Assessment of the Role of Interstitial Glucagon in the
Acute Glucose Secretory Responsiveness of In Situ
Pancreatic ␤-Cells
Karen Moens,1 Veerle Berger,1 Jung-Mo Ahn,2 Chris Van Schravendijk,1 Victor J. Hruby,2
Daniel Pipeleers,1 and Frans Schuit1
Glucagon is a potent stimulator of insulin release in the
presence of a permissive glucose concentration, activating
␤-cells in vitro via both glucagon- and glucagon-like peptide-1 (GLP-1)–receptors. It is still unclear whether
locally released glucagon amplifies the secretory responsiveness of neighboring ␤-cells in the intact pancreas. The
present study investigates this question in the perfused
pancreas by examining the effects of antagonists for glucagon receptors ([des-His1,des-Phe6,Glu9]glucagon-NH2,
10 ␮mol/l) and GLP-1–receptors [exendin-(9-39)-NH2, 1
␮mol/l] on the insulin secretory response to glucose. The
specificity of both antagonists was demonstrated by their
selective interaction with glucagon-receptor signaling in
rat hepatocytes and GLP-1–receptor signaling in Chinese
hamster lung (CHL) fibroblasts. In purified rat ␤-cells, the
glucagon-receptor antagonist (10 ␮mol/l) inhibited the
effect of 1 nmol/l glucagon upon glucose-induced insulin
release by 78 ⴞ 6%. In the perfused rat pancreas, neither
of these antagonists inhibited the potent secretory response to 20 mmol/l glucose, although they effectively
suppressed the potentiating effect of, respectively, an
infusion of glucagon (1 nmol/l) or GLP-1 (1 nmol/l) on
insulin release. When endogenous glucagon release was
enhanced by isoproterenol (100 nmol/l), no amplification
was seen in the simultaneous or subsequent insulin secretory response to glucose. It is concluded that, at least
under the present selected conditions, the glucose-induced insulin release by the perfused rat pancreas seems
to occur independent of an amplifying glucagon signal
from neighboring ␣-cells. Diabetes 51:669 – 675, 2002
W
e have previously reported that the potent
insulin secretory response to glucose is
markedly diminished when pancreatic ␤-cells
are isolated from the islet structure. A number of previous in vitro observations using isolated islets
or purified islet cells have supported the idea that the
amplitude of glucose stimulation of insulin release is
strongly influenced by the intracellular cAMP level, which
From the 1Diabetes Research Center, Vrije Universiteit Brussel, Brussels,
Belgium; and the 2Department of Chemistry, University of Arizona, Tucson,
Arizona.
Address correspondence and reprint requests to Dr. Frans Schuit, Molecular Pharmacology Unit, Diabetes Research Center, Faculty of Medicine, Vrije
Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium. E-mail:
fschuit@minf.vub.ac.be.
Received for publication 10 July 2001 and accepted in revised form 26
November 2001.
AUC, area under the curve; CHL, Chinese hamster lung; GLP-1, glucagonlike peptide-1; IBMX, 3-isobutyl-1-methylxanthine; RIA, radioimmunoassay.
DIABETES, VOL. 51, MARCH 2002
is regulated by various hormones and neurotransmitters
(1–5). It remained unclear, however, which agents provide
the amplifying cAMP signals in physiological conditions
(6). There is sufficient evidence for a major role of
glucagon in isolated islet preparations (1–3,7), which can
potentiate glucose-induced insulin release via two different receptors (8).
In vitro studies have indicated that separation of ␤-cells
from ␣-cells explains in part the lower secretory activity of
purified ␤-cells as compared with isolated islets (1,2).
Locally released glucagon was indeed found to amplify
glucose-induced insulin release in isolated islet cell preparations (1,7,8). This also explains the higher secretory
responsiveness of ␣-cell– containing islets isolated from
the dorsal pancreatic lobe when compared with that of
␣-cell poor islets from the ventral lobe (5,9). It is conceivable that circulating glucagon levels influence the cAMP
levels in ␤-cells in vivo. Other peptides, in particular
glucagon-like peptide-1 (GLP-1) and glucose-dependent
insulinotropic polypeptide, are likely candidates to contribute in this regulatory pathway through variations in
their circulatory concentrations (3,10). In addition to this
endocrine pathway, peptides such as glucagon may act via
interstitial interactions. It is not yet possible to investigate
such influences in the endocrine pancreas in vivo. Experiments have been undertaken in the isolated perfused
pancreas to examine the possible role of ␣-cells on neighboring ␤-cells. These studies have indicated that islet
blood flows in the direction from the ␤-cell core toward
the non–␤-cell mantle (11,12). However, none of these
data could provide final conclusions concerning interstitial
glucagon interacting at sites where both ␣- and ␤-cells are
in close contact.
As was suggested 10 years ago, specific glucagon-receptor antagonists would be required to assess the contribution of glucagon in the ␤-cell response to glucose (13).
Such antagonists would have the advantage over the
previously used antibodies to glucagon in that they can
penetrate into the interstitial space (14,15). In the present
work, we examined the effect of a glucagon-receptor
antagonist, [des-His1,des-Phe6,Glu9]glucagon-NH2 (16), to
assess the role of interstitial glucagon in the acute responsiveness to glucose of in situ pancreatic ␤-cells in the
glucagon-rich dorsal part of the rat pancreas. Because high
nanomolar glucagon concentrations were reported to
activate GLP-1 receptors on ␤-cells (8), we also used a
GLP-1–receptor antagonist, exendin-(9-39)-NH2 (17). Our
669
INTERSTITIAL GLUCAGON AND THE SECRETORY STATE OF IN SITU ␤-CELLS
data show that, at least in the condition of rat pancreas
perfusion, there is little or no contribution of locally
secreted glucagon to the acute secretory response of islet
␤-cells to glucose.
