Mechanism of sympathetic neural regulation of insulin, somatostatin, and glucagon secretion TAKESHI KUROSE, YUTAKA SEINO, SHIGEO NISHI, KAZUO TSUJI, TOMOHIKO TAMINATO, KINSUKE TSUDA, AND HIROO IMURA Division of Metabolism and Clinical Nutrition, and Second Division, Department of Medicine, Kyoto University School of Medicine, Kyoto 606, Japan

KUROSE, TAKESHI, YUTAKA SEINO, SHIGEO NISHI, KAZUO TSUJI, TOMOHIKO TAMINATO, KINSUKE TSUDA, AND HIROO IMURA. Mechanism of sympathetic neural regulation of insulin, somatostatin, and glucagon secretion. Am. J. Physiol. 258 (En-

docrinol. Metab. 21): E220-E227,1990.-The effects of electrical stimulation of the left splanchnic nerve on insulin, somatostatin, and glucagon secretion from the isolated perfused rat pancreas were investigated. Electrical splanchnic nerve stimulation (SNS), performed by square-wave impulses, produced a 25% decrease in effluent flow and a IO-fold increase in perfusate norepinephrine. Both insulin and somatostatin output in the presence of 16.7 mM glucose were inhibited during SNS by 85 and 56% of the basal value, respectively. Glucagon output in the presence of 5.5 mM glucose was stimulated 20fold by SNS. Perfusion with 10B6 M propranolol further decreased insulin and somatostatin output during SNS, when expressed as the total decrement beneath basal during stimulation. The glucagon response to SNS tended to be enhanced, although not significantly, by simultaneous infusion of low6 M propranolol. However, 10m6M phentolamine (Phe) attenuated the SNS-induced inhibition of insulin and somatostatin output by 50 and 40%, respectively. However, insulin output remained decreased after SNS with Phe. The SNS-induced glucagon response was completely abolished by 10m6M Phe alone or by low6 M Phe plus 10m6M propranolol. With lo-” M Phe plus low6 M propranolol, insulin and somatostatin output remained decreased after SNS. These results suggest that insulin and somatostatin secretions induced by glucose are inhibited during SNS through the cu-adrenergic mechanism and also that the ,& adrenergic mechanism exerts a stimulatory action. SNS-induced glucagon secretion occurs mainly through a-adrenergic activation. The involvement of a nonadrenergic mechanism in the inhibition of glucose-stimulated insulin and somatostatin secretion induced by SNS was suggested by the incomplete reversal of insulin and somatostatin inhibition in the presence of Phe. splanchnic propranolol

nerve; autonomic

nervous system; phentolamine;

NERVOUS SYSTEM is known to play an important role in the rapid adjustment of various metabolic pathways to physiological and pathophysiological demands (8, 37). This function is accomplished partly within the endocrine pancreas, where a close relationship between sympathetic nerve fibers and endocrine cells has been well documented (40). A large number of studies have shown that insulin, somatostatin, and glucagon release are modified substantially in response THE SYMPATHETIC

E220

0193-1849/90

$1.50 Copyright

to stimulation of the splanchnic nerve (1, 26). Because most of these studies were performed in vivo, extrapancreatic events such as metabolic changes in various tissues and changes in the secretion of other hormones could have been induced (17, 30), making it difficult to evaluate the direct neural effect on the endocrine pancreas. In recent years, it has been demonstrated that intrapancreatic nerves contain peptides other than classic substances, which may function also as neurotransmitters or neuromodulators (1, 26). In particular, neuropeptide Y (28,31) and galanin (11) are candidates for peptide modulators. If these peptides mediate the response of pancreatic hormones to splanchnic nerve stimulation (SNS), combined adrenergic blockade must be incomplete or ineffective (12). In the present study, we have investigated the direct effect of electrical stimulation of the sympathetic nerve on the endocrine pancreas using the isolated and vascularly perfused rat pancreas with intact left splanchnic nerve innervation. We have also studied the effect of adrenergic blockade on SNS-induced hormone release to verify the possible involvement of a nonadrenergic neurotransmitter in the splanchnic nerve. MATERIALS

