Am J Physiol Endocrinol Metab 307: E644–E652, 2014.
First published August 12, 2014; doi:10.1152/ajpendo.00244.2014.
Hepatic portal vein denervation impairs oral glucose tolerance but not
exenatide’s effect on glycemia
Viorica Ionut,1 Ana Valeria B. Castro,1 Orison O. Woolcott,1 Darko Stefanovski,1 Malini S. Iyer,1
Josiane L. Broussard,1 Miguel Burch,2 Ram Elazary,2 Cathryn M. Kolka,1 Hasmik Mkrtchyan,1
Isaac Asare Bediako,1 and Richard N. Bergman1
1
Diabetes and Obesity Research Institute and 2Department of Surgery, Cedars Sinai Medical Center, Los Angeles, California
Submitted 29 May 2014; accepted in final form 11 August 2014
Ionut V, Castro AV, Woolcott OO, Stefanovski D, Iyer MS,
Broussard JL, Burch M, Elazary R, Kolka CM, Mkrtchyan H,
Bediako IA, Bergman RN. Hepatic portal vein denervation impairs
oral glucose tolerance but not exenatide’s effect on glycemia. Am J
Physiol Endocrinol Metab 307: E644 –E652, 2014. First published
August 12, 2014; doi:10.1152/ajpendo.00244.2014.—The hepatoportal area is an important glucohomeostatic metabolic sensor, sensing
hypoglycemia, hyperglycemia, and hormones such as glucagon-like
peptide-1 (GLP-1). We have reported previously that activation of
hepatoportal sensors by intraportal infusion of glucose and GLP-1 or
by subcutaneous administration of GLP-1 receptor activator exenatide
and of intraportal glucose improved glycemia independent of corresponding changes in pancreatic hormones. It is not clear whether this
effect is mediated via the portal vein (PV) or by direct action on the
liver itself. To test whether receptors in the PV mediate exenatide’s
beneficial effect on glucose tolerance, we performed 1) paired oral
glucose tolerance tests (OGTT) with and without exenatide and 2)
intravenous glucose tolerance tests before and after PV denervation in
canines. Denervation of the portal vein affected oral glucose tolerance; post-denervation (POST-DEN) OGTT glucose and insulin AUC
were 50% higher than before denervation (P ⫽ 0.01). However, portal
denervation did not impair exenatide’s effect to improve oral glucose
tolerance (exenatide effect: 48 ⫾ 12 mmol·l⫺1·min before vs. 64 ⫾ 26
mmol·l⫺1·min after, P ⫽ 0.67). There were no changes in insulin
sensitivity or secretion during IVGTTs. Portal vein sensing might play
a role in controlling oral glucose tolerance during physiological
conditions but not in pharmacological activation of GLP-1 receptors
by exenatide.
hepatic portal vein; denervation; glucagon-like peptide-1; exenatide
that the gastrointestinal tract plays an
important role in controlling glucose homeostasis. The gut
releases signal molecules (nutrients, hormones, metabolites)
that directly, or via neural connections, alert the body to the
presence of a meal and prepare an appropriate response to
maintain glycemia. We and others (11, 15) have shown that the
presence of glucose and of the intestinal hormone glucagonlike peptide-1 (GLP-1) in the portohepatic area triggers a series
of events that result in the lowering of peripheral glycemia.
The latter effect is mediated via changes in glucose uptake/
output by the liver and by increased systemic glucose uptake
(28). The lowering of glucose is independent of the “ileal break
effect” of GLP-1 and peptide YY (PYY) and independent of
corresponding changes in pancreatic hormones (insulin, glucagon). GLP-1-dependent glucose lowering could be mediated
via GLP-1 receptors in the hepatoportal area (13). Vahl et al.
IT IS NOW RECOGNIZED
Address for reprint requests and other correspondence: V. Ionut, Diabetes
and Obesity Research Institute, Cedars Sinai Medical Center, 8700 Beverly
Blvd., THAL 104, Los Angeles, CA 90048 (e-mail: Viorica.Ionut@cshs.org).