RESEARCH DESIGN AND METHODS
Materials. The glucagon-receptor antagonist [des-His1,des-Phe6,Glu9]
glucagon-NH2 was prepared as described previously (16,18). Synthetic glucagon was purchased from Penninsula Laboratories Europe (Merseyside, U.K.).
Dextran T70 was from Amersham Pharmacia Biotech AB (Uppsala, Sweden);
HEPES was from Gibco BRL (Gaithersburg, MD); 3-isobutyl-1-methylxanthine
(IBMX) was from Aldrich (Janssen Chimica, Beerse, Belgium), and BSA was
from Roche Diagnostics (Mannheim, Germany). Exendin-(9-39)-NH2 was
provided by Dr. J. Vandekerckhove (Department of Physiological Chemistry,
State University of Ghent, Belgium). All other products were analytical grade
and obtained from either Merck (Darmstadt, Germany) or Sigma (St. Louis,
MO).
All studies involving animals were carried out according to the Belgium
regulation of animal welfare and after approval by the institution’s commission for animal experiments.
Cells and cell culture. Rat hepatocytes were isolated from adult male
Sprague-Dawley rats (200 –250 g) as described previously (19). Chinese
hamster lung (CHL) fibroblasts stably transfected with the rat GLP-1 receptor
cDNA (donated by Dr. B. Thorens, Department of Pharmacology and Toxicology, University of Lausanne, Switzerland [20]) were cultured in Dulbecco’s
modified Eagle’s medium (Gibco BRL) with 10% FCS (Gibco BRL), 2 mmol/l
glutamine, 25 mmol/l glucose, 100 mg/l streptomycin, 60 mg/l penicillin, and
500 mg/ml geneticin (Gibco BRL).
Measurement of cAMP production in hepatocytes and transfected CHL
cells. In rat hepatocytes, cAMP levels were measured as described previously
(8). Cells (5 ⫻ 104 cells/condition) were incubated for 15 min at 37°C in Earle’s
HEPES medium supplemented with 5 mmol/l glucose, 0.5% BSA, 250 ␮mol/l
IBMX, and different peptide hormones at the indicated concentrations.
Cellular cAMP content was measured using a commercially available
[125I]cAMP radioimmunoassay (RIA) kit (Amersham, Buckinghamshire, U.K.).
Transfected CHL cells were grown in 24-well plates (2 ⫻ 104 cells/
condition) until 80% confluence (⬃24 h) in 1 ml of culture medium. Cells were
washed three times with Earle’s HEPES containing 25 mmol/l glucose and
0.5% BSA and then incubated in a Steri-cult 200 incubator (Forma Scientific,
Parker, CO) for 15 min at 37°C in 1 ml of this medium, supplemented with 250
␮mol/l IBMX, glucagon, and GLP-1, with or without receptor antagonists.
After 15 min, the medium was aspirated, 500 ␮l of trichloroacetic acid (8%
[wt/vol] in water) was added, and each cup was sonicated. Of this sonicated
sample, 400 ␮l was collected in 5-ml plastic tubes (Falcon, Oxnard, CA) and
centrifuged at 2000g for 15 min. The soluble fraction was extracted three times
with 4 ml of water-saturated ether, lyophilized, redissolved in sodium acetate
buffer (0.05 mol/l, pH 5.8), and assayed for cAMP content. Results were
expressed as femtomoles per 103 initial cells at the beginning of cell culture.
Pancreas perfusion. Insulin secretion by in situ ␤-cells was investigated
using the isolated perfused rat pancreas model (21). Male Wistar rats (250 –350
g) were anesthetized with 60 mg/kg Nembutal (Sanofi Santé Animale, Libourne Cedex, France). To ensure perfusion of only the dorsal part of the
pancreas, we inserted a catheter (24 G; Neoflon, Ohmeda, Sweden) immediately into the celiac artery. The perfusate was collected via a 22-G Surflo
intravenous catheter (Terumo Europe NV, Leuven, Belgium) inserted and tied
into the portal vein. Anoxia time was kept to a minimum and was always ⬍3
min. The pancreas preparation was placed on a heated, moist surface
(surrounded by basal medium) and covered with Parafilm M (American
National Can Company, Chicago, IL), and the temperature was kept constant
at 37°C. Flow rate was constant throughout the perfusion experiment (2.5
ml/min); perfusions in which flow rates dropped below 2.5 ml/min before the
end of the experiment were excluded. The perfusion medium consisted of
Earle’s HEPES buffer (pH 7.35) (22) continuously gassed with 95% O2/5% CO2,
supplemented with 0.2% charcoal-treated BSA, 4% dextran T-70, and 1.4
mmol/l glucose (basal medium). At the time points indicated in the figures, 20
mmol/l glucose or glucose⫹peptides (glucagon/GLP-1 and/or receptor antagonists) was added to the medium. Perfusions were started with a 20-min/
equilibration period, followed by the experimental conditions, and ended with
neutral red perfusion to control the completeness of tissue perfusion. The
perfusate was collected every minute and analyzed for insulin and glucagon
content using RIAs (1,23), the detection limits being 20 pg and 4 pg,
respectively. To minimize degradation of peptides, fractions were collected in
50 ␮l of 7.5% precooled EDTA and immediately kept on ice. Data were
expressed as nanograms of hormone released per minute, taking into account
the dead space of perfusion, i.e., the time for neutral red to stain the pancreas
670
(1 min). Areas under the curve (AUC) in response to 20 mmol/l glucose
stimulations (min 11–20, 36 – 45, and 61–70), with or without GLP-1/glucagon
receptor agonists/antagonists (AUC1, AUC2, and AUC3) or 100 nmol/l isoproterenol subtraction were calculated after basal insulin release at 1.4 mmol/l
glucose. To adjust for the differences of the first glucose stimulation between
different experiments, we expressed AUC2 and AUC3 as a percentage of AUC1.
The potentiating effect of glucagon and GLP-1 and the effects of receptor
antagonists were subsequently calculated by subtracting the response to
glucose alone from the corresponding peak, in this way taking into account
the memory effect of glucose alone.