AND METHODS

Animals and pancreas perfusion (Fig. 1). Male Wistar rats weighing 200-250 g were housed for at least 1 wk before the experiments in a temperature- and lightconditioned room. After an overnight fast, anesthesia was introduced by intraperitoneal administration of 6 mg/lOO g body wt pentobarbital sodium. The pancreas was isolated as previously described (38), with some modifications. In brief, laparotomy was performed, and the omentum, which connects the pancreas to the transverse colon, was dissected. After the ligation of anterior and posterior mesenteric arteries and veins, the entire intestine below the duodenum was cut and removed from the rat to improve the exposure. The splenic arteries near the splenic hilus, the right and left gastric arteries, and the gastric branches of the gastroepiploic artery were ligated individually. The left splanchnic nerve was cautiously exposed, isolated, and cut beneath the diaphragm. The abdominal aorta was carefully separated from the connective tissues and ligated well above the point of origin of the celiac axis and beneath the renal arteries.

0 1990 the American

Physiological

Society

Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (133.006.082.173) on August 2, 2018. Copyright © 1990 American Physiological Society. All rights reserved.

SYMPATHETIC

Celiac artery

NEURAL

REGULATION

‘ILL& ,-a-

II

OF INSULIN,

Perfusate Infusion

h

pump

Pressure monitor

FIG. 1. Schematic representation of isolated perfused preparation with intact left splanchnic nerve.

rat pancreas

Then, an arterial cannula was inserted into the aorta through a slit made just above the lower ligation of the aorta, and the cannulation tip was placed into the celiac artery. At that moment, circulation was initiated, and then the pancreas, spleen, stomach, duodenum, liver, lower part of the esophagus, and the left splanchnic nerve were removed in one block from the rat. Much care was taken not to damage the celiac and mesenteric plexuses in this procedure. The portal vein was then cannulated, and the block was placed into a plastic chamber. The temperature was maintained at 37OC, and the humidity was kept constant in the chamber. A bipolar platinum electrode was attached to the left splanchnic nerve just beneath the origin of adrenal branch. All perfusions were accomplished with Krebs-Ringer bicarbonate buffer containing 0.25% bovine serum albumin (fraction V, Sigma Chemical, St. Louis, MO) and 4.6% dextran (mean mol wt 70,000, Pharmacia, Sweden). The perfusion medium was gassed with 95% Oz-5% CO2 and maintained at pH 7.4 at 37OC. The arterial pressure of O2 ranged from 450 to 500 mmHg. The flow rate through the pancreas was usually 1.8-1.9 ml/min at perfusion pressures of 60-100 mmHg, measured with the transducer (model T12, Gould) and monitor. Each 1-min effluent from the portal vein was collected into chilled tubes containing 1,000 U of Trasylol (Bayer, Leverkusen, FRG), frozen immediately, and stored at -20°C until assayed. Experimental protocol. After an equilibration period of 20 min, samples were taken for 20 min. Three minutes after the start of sampling, electrical SNS was performed for 5 min as 1-ms, 30-V square-wave impulses at a frequency of 10 Hz. The completeness of the electrical SNS was confirmed by ascertaining an increase of ~30 mmHg in the pressure of the arterial cannulation tube. To investigate the response of insulin and somatostatin to SNS, the glucose concentration was kept constant at 16.7 mM. On the other hand, the response of glucagon to SNS was investigated with a constant glucose concentration of 5.5 mM. To study the effect of adrenergic antagonists on hormone release induced by SNS, a final concentration of 10e6 M propranolol hydrochloride (Inderal, ICI Pharmacia), 10v6 M phentolamine mesylate (Regitin, Ciba-Geigy Japan), or 10v6 M propranolol plus 10e6 M phentolamine was infused over whole experimen-