E644
(25) identified GLP-1 receptor mRNA in the nodose ganglia
and GLP-1 protein in the portal vein of rats. Intraportal infusion of glucose and a GLP-1 receptor antagonist resulted in a
worsening of glucose tolerance, which could have been mediated by portal receptors (25). Alternatively, GLP-1-induced
signal initiation could happen not in the portal vein area but
directly in the liver. Although controversial, the presence of
GLP-1 liver receptors is suggested by some studies (6). GLP-1
could bind to hepatic GLP-1 receptors and either directly
activate signaling cascades that increase glucose storage and/or
decrease glucose output (intrahepatic communication), or it
could activate autonomic nerve endings that signal to the brain.
The brain in turn would signal back to liver and perhaps other
organs (muscle, adipose). However, endogenous GLP-1 might
contribute to glucohomeostatic regulation via portally initiated
signaling, and equally important is that this effect could be
shared by the GLP-1 receptor activators used currently in type
2 diabetes treatment, such as exenatide. GLP-1 is synthesized
in the L cells of the distal ileum and colon and released in
response to a meal in the intestinal capillaries (7). The portal
vein and its tributaries collect blood from the intestinal area
and direct it to the liver; in this way the hepatoportal area is
exposed to a relatively high concentration of endogenous
GLP-1. Unlike GLP-1, exenatide and other GLP-1 mimetics
are administered subcutaneously, activating both portal and
nonportal receptors. Thus, the contribution of GLP-1 receptors
in the portal vein to the control of glycemic excursions by
subcutaneous exenatide is currently unknown. To date, there is
no direct proof via portal vein denervation that GLP-1 activates
receptors in the portal vein or that neural connections between
portal vein and the central nervous system are relaying information that ultimately leads to changes in glucose uptake/
release in various organs.
The aim of the current study was to directly test, via portal
vein denervation, the hypothesis that sensors for glucose and
GLP-1 in the portal vein mediate the lowering of glycemia seen
with subcutaneous exenatide administration.
METHODS
Animals
Experiments were performed in male mongrel dogs (1 yr old, 28.2 ⫾ 0.9
kg body wt) that were in the conscious relaxed state (n ⫽ 13), housed
under controlled kennel conditions in the Cedars Sinai Medical Center
vivarium, and fed once daily with a standard diet (27% protein, 32%
fat, and 41% carbohydrate; Labdiet; PMI Nutrition International,
Richmond, IN). The animals were used for experiments only if judged
to be in good health as determined by body temperature, hematocrit,
regularity of food intake, and direct observation. All surgical and
0193-1849/14 Copyright © 2014 the American Physiological Society
Downloaded from journals.physiology.org/journal/ajpendo (068.004.044.057) on November 11, 2020.
http://www.ajpendo.org
PORTAL VEIN DENERVATION AND EXENATIDE’S EFFECT
experimental procedures were approved by the Cedars Sinai Medical
Center Institutional Animal Care and Use Committee.
Experimental Design
Experiments were performed in the morning after 12–16 h of
fasting. In each animal, before and after portal denervation, we
performed a metabolic assessment [oral glucose tolerance test
(OGTT) with and without exenatide and intravenous glucose tolerance test (IVGTT)] and a hypoglycemic clamp (for physiological
verification of portal vein denervation). At the end of the experimental
period, the animals were euthanized and portal veins harvested for
measurement of tyrosine hydroxylase by immunohistochemistry.