Statistical analysis. Data are expressed as mean values ⫾ SD of n
independent experiments in the table and the text and as mean values ⫾ SE
in the figures. Significance of differences between experimental conditions
was assessed by the two-tailed Student’s t tests. In the pancreas perfusions,
the magnitude of the secretory response after the first glucose stimulation
(AUC1) was normally distributed for the 36 tested pancreata in our study (data
not shown), validating the use of Student’s t tests for further analysis.
RESULTS
Characterization of the glucagon receptor antagonist
[des-His1,des-Phe6,Glu9]glucagon-NH2. The specificity
and potency of [des-His1,des-Phe6,Glu9]glucagon-NH2 as a
glucagon receptor antagonist was tested on primary rat
hepatocytes, a well-characterized reference system for
glucagon receptor signaling (24), and on CHL cells stably
transfected with the rat GLP-1–receptor gene (20). Significant antagonism of 1 nmol/l glucagon-induced cAMP
production in hepatocytes was noted between 0.3 ␮mol/l
peptide (17 ⫾ 5% inhibition; P ⬍ 0.005) and 10 ␮mol/l (80%
inhibition; P ⬍ 0.001). At 10 ␮mol/l [des-His1,des-Phe6,Glu9]
glucagon-NH2, antagonism was almost complete at 0.1
nmol/l glucagon (97 ⫾ 1% inhibition; P ⬍ 0.005) but
became nonsignificant at 10 nmol/l glucagon. The overall
effect was a shift of the concentration response curve to
the right (Fig. 1A), resulting in higher apparent EC50 for
glucagon (12 ⫾ 4 nmol/l) than in control cells (EC50 ⫽
0.9 ⫾ 0.4 nmol/l; P ⬍ 0.05). We previously reported that 1
␮mol/l exendin-(9-39)-NH2 had no effect on glucagoninduced cAMP production by rat hepatocytes (8).
In CHL cells expressing rat GLP-1 receptors, cAMP
content was elevated 4.5 ⫾ 0.3 times by 1 nmol/l glucagon
(Fig. 1B), confirming the observation (8) that glucagon can
activate GLP-1 receptors. Because these cells do not
express glucagon receptors, no inhibition of glucagoninduced cAMP production was observed with 10 ␮mol/l
[des-His1,des-Phe6,Glu9]glucagon-NH2, whereas 1 ␮mol/l
exendin-(9-39)-NH2 could inhibit this effect by 99 ⫾ 2%.
Likewise, 10 pmol/l GLP-1–induced cAMP production was
decreased by exendin-(9-39)-NH2, with 90 ⫾ 4% (stimulation above basal cAMP production: 285 ⫾ 87 vs. 28 ⫾ 18
fmol 䡠 103 initial cells⫺1 䡠 15 min⫺1 in the absence or
presence, respectively, of 1 ␮mol/l exendin-(9-39)-NH2;
n ⫽ 4; P ⬍ 0.01), whereas 10 ␮mol/l [des-His1,desPhe6,Glu9]glucagon-NH2 had no effect (312 ⫾ 113 fmol 䡠
103 initial cells⫺1 䡠 15 min⫺1 in the presence of [desHis1,des-Phe6,Glu9]glucagon-NH2; n ⫽ 4; not significant).
The effect of [des-His1,des-Phe6,Glu9]glucagon-NH2 was
subsequently tested in fluorescence-activated cell sorter–
purified ␤-cells (expressing both glucagon- and GLP-1receptors [8]), which were perifused with 20 mmol/l
glucose together with 1 nmol/l glucagon. This glucagon
concentration was chosen because 1) it has previously
been shown to potentiate significantly glucose-induced
insulin release in ␤-cells (1,8) and 2) 10 ␮mol/l [desHis1,des-Phe6,Glu9]glucagon-NH2 could inhibit its stimulaDIABETES, VOL. 51, MARCH 2002
K. MOENS AND ASSOCIATES
FIG. 1. Effect of [des-His1,des-Phe6,Glu9]Glucagon-NH2 and exendin-(9-39)-NH2 on glucagon-induced cAMP production. Glucagon-induced cAMP
production in rat hepatocytes (A) or CHL cells transfected (20) with rat GLP-1 receptor cDNA (B), either in the absence of antagonists (E) or
in the presence of 10 ␮mol/l [des-His1,des-Phe6,Glu9]Glucagon-NH2 (f) or 1 ␮mol/l exendin-(9-39)-NH2 (F) (B only). cAMP production was
measured after 15 min of incubation in buffer containing 250 ␮mol/l IBMX. Data are means ⴞ SE of four (A) or three (B) independent
experiments.
tory effect on cAMP production in rat hepatocytes by 80%
(Fig. 1B). Administration of 10 ␮mol/l [des-His1,desPhe6,Glu9]glucagon-NH2 counteracted the potentiating effect of 1 nmol/l glucagon on 20 mmol/l glucose-induced
insulin secretion from perifused cells by 78 ⫾ 6% (glucagon-stimulated insulin release, without versus with antagonist: 476 ⫾ 218 vs. 110 ⫾ 69 pg 䡠 103 cells⫺1; n ⫽ 4; P ⬍
0.02). As shown previously (8), this glucagon concentration was not affected by 1 ␮mol/l exendin-(9-39)-NH2,
whereas on the contrary, exendin-(9-39)-NH2 could inhibit
the potentiating effect of 10 nmol/l glucagon by ⬎40% (8).