SOMATOSTATIN,

AND

GLUCAGON

E221

tal periods, including the equilibration period. Assays and data analyses. The concentration of immunoreactive insulin was measured by radioimmunoassay using a polyethylene glycol precipitation technique (9), with rat insulin (Novo, Bagsvaerd, Denmark) as standard. Immunoreactive somatostatin was measured by specific radioimmunoassay (38), with a modification of the method described by Arimura et al. (4), with an antiserum T-316 and synthetic cyclic somatostatin-14 as standard. These antisera did not cross-react with insulin, glucagon, gastrin, motilin, vasoactive intestinal peptide, secretin, or substance P. Immunoreactive glucagon was measured by radioimmunoassay using the talcum adsorption technique of Sakurai et al. (35), with an antiserum specific for pancreatic glucagon (OAL-123, kindly provided by Ohtsuka Assay Laboratory, Tokushima, Japan) in a final dilution of 1:120,000 (18). Porcine glucagon (Novo) was the standard. The norepinephrine concentration was determined by high-performance liquid chromatography combined with the trihydroxyindole method (14). The insulin, somatostatin, and glucagon outputs were calculated as the concentration of each hormone multiplied by the volume of each l-min effluent. Percent basal was calculated as follows: the value of hormone output at each time was divided by the basal value at 2-3 min (hereafter expressed as 3 min) and then multiplied by 100. For the evaluation of the decrease in perfusate flow and insulin and somatostatin output during SNS, the percent basal at nadir was calculated with the mean of the values at the lowest point in each experiment. The total decrement beneath the basal level of insulin and somatostatin output and the total increment above the basal level of glucagon output are the sums of the difference between each value during the 5-min period of SNS and the basal value at 3 min, expressing as negative values for decrease and positive values for increase. Statistical comparisons of means within a group were made using Student’s paired t test. Comparisons between means of different groups were made by analysis of variance followed by Duncan’s multiple range test (10) if null hypotheses were rejected by the former. Differences were accepted as significant at P < 0.05. RESULTS

Effects of SNS on effluent perfusate flow in the presence or absence of adrenoceptor antagonists. As shown in Fig.

2, SNS for 5 min decreased the effluent flow from 1.90 t 0.03 to 1.46 t 0.15 ml/min (P < 0.05), the maximal decrease being 25.5 t 8.9% of the basal value (Table 1). In the presence of low6 M propranolol, SNS caused a greater decrease in the effluent flow, from 1.89 t 0.03 to 1.24 t 0.18 ml/ min (P < 0.05). On the other hand, SNS induced a less remarkable decrease in the effluent flow, from 1.82 t 0.04 to 1.64 t 0.04 ml/min (P < O.Ol), in the presence of lOa M phentolamine. Furthermore, the effluent flow was still reduced by SNS during combined perfusion with 10m6M propranolol and 10m6M phentolamine, from 1.92 t 0.04 to 1.66 t 0.10 ml/min (P C 0.05). Among these four groups there were no significant differences in the degree of effluent flow reduction induced by SNS (Table 1).

Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (133.006.082.173) on August 2, 2018. Copyright © 1990 American Physiological Society. All rights reserved.

E222

SYMPATHETIC

NEURAL

REGULATION

OF INSULIN,

SOMATOSTATIN,

AND

GLUCAGON

1. Percent basal at nadir during SNS

TABLE

Effluent Volume,

Insulin, %

Somatostatin, %

14.8-+4.8* 66.2+2.9*$ 33.6*6.5* 70.0&6.9t$

43.3zk9.6t 82.5+8.4$ 13.7t9.4" 68.8k7.07

YO

o

J

,

j .::::

1

.i:.:.;.:...:.;.:

..:...;I

,

5

-

STIM.

o--e

STIM.+PHENTOLAMINE

b--d

STIM.+PROPRANOLOL

o-*-a

STIM.+PHENTOLAMINE + PROPRANOLOL

ALONE

I

10 TIME

Stim. alone Stim. + phentolamine Stim. + propranolol Stim. + phentolamine propranolol

Means t SE of 5 individual experiments are shown. * P < 0.01, t P < 0.05 vs. effluent volume; $ P < 0.01, QP < 0.05 vs. stimulation (stim.) alone. TABLE 2. Norepinephrine before and during SNS

4

15

+

20 (mid

-

STIM.