Surgery
Portal vein denervation was performed under general anesthesia, as
described previously (10). After a midline incision, the hepatic portal
vein was exposed. A myelin-specific dye (toluidine blue 1%) was used
to visualize nerve bundles along the portal vein, hepatic artery,
common bile duct, and liver hilum. Only the visible nerves along the
portal vein were cut, and the vein was painted with 10% phenol/
alcohol vol/vol. Special care was taken to preserve the nerves running
along the hepatic artery and its branches, the hepatic hilum nerves
where nerve fibers enter the liver, the hepatic branch of the vagus
nerve, and innervation of the common bile duct. Specifically, the
portal vein was denervated distally (1–2 cm) from the liver hilum. It
has been shown that, in rat, the portal vein glucose sensor is situated
1–3 cm from the liver hilum (8, 9), so this distance is likely to be even
longer in larger mammals. The abdomen was then sutured closed.
Animals recovered for ⱖ7 days after surgery before experiments were
performed in the conscious state (see below).
OGTT
Basal blood samples were drawn from a peripheral vein at ⫺20,
⫺10, and ⫺1 min. At t ⫽ 0, 25 g of glucose (55 ml of 50%dextrose,
454 mg/ml) was given as a bolus by oral gavage. Additional blood
samples were taken at 15, 30, 45, 60, 90, 120, and 180 min. In a subset
of animals (n ⫽ 9), 640 mg of acetaminophen was added to the oral
gavage solution to measure gastric emptying (24). The OGTTs were
performed in 12 animals.
OGTT With Exenatide
The protocol is identical to a regular OGTT, except that after basal
sampling, at t ⫽ 0 an injection of exenatide (20 ␮g of Byetta; Amylin,
San Diego, CA) was administered subcutaneously in the interscapular
region. The OGTT with exenatide was performed in a subset of n ⫽
6 animals.
IVGTT
Blood samples were drawn from a peripheral vein at ⫺20, ⫺10,
and ⫺1 min. Fasting values were defined as the average of the three
basal samples. At t ⫽ 0, 0.3 g/kg body wt glucose (50% dextrose, 454
mg/ml; B Braun, Irvine, CA) was injected into a peripheral vein.
Additional samples were taken at 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, and
19 min. At t ⫽ 20 min, a bolus of 0.03 U/kg porcine insulin was
injected into a peripheral vein, followed by sampling at 22, 23, 24, 25,
30, 40, 50, 60, 70, 80, 90, 100, 120, 140, 160, and 180 min. IVGTTs
were performed in a subset of n ⫽ 6 animals.
Hypoglycemic Clamps
Intracatheters were acutely placed in peripheral veins for infusion
of insulin and glucose and for blood sampling. After basal sampling
(⫺20 to 0 min) at t ⫽ 0, insulin was infused at 30 pmol·kg⫺1·min⫺1
to slowly decrease glycemia to 2.5–3 mmol/l. Glucose (50% dextrose)
was infused at a variable rate. The glucose infusion rate was adjusted
E645
every 10 min to achieve ⬃0.55 mmol/l reductions in blood glucose
over 40-min periods. Thus, peripheral glucose decreased stepwise to
a nadir that was reached between 120 and 160 min (steady state).
Blood samples were taken every 10 min for measurements of glucose
and insulin. Blood samples for measurement of catecholamines were
taken at basal and at steady state. A subset of the hypoglycemic clamp
data (8 out of 12) has been published previously as part of a larger
study (n ⫽ 17) (10).
Blood Samples and Assay Measurements
Samples were collected into tubes coated with lithium fluoride
and heparin containing EDTA (for glucose and insulin), EDTAaprotinin (for glucagon), and EDTA-SMS and flash-frozen (for
catecholamines). Plasma glucose concentrations were determined
using a YSI 2300 autoanalyzer (YSI, Yellow Springs, OH). Insulin
was measured using a human ELISA kit (Millipore, Billerica, MA)
adapted in our laboratory for dog plasma (11). We used ELISA kits to
measure catecholamines (Rocky Mountain Diagnostics, Colorado
Springs, CO), acetaminophen (Immunalysis, Pomona, CA), and total
and intact GLP-1 (Millipore). PYY was measured using RIA (Millipore).
Tyrosine Hydroxylase Immunohistochemistry
In a subset of animals, tyrosine hