (1984) Nature 3 1 1 , 271-273 24. Hellman, B. (1965) Ann. N.Y. Acad. Sci. 131,541-558 25. Nilsson, T.,Arkhammar, P., Hallberg, A., Hellman, B., and 55. Ashcroft, F. M., Harrison, D. E., and Ashcroft, S. J. H. (1984) Berggren, P.-0. (1987) Biochem. J . 248,329-336 Nature 312,446-448 56. Rorsman, P., and Trube, G. (1985) Pfl@ers Arch. 405,305-309 26. Lernmark, A. (1974) Diabetologiu 10, 431-438 27. Arkhammar, P., Berggren, P.-O., and Rorsman, P. (1986) Biosci. 57. Kakei, M., Kelly, R. P., Ashcroft, S. J. H., and Ashcroft, F. M. Rep. 6,355-361 (1986) FEBS Lett. 208,63-66 58. Cook, D.L., Satin, L. S., Ashford, M.L. J., and Hales, C.N. 28. Hellman, B. (1975) Endocrinology 97,392-398 (1988) Diabetes 37, 495-498 29. Beaven, M. A,, Rogers, J., Moore, J. P., Hesketh, T. R., Smith, 59. Trube, G., Rorsman, P., and Ohno-Shosaku, T. (1986) Pflugers G. A., and Metcalfe, J. C. (1984) J. Biol. Chem. 2 5 9 , 71297136 Arch. 407,493-499 30. Rink, T. J., Montecucco, C., Hesketh, T. R., and Tsien, R. Y. 60. de Weille, J., Schmid-Antomarchi, H., Fosset,M., Lazdunski, M. (1988) Proc. Natl. Acud. Sci. U. S. A. 85, 1312-1316 (1980) Biochim. Biophys. Acta 595, 15-30 31. Arkhammar, P., Nilsson, T., and Berggren, P.-0. (1986) Biochim. 61. Henquin, J. C., and Meissner, H. P. (1982) Biochem. Phurmacol. Biophys. Acta 887,236-241 31,1407-1415 32. Kanatsuna, T., Lernmark, A., Rubenstein, A. H., and Steiner, D. 62. Ahrkn, B., Rorsman, P., and Berggren, P.-0. (1988) FEBS Lett. F. (1981) Diabetes 30, 231-234 229,233-237 33. Tsien, R. Y., Pozzan, T., and Rink, T. J. (1982) J. Cell Biol. 9 4 , 63. Tamagawa, T., and Henquin, J. C. (1983) Am. J. Physiol. 2 4 4 , 325-334 E245-E252 34. Arkhammar,P., Nilsson, T., and Berggren, P.O. (1988) Cell 64. Neer, E. J., and Clapham, D. E. (1988) Nature 333,129-134 Calcium, in press 65. Brown, A. M., and Birnbaumer, L. (1988) Am. J . Physiol. 2 6 4 , 35. Tsunoo, A,, Yoshii, M., and Narahashi, T. (1986) Proc. Natl. H401-H410