Effect of [des-His1,des-Phe6,Glu9]glucagon-NH2 or
exendin-(9-39)-NH2 on glucose-induced insulin release. Having characterized the molecular tools to block
glucagon activation of glucagon- and GLP-1–receptors on
␤-cells, we next assessed their effect on glucose-induced
insulin release from the isolated perfused rat pancreas. As
expected (25), 20 mmol/l glucose stimulation resulted in
biphasic insulin secretory response (Fig. 2A), with a rapid
return to basal values when glucose was lowered to 1.4
mmol/l. Using a protocol with three successive 10-min
glucose stimulation periods alternated by 15-min intervals
of basal glucose (Fig. 2), we observed that the second
glucose stimulation resulted in a 1.5 ⫾ 0.2-fold larger
insulin secretory response than the first stimulation (AUC1
561 ⫾ 191 ng vs. AUC2 819 ⫾ 275 ng; n ⫽ 4; P ⬍ 0.02). As
is shown in Fig. 2A, the mean secretory response to the
third glucose stimulation was again 30% higher than the
second (AUC3 1,044 ⫾ 395 ng; P ⬍ 0.05). These results are
compatible with a glucose memory effect described by
Grill et al. (26,27). The potent insulin secretory response to
glucose— on average 55 ⫾ 18 ng/min for the first glucose
stimulation, 80 ⫾ 27 ng/min for the second, and 102 ⫾ 39
ng/min for the third— occurred in the presence of an
undetectable pancreatic glucagon output before or during
the 20 mmol/l glucose stimulation (⬍6 pmol/l). To investigate whether locally released glucagon may have influenced the observed secretory responsiveness, we added a
second series of perfusions, the glucagon receptor antagDIABETES, VOL. 51, MARCH 2002
onist [des-His1,des-Phe6,Glu9]glucagon-NH2 (10 ␮mol/l) to
the perfusion medium 15 min before and until the end of
the second period of glucose stimulation. This addition
had no effect on the time kinetics and magnitude of
glucose-stimulated insulin release (Fig. 2B) because AUC2
was again 1.3 ⫾ 0.2-fold higher (905 ⫾ 166 ng) than the
AUC1 caused by the first stimulation (682 ⫾ 75 ng; P ⫽
0.05). Furthermore, as in the control experiment, the third
glucose stimulation, performed in the absence of the
glucagon-receptor antagonist, resulted in 1.4 ⫾ 0.3-fold
higher release (AUC3 1,290 ⫾ 319 ng) than the second
glucose stimulation (mean ⫾ SD, n ⫽ 4; P ⫽ 0.056).
The absence of an antagonizing effect of [des-His1,desPhe6,Glu9]glucagon-NH2 might be attributable to the presence of local glucagon levels above 10 nmol/l, in which
range the antagonist was found to be ineffective. However,
such local levels should have been detected in the effluent.
Furthermore, at such high concentrations, glucagon is
known to activate ␤-cells via GLP-1 receptors (8), which
would have resulted in exendin-(9-39)-NH2 antagonizing
such effects, and this was not the case (Fig. 2C). Therefore,
neither the presence of the glucagon-receptor antagonist
[des-His1,des-Phe6,Glu9]glucagon-NH2 nor that of the
GLP-1 receptor antagonist exendin-(9-39)-NH2 could affect
the amplitude and time kinetics of glucose-induced insulin
secretory response from the perfused glucagon-rich lobe
of the rat pancreas. This could mean either that these
antagonists do not block their respective receptors on
␤-cells because they are degraded before binding occurs
or that occupancy of glucagon- and/or GLP-1–receptors
with endogenous glucagon is very low in the studied
conditions.
Effect of [des-His1,des-Phe6,Glu9]glucagon-NH2 and
exendin-(9-39)-NH2 on glucagon and GLP-1 potentiation of glucose-induced insulin release. To exclude
the possibility of antagonist degradation in the chosen
experimental conditions (Fig. 2), we added in a new series
of experiments with exogenous glucagon or GLP-1 in the
presence or absence of antagonist (Table 1). Glucagon (1
671
INTERSTITIAL GLUCAGON AND THE SECRETORY STATE OF IN SITU ␤-CELLS
FIG. 2. Effect of [des-His1,des-Phe6,Glu9]glucagon-NH2 and exendin-(939)-NH2 on glucose-induced insulin secretion from the perfused pancreas. Changes in the glucose concentration of the perfusion medium
are indicated in the top of each panel. A: Control pancreata stimulated
three times with 20 mmol/l glucose. The receptor antagonists [desHis1,des-Phe6,Glu9]glucagon-NH2 (desHisdesPhe; B) and exendin-(939)-NH2 (Exendin; C) were added 15 min before and during the second
glucose stimulation, as indicated by the bold line. Insulin secretion in
response to 1.4 mmol/l glucose was below the detection limit of the RIA
assay (20 pg). Data represent means ⴞ SE of four independent
experiments.
nmol/l) potentiated the second glucose stimulation more
than twofold (AUC2 as % of AUC1: 1 nmol/l glucagon 300 ⫾
4, n ⫽ 12, vs. control 146 ⫾ 16, n ⫽ 4; P ⬍ 0.001). Taking
into account the memory effect of glucose alone, a second
combined glucagon/glucose stimulation resulted in a 57%
larger insulin output than the first glucagon/glucose stimulation (AUC3 as % of AUC1: 1 nmol/l glucagon 427 ⫾ 65,
n ⫽ 4, vs. control 185 ⫾ 14, n ⫽ 4; P ⬍ 0.01). Presence of
the glucagon-receptor antagonist [des-His1,des-Phe6,Glu9]
glucagon-NH2 prevented 81 ⫾ 10% of this potentiating
effect of glucagon on insulin release during the third
stimulation with glucose (AUC3 as % of AUC1: 230 ⫾ 21 in
the presence vs. 427 ⫾ 65 in the absence of antagonist, n ⫽
4; P ⬍ 0.002). In contrast, we observed no antagonism of
the glucagon-potentiation of glucose-induced insulin release using 1 ␮mol/l exendin-(9-39)-NH2 (AUC3 as % of
672
AUC1: 389 ⫾ 61 in the presence vs. 427 ⫾ 65 in the absence
of antagonist, n ⫽ 4; not significant). As expected (28), 1
nmol/l GLP-1 potentiated glucose-induced insulin release
more than threefold (AUC2 as % of AUC1: 1 nmol/l GLP-1
500 ⫾ 66 ng, n ⫽ 8, vs. control 146 ⫾ 16, n ⫽ 4; P ⬍ 0.001).
Similar to the gain in secretory response after repeated
glucagon/glucose stimulation, the second GLP-1/glucose
exposure of the pancreas resulted in a 58% more efficient
amplification of the glucose-induced insulin release (AUC3
as % of AUC1: 1 nmol/l GLP-1 742 ⫾ 94, n ⫽ 4, vs. control
185 ⫾ 14, n ⫽ 4) than the first GLP-1 addition (P ⬍ 0.002).
This second GLP-1 stimulation was antagonized by 89 ⫾
9% (P ⬍ 0.0001) by simultaneous perfusion with exendin(9-39)-NH2. Therefore, the present data indicate that both
[des-His1,des-Phe6,Glu9]glucagon-NH2 and exendin-(9-39)NH2 effectively antagonize the effect of, respectively, glucagon on glucagon receptors and GLP-1 on GLP-1–
receptors. Together with the data in Fig. 2, these results
suggest that the glucose-induced insulin release from the
perfused pancreas can occur in the absence of significant
activity of occupied glucagon- and/or GLP-1–receptors.
Effect of ␣-cell stimulation on glucose-induced insulin release. To investigate the possible contribution of
local glucagon to glucose-induced insulin secretion in a
condition of elevated glucagon release, we performed a
perfusion experiment in which the second glucose stimulation and the 15 min preceding this period were supplemented with 100 nmol/l isoproterenol (Fig. 3). This
␤-adrenergic agonist has been shown to stimulate directly
the glucagon release from rat ␣-cells, without affecting
insulin secretion from ␤-cells (4). Furthermore, in contrast
to the effects observed in humans whereby the potentiation of glucose-induced insulin release by isoproterenol is
strong (29), interaction between isoproterenol and glucose
at the level of insulin release has been reported to be weak
and both stimulatory and inhibitory in the rat perfused
pancreas (30,31). As could be expected (4,32), perfusion
with 100 nmol/l isoproterenol induced a strong glucagon
secretory response at 1.4 mmol/l glucose, with a maximal
release of 0.24 ng/min that was at least fivefold above basal
and returned to levels below the detection limit at 40 min
(Fig. 3). This resulted in an AUC of 1.7 ⫾ 0.4 ng glucagon
and no detectable insulin during the 15-min perfusion with
1.4 mmol/l glucose and 100 nmol/l isoproterenol. Despite
this previous elevated glucagon secretion, glucoseinduced insulin secretion was not augmented as compared
with the control experiment without isoproterenol (AUC2
⫽ 148 ⫾ 18% of AUC1 in the presence vs. 146 ⫾ 16% in the
absence of isoproterenol, n ⫽ 4; not significant). Furthermore, glucagon remained elevated for 3 min more during
the 20 mmol/l glucose stimulation, resulting in an average
glucagon level of 0.09 ⫾ 0.02 ng/min (10 pmol/l). Because
locally secreted glucagon is diluted into the islet blood
flow, which represents in itself only 10% of the total
pancreatic blood flow (33), we can assume that locally
these glucagon concentrations may rise up to at least 100
pmol/l. We have previously shown that such glucagon
concentration can significantly potentiate glucose-induced
insulin secretion (1). However, during this period of
elevated pancreatic glucagon output (min 37–39) (Fig. 3),
glucose-induced insulin secretion was not significantly
altered. In isoproterenol-stimulated pancreata, insulin reDIABETES, VOL. 51, MARCH 2002
K. MOENS AND ASSOCIATES
TABLE 1
Glucagon and GLP-1 potentiation of 20 mmol/l glucose-induced insulin release from the isolated perfused rat pancreas: effect of the
receptor antagonists [des-His1,des-Phe6,Glu9]glucagon-NH2 and exendin-(9-39)-NH2
Condition
Released insulin (AUC)
% of stimulation 1 (min 11–20)
Stimulation 2
Stimulation 3
(min 36–45)
(min 61–70)
Control
Glucagon (1 nmol/l)
Glucagon (1 nmol/l) ⫹ [des-His1,des-Phe6,Glu9]Glucagon-NH2 (10 ␮mol/l)
Glucagon (1 nmol/l) ⫹ exendin-(9-39)-NH2 (1 ␮mol/l)
GLP-1 (1 nmol/l)
GLP-1 (1 nmol/l) ⫹ exendin-(9-39)-NH2 (1 ␮mol/l)
146 ⫾ 16 (4)
300 ⫾ 46 (12)
NT
NT
500 ⫾ 66 (8)
NT
185 ⫾ 14 (4)
427 ⫾ 65 (4)
230 ⫾ 21 (4)*
389 ⫾ 61 (4) NS
742 ⫾ 94 (4)
242 ⫾ 44 (4)†
Data represent the means ⫾ SD of insulin released (AUC after glucose stimulation 2 [min 36 – 45] or 3 [min 61–70] and expressed as % of AUC
after glucose stimulation 1 [min 11–20]) from (n) individual pancreata. Statistical significance of difference between absence or presence of
antagonists versus control (stimulation 2), glucagon, or GLP-1 (stimulation 3) was calculated via the two-tailed Student’s t test. *P ⬍ 0.02;
†P ⬍ 0.001. NS, not significant; NT, not tested.
lease during this period (298 ⫾ 115 ng) was 1.5 ⫾ 0.1-fold
higher than release during the comparable period of the
first glucose stimulation (193 ⫾ 76 ng). This memory effect
was similar to control pancreata (Fig. 2A), from which
release during min 37–39 (224 ⫾ 103 ng) was 1.4 ⫾ 0.3-fold
higher than that observed during the first 3 min of the first
glucose stimulation (158 ⫾ 59 ng).