-

STIM. +PHENTOLAMINE

Pdml-

nificant differences in the basal insulin output among the four groups: stimulation alone, stimulation plus phentolamine, stimulation plus propranolol, and stimulation plus phentolamine and propranolol (Fig. 3). In the presence of 16.7 mM glucose, the insulin concentration

ALONE

B

-

STIM.

6--d

STIM. + PROPRANOLOL

ALONE

C I

1

. . . . . . ..I

5

I

I

I

10

15

20

TIME

(min

. ...::.:. .

5

Stimulation

Stim. alone 36-+10 326&107 Stim. + phentolamine 16&4 1,706f654 1,570f835 Stim. + propranolol 2020 2,107f1,085 Stim. + phentolamine + 23k3 propranolol Values are means + SE, Norepinephrine concentration measured in

Effects of SNS on norepinephrine concentration of the perfusate (Table 2). SNS produced a lo-fold increase in norepinephrine concentration, from the prestimulatory level of 36 k 10 to 326 t 107 pg/ml during the stimulation (P < 0.05). Phentolamine (10e6 M) and/or propranolol (10e6 M) enhanced the increase of norepinephrine concentration during SNS. Effects of SNS on insulin secretion in the presence or absence of adrenoceptor antagonists. There were no sig-

b

concentration Prestimulation

2. Effect of electrical splanchnic nerve stimulation on effluent volume in presence or absence of 10m6 M phentolamine and/or 10e6 M propranolol. Means of 5 individual experiments ‘are shown. FIG.

1

74.5t8.9 89.8k1.3 66.4k8.4 86.5k3.8

.

10 TIME

I

4

15

20 (min

STIM. ~.*.*.‘.*.*.*.~.*.* . . . .. . .. . . Fl‘.*.~.*.~.~.*.~.~.~.

TIME

-

STIM.

-

STIM. +PHENTOLAMINE +PROPRANOL~

ALONE

(mid

FIG. 3. Effect of electrical splanchnic nerve stimulation on insulin concentration (top) and insulin output (bottom) from isolated perfused rat pancreas in presence or absence of low6 M phentolamine (A), in presence or absence of low6 M propranolol (B), and in presence or absence of 10m6 M phentolamine plus lo-' M propranolol (C). Glucose concentration in perfusate was kept constant at 16.7 mM throughout experimental period. Means & SE of 5 individual experiments are shown.

Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (133.006.082.173) on August 2, 2018. Copyright © 1990 American Physiological Society. All rights reserved.

SYMPATHETIC

NEURAL

REGULATION

OF INSULIN,

was decreased by SNS from the basal concentration

of

2.41 t 0.31 n g/ m 1 t o a nadir of 0.85 & 0.13 rig/ml at 5

min (P c 0.01) (Fig. 3, stimulation alone). The insulin output during each l-min collection period also markedly decreased, as shown in Figs. 3 and 4. In the presence of 10m6M phentolamine, the basal insulin output increased slightly but significantly (8.10 t 1.30 ng/min vs. control value 4.98 t 0.64 ng/min), as shown in Fig. 3A. The phentolamine perfusion partially reversed the decrease of insulin concentration and output during SNS and the maximal decrease, 33.8 t 2.9%, was significantly smaller than in the group given stimulation alone (Table 1). When compared by total decrement during SNS, the decrease in the group given stimulation plus phentolamine was still smaller, although not significantly, than in the group given stimulation alone (Table 3). On theother hand, the perfusion with 10m6M propran0101did not significantly alter the extent of inhibition of insulin output induced by SNS (Figs. 3B and 4, Table 1). However, the total decrement of insulin output during the 5-min stimulation period was greater in the propran0101 group than in the group given stimulation alone (Table 3), suggesting augmentation of the inhibitory action of the nerve stimulation. Insulin output during combined administration of low6 M phentolamine and 10m6M propranolol still decreased from 5.95 t 1.12 to 4.33 t 0.65 ng/min in response to SNS (P < 0.05) (Fig. 3C). The total decrement of insulin output during the SNS period also was smaller in the

,

1

-

STIM.

ALW

-

WM.

+PHENTOLAMINE

. . .. . .. . .. . . .. . . ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~ ~.~.~.~.~.~,~.~.~.~.~.~.~.~.~.~.~ ~.~.~.~.~.~,~.~.~.~.~,~.~.~.~.~.~ . . . .* . .... *..... ...... ...... ...... ...... .,.... ::::*:* I,. . :: ~~~~‘.‘.““‘.“‘.“‘.‘.‘.‘.“‘.’ ~.~.~.~.~.~.~.~.~.~.~.~.~.‘.~ ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~ ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~ ~.~,~.~,~.~.~,~.~,~.~.~.~.~.~.~.~ ~,~.~.~,~.~.~.~.~,~.‘.~.~.~.~.~,f ~.‘.~.~.~.~.‘.~.~.~.‘.~.~.‘.~.~.~ ~.~.~.~.~.~,~.~.~.~.‘,~.~.~.~.~.~ ~.~.~.~.~.~.~.~.~.~.~.~,~.~.~.~.~ ~~~~‘~~~‘~~~‘~~~~~~~~~~~~~~~~~‘~~ ~.~.~.‘.‘.~.~.‘.~.‘.‘,~.~.~.~.~.~ ~,~.~.~.~.~.~.~.~.~,~.~.~.~.~.~.~ ~.~.~.~,~.~.~.~.~.~.~,~.~.~.~.~.~ ~.~.~.~.~.~.~.~.~.~.‘.~.~.‘.‘.‘.’

5

10 TIME

15

20 (mid

FIG. 4. Percent basal insulin output from isolated perfused rat pancreas before, during, and after electrical stimulation of splanchnic nerve with or without 10B6 M phentolamine and/or 10m6M propranolol. Means of 5 individual experiments are shown.

SOMATOSTATIN,

AND

GLUCAGON

E223

3. Total decrements beneath basal level of insulin and somatostatin output and total increment above basal level of glucagon output during SNS

TABLE

Z Aid/S min, ng/5 min

Stim. alone Stim. + phentolamine Stim. + propranolol Stim. + phentolamine propranolol

+

-14.4k1.2 -7.8t2.7 -24.5+4.5avd -5.lk2.0""

2 Airs/5 min, pg/5 min

2 AirG/S min, pg/5 min

-37.3k12.3 2,557+841 -0.5k20.4 139266" - 139.62 18.0bld 4,539-c 1,244d -6.3+12.3f 58+31""

Values are means t SE. Total decrements beneath basal level of insulin and somatostatin output during splanchnic nerve stimulation are expressed as ZAirI/5 min and ZAirS/5 min, and total increment above basal level of glucagon output during nerve stimulation is expressed as ZAirG/5 min. a P C 0.05, b P < 0.01 vs. stim. alone; ’ P C 0.05, d P < 0.01 vs. stim. + phentolamine; e P < 0.05, f P < 0.01 vs. stim. + propranolol.

group given stimulation plus phentolamine and propran0101than in the group given stimulation alone (P < 0.05) or in the stimulation plus propranolol group (P c 0.01). Effects of SNS on somatostatin secretion in the presence or absence of adrenoceptor antagonists. In the presence of 16.7 mM glucose, SNS reduced both somatostatin concentration (13.8 t 2.0 to 8.8 $- 3.8 pg/ml) and somatostatin output (27.4 t 4.1 to 15.1 t 3.3 pg/min, P < 0.02) (Figs. 5 and 6, Table 1). The perfusion with phentolamine significantly elevated basal somatostatin output to 62.3 t 10.3 pg/min (P < 0.01 vs. group given stimulation alone) and abolished the changes in the somatostatin concentration and output during SNS (Fig. 5A). On the contrary, the SNS-induced inhibition of somatostatin secretion during the perfusion with 10m6M propranolol appeared to be larger than in the group given stimulation alone (Figs. 5B and 6). In addition, the total somatostatin output during the stimulation period was significantly lower in the presence of propranolol than in the control group (P