Together, our results indicate that an episode of stimulated glucagon release, before or during glucose stimulation, does not potentiate the insulin response from the
pancreas, further supporting the idea that locally secreted
glucagon may not play a role in the responsiveness of
pancreatic ␤-cells to glucose through interstitial interactions.
DISCUSSION
In rats and mice, ␤-cells are located in the islet center,
whereas the non–␤-cells are situated in the islet mantle
(34). For the dorsal rat pancreas, which contains most
glucagon cells of the gland, this means that the ␤-cell core
of the islet is surrounded by a mantle of glucagon-secreting and somatostatin-secreting cells (9,35). Because both
glucagon and somatostatin exert powerful effects on the
glucose response of isolated rat ␤-cells (1,36), the possibility was proposed that local (paracrine) interactions of
glucagon and somatostatin with islet ␤-cells modulate the
secretory response of the pancreas to elevated glucose
(37). The data in the present study suggest that the strong
insulin secretory response to glucose of the isolated
perfused pancreas can occur in the absence of interstitial
influences by locally secreted glucagon. A potent insulin
secretory response of ␤-cells to glucose was demonstrated
in the isolated rat perfused pancreas in a condition of
glucagon levels ⬍6 pmol/l in the pancreatic effluent. However, locally secreted glucagon is diluted first in the islet
blood circulation and subsequently in the total pancreatic
blood circulation. Furthermore, in vivo normal blood
glucagon levels are reported to be between 1 and 100
pmol/l (38 – 40), whereas 1 pmol/l glucagon is enough to
stimulate significantly insulin secretion from the perfused rat pancreas (41). The use of a specific glucagonand GLP-1–receptor antagonist, [des-His1,des-Phe6,Glu9]
glucagon-NH2 and exendin-(9-39)-NH2 respectively, allowed us to assess interstitial influences of locally
secreted glucagon, because these antagonists can enter
the interstitial space (14,15).
We demonstrate here that [des-His1,des-Phe6,Glu9]
glucagon-NH2 inhibits specifically glucagon receptors,
blocking 80% of 1 nmol/l glucagon activation of hepatocytes, and that it does not interfere with GLP-1–receptor
activation by either glucagon or GLP-1. We acknowledge
the problem of low potency of [des-His1,des-Phe6,Glu9]
FIG. 3. Effect of isoproterenol on glucagon
secretion and subsequent glucose-induced
insulin secretion from the perfused pancreas. Insulin (E) and glucagon (F) secretion were assessed as outlined in RESEARCH
DESIGN AND METHODS. Changes in the perfusion
medium are indicated in the top. The ␤-adrenergic agonist isoproterenol (100 nmol/l)
was added 15 min before and during the
second glucose stimulation, as indicated by
the bold line. Insulin secretion in response
to 1.4 mmol/l glucose was below the detection limit of the RIA assay (20 pg); glucagon
secretion values below the detection limit of
the RIA assay (4 pg) are not shown. Data
represent the means ⴞ SE of four independent experiments.
DIABETES, VOL. 51, MARCH 2002
673
INTERSTITIAL GLUCAGON AND THE SECRETORY STATE OF IN SITU ␤-CELLS
glucagon-NH2 for the glucagon receptors. In a concentration-response experiment performed with hepatocytes
(data not shown), we observed concentration-dependent
effects of the antagonist on 1 nmol/l glucagon-induced
cAMP production. Because of solubility and pH effects on
the buffers, we could not use concentrations ⬎10 ␮mol/l.
This concentration could significantly antagonize glucagon
concentrations up to 1 nmol/l in hepatocytes (Fig. 1A).
However, if higher glucagon levels exist locally, it could be
expected that, in addition to glucagon receptors, GLP-1
receptors will be activated (8). Therefore, 1 ␮mol/l exendin-(9-39)-NH2 was used in parallel (Fig. 2C). Because
glucagon amplifies 20 mmol/l glucose-induced insulin release (1,8), perfusion of the pancreas with either [desHis1,des-Phe6,Glu9]glucagon-NH2 or exendin-(9-39)-NH2
should result in a lower glucose-induced insulin secretory
response than in the control pancreas on the condition
that local glucagon can activate ␤-cells. That we did not
observe any inhibition of the pancreatic glucose-induced
insulin response by either antagonist indicates that if
glucagon interacts with ␤-cells, then it occurs after dilution in the general circulation (42,43), i.e., after at least one
passage through the liver. Therefore, the lack of measurable glucagon levels in the pancreatic effluent may be
explained by an extremely low glucagon secretion in the
present experimental conditions. The observed effect of 1
nmol/l exogenous glucagon on insulin release (Table 1)
further supports the absence of high glucagon concentrations influencing neighboring ␤-cells. This argument was
already used by Kawai et al. (42). It is unlikely that our
negative results were caused by antagonist degradation
before binding on their respective receptors. First, high
concentrations were used, so degradation would have
been extremely rapid. Second, [des-His1,des-Phe6,Glu9]
glucagon-NH2 and exendin-(9-39)-NH2 could block ⬎80%
of the insulin-stimulating effect of, respectively, 1 nmol/l
glucagon and GLP-1 that was added to the perfusion
medium. Positive controls for the action of these antagonists in the perfused pancreas were therefore obtained.
The same degree of inhibition was also observed with
hepatocytes stimulated with 1 nmol/l glucagon (Fig. 1A),
further supporting the idea that in the present model the
used exogenous glucagon is not diluted extensively by
endogenous glucagon.
The low rates of glucagon release under the chosen
experimental conditions are a limitation of the present
experimental system. We therefore induced glucagon release with the ␤-adrenergic agonist isoproterenol, a condition that is known to stimulate ␣-cells (4,32), while at the
same time this agonist is not expected to stimulate ␤-cells
directly (4). We observed that this condition did not
potentiate the insulin secretory response of pancreatic
␤-cells to 20 mmol/l glucose, despite a clearly elevated rate
of glucagon release before and during the first minutes of
glucose stimulation. These observations further support
the idea that the secretory response of the intact pancreas
is not dependent on locally released glucagon. It is extremely difficult to rule out the possibility that tight
junctions between neighboring cells seal off intercellular
compartments where locally released glucagon is allowed
to interact with receptors on ␤-cells. Considering our data
(Fig. 3), the release of such glucagon pools should not be
674
isoproterenol-sensitive, which is very unlikely. It can thus
be proposed that under the present conditions, local
release of glucagon has no major effect on the magnitude
of glucose-induced insulin release from the isolated rat
pancreas. Extrapolation of these observations to in vivo
conditions will require appropriate models to measure
glucagon levels within the interstitial space, where it can
potentiate insulin secretion.
In summary, our data demonstrate that the potent
glucose-induced insulin release from the isolated perfused
rat pancreas is not dependent on the priming effect of
locally released glucagon, at least under the chosen experimental conditions. This both supports the view that islet
blood flow direction is not from ␣- to ␤-cells (11) and
indicates that glucagon in the islet interstitium is not in
contact with glucagon- and GLP-1–receptors on ␤-cells. If
the present observations on isolated rat pancreas (where
no evidence for interstitial glucagon is found) can be
extrapolated to human pancreas, then these data can be
relevant for the clinical development of glucagon-receptor
antagonists (16), which aim to suppress glucagon-stimulated liver metabolism rather than glucose-induced insulin
secretion.
ACKNOWLEDGMENTS
The work in this article has been supported by grants
G.3127.93 and G.0376.97 from the Flemish Fund for Scientific Research (FWO Vlaanderen), the Ministerie van de
Vlaamse Gemeenschap, departement Onderwijs (Geconcerteerde Onderzoeksactie 1807), the Services of the Belgian Prime Minister (Interuniversity Attraction Pole P4/
21), and the U.S. Public Health Service (Grant DK-21085).
We thank Dr. V. Rogiers, Sonia Beken, and Karen De
Smet for isolating rat hepatocytes; Dr. B. Thorens (Institut
de Pharmacologie, Université de Lausanne, Switzerland)
for kindly donating CHL cells transfected with the GLP-1
receptor; and Dr. J. Vandekerckhove (Department of Physiological Chemistry, State University of Ghent, Belgium)
for providing synthetic exendin-(9-39)-NH2. Dr. L. Kaufman is acknowledged for statistical advice.
REFERENCES
1. Pipeleers DG, Schuit FC, in’t Veld P, Maes E, Hooghe-Peters EL, Van De
Winkel M, Gepts W: Interplay of nutrients and hormones in the regulation
of insulin release. Endocrinology 117:824 – 833, 1985
2. Schuit FC, Pipeleers DG: Regulation of adenosine 3⬘,5⬘-monophosphate
levels in the pancreatic B cell. Endocrinology 117:834 – 840, 1985
3. Moens K, Heimberg H, Flamez D, Huypens P, Quartier E, Ling Z, Pipeleers
D, Gremlich S, Thorens B, Schuit F: Expression and functional activity of
glucagon, glucagon-like peptide I, and glucose-dependent insulinotropic
peptide receptors in rat pancreatic islet cell. Diabetes 45:257–261, 1996
4. Schuit FC, Pipeleers DG: Differences in adrenergic recognition by pancreatic A and B cells. Science 232:875– 877, 1986
5. Trimble ER, Halban PA, Wollheim CB, Renold AE: Functional differences
between rat islets of ventral and dorsal pancreatic origin. J Clin Invest
69:405– 413, 1982
6. Pipeleers D: Purified islet cells in diabetes research. Horm Res 23:225–234,
1986
7. Huypens P, Ling Z, Pipeleers D, Schuit, F: Glucagon receptors on human
islet cells contribute to glucose competence of insulin release. Diabetologia 43:1012–1019, 2000
8. Moens K, Flamez D, Van Schravendijk C, Ling Z, Pipeleers D, Schuit F:
Dual glucagon recognition by pancreatic ␤-cells via glucagon and glucagon-like peptide 1 receptors. Diabetes 47:66 –72, 1998
9. Baetens D, Malaisse-Lagae F, Perrelet A, Orci L: Endocrine pancreas:
three-dimensional reconstruction shows two types of islets of Langerhans.
Science 206:1323–1324, 1979
DIABETES, VOL. 51, MARCH 2002
K. MOENS AND ASSOCIATES
10. Creutzfeldt W: The incretin concept today. Diabetologia 16:75– 85, 1979
11. Bonner-Weir S, Orci L: New perspectives on the microvasculature of the
islets of Langerhans in the rat. Diabetes 31:883– 889, 1982
12. Samols E, Stagner JI: Intra-islet regulation. Am J Med 85:31–35, 1988
13. Weir GC, Bonner-Weir S: Islets of Langerhans: the puzzle of intraislet
interactions and their relevance to diabetes. J Clin Invest 85:983–987, 1990
14. Kvietys PR, Perry MA, Granger DN: Permeability of pancreatic capillaries
to small molecules. Am J Physiol 245:G519 –G524, 1983
15. Rippe B, Haraldsson B: Transport of macromolecules across microvascular walls: the two-pore theory. Physiol Rev 74:163–219, 1994
16. Van Tine BA, Azizeh BY, Trivedi D, Phelps JR, Houslay MD, Johnson DG,
Hruby VJ: Low level cyclic adenosine 3⬘,5⬘-monophosphate accumulation
analysis of [des-His1,des-Phe6,Glu9]glucagon-NH2 identifies glucagon antagonists from weak partial agonists/antagonists. Endocrinology 137:
3316 –3322, 1996
17. Göke R, Fehmann H-C, Linn T, Schmidt H, Krause M, Eng J, Göke B:
Exendin-4 is a high potency agonist and truncated exendin-(9-39)-amide an
antagonist at the glucagon-like peptide 1-(7-36)-amide receptor of insulinsecreting ␤-cells. J Biol Chem 268:19650 –19655, 1993
18. Azizeh BY, Van Tine BA, Sturm NS, Hutzler AM, David C, Trivedi D, Hruby
VJ: [des-His1,des-Phe6,Glu9]glucagon amide: a newly designed “pure” glucagon antagonist. Bioorg Med Chem Lett 5:1849 –1852, 1995
19. Rogiers V, Paeme G, Vercruysse A, Bouwens L: Procycline metabolism in
isolated rat liver cells. In Pharmacological, Morphological and Physiological Aspects of Liver Ageing, Van Bezooyen CFA, Ed. Eurage, Rijswijk,
Holland, 1984, p. 155–163
20. Widmann C, Bürki E, Dolci W, Thorens B: Signal transduction by the
cloned glucagon-like peptide-1 receptor: comparison with signaling by the
endogenous receptors of ␤ cell lines. Mol Pharmacol 45:1029 –1035, 1994
21. Grodsky GM, Fanska R: The in vitro perfused pancreas. In Methods in
Enzymology, Vol. 39: Hormone Action, Part D: Isolated Cells, Tissues,
and Organ Systems. Hardman JG, O’Malley BW, Eds. New York, NY,
Academic Press, 1975, p. 364 –372
22. Pipeleers DG, in’t Veld P, Van De Winkel M, Maes E, Schuit FC, Gepts W:
A new in vitro model for the study of pancreatic A and B cells.
Endocrinology 117:806 – 816, 1985
23. Pipeleers DG, Schuit FC, Van Schravendijk CFH, Van De Winkel M:
Interplay of nutrients and hormones in the regulation of glucagon release.
Endocrinology 117:817– 823, 1985
24. Fehlman M, Morin O, Kitagbi P, Freychet P: Insulin and glucagon receptors
in isolated rat hepatocytes: comparison between hormone binding and
amino acid transport stimulation. Endocrinology 109:253–261, 1981
25. Curry DL, Bennett LL, Grodsky GM: Dynamics of insulin secretion by the
perfused rat pancreas. Endocrinology 83:572–584, 1968
26. Grill V, Adamson U, Cerasi E: Immediate and time dependent effects of
glucose on insulin release from rat pancreatic tissue. J Clin Invest
61:1034 –1043, 1977
DIABETES, VOL. 51, MARCH 2002
27. Grill V: Time and dose dependencies for priming effect of glucose on
insulin secretion. Am J Physiol 240:E24 –E31, 1981
28. Weir GC, Mojsov S, Hendrick GK, Habener JF: Glucagonlike peptide I
(7-37) actions on endocrine pancreas. Diabetes 38:338 –342, 1989
29. Halter JB, Graf RJ, Porte D Jr: Potentiation of insulin secretory responses
by plasma glucose levels in man: evidence that hyperglycemia in diabetes
compensates for impaired glucose potentiation. J Clin Endocrinol Metab
48:946 –954, 1979
30. Loubatieres-Mariani MM, Chapa J, Puech R, Manteghetti M: A different
action of hypothermia on insulin release from the isolated, perfused rat
pancreas, depending on the stimulating agent. Diabetes 29:895– 898, 1980
31. Weir GC, Clore ET, Zmachinski CJ, Bonner-Weir S: Islet secretion in a new
experimental model for non-insulin-dependent diabetes. Diabetes 30:590 –
595, 1981
32. Gerich JE, Lovinger R, Grodsky GM: Inhibition by somatostatin of glucagon and insulin release from the perfused rat pancreas in response to
arginine, isoproterenol and theophylline: evidence for a preferential effect
on glucagon secretion. Endocrinology 96:749 –754, 1975
33. Jansson L, Hellerström C: Stimulation by glucose of the blood flow to the
pancreatic islets of the rat. Diabetologia 25:45–50, 1983
34. Orci L, Unger RH: Functional subdivision of islets of Langerhans and
possible role of D cells. Lancet 2:1243–1244, 1975
35. Orci L, Baetens D, Ravazzola M, Stefan Y, Malaisse-Lagae F: Pancreatic
polypeptide and glucagon: non-random distribution in pancreatic islets.
Life Sci 19:1811–1816, 1976
36. Schuit FC, Derde MP, Pipeleers DG: Sensitivity of pancreatic A and B cells
to somatostatin. Diabetologia 32:207–212, 1989
37. Unger RH, Orci L: Possible roles of the pancreatic D-cell in the normal and
diabetic states. Diabetes 26:241–244, 1977
38. Yamamoto T, Raskin P, Aydin I, Unger R: Effects of insulin on the response
of immunoreactive glucagon to an intravenous glucose load in human
diabetes. Metabolism 28:568 –574, 1979
39. Hansen CB, Jen KC, Pek SB, Wolfe RA: Rapid oscillations in plasma
insulin, glucagon, and glucose in obese and normal weight humans. J Clin
Endocrinol Metab 54:785–792, 1982
40. Golay A, Swislocki ALM, Chen YI, Jaspan JB, Reaven GM: Effect of obesity
on ambient plasma glucose, free fatty acid, insulin, growth hormone, and
glucagon concentrations. J Clin Endocrinol Metab 63:481– 484, 1986
41. Kieffer TJ, Heller RS, Unson CG, Weir GC, Habener JF: Distribution of
glucagon receptors on hormone-specific endocrine cells of rat pancreatic
islets. Endocrinology 137:5119 –5125, 1996
42. Kawai K, Ipp E, Orci L, Perrelet A, Unger RH: Circulating somatostatin acts
on the islets of Langerhans by way of a somatostatin-poor compartment.
Science 218:477– 478, 1982
43. Bonner-Weir S: Morphological evidence for pancreatic polarity of ␤-cell
within islets of Langerhans. Diabetes 37:616 – 621, 1988
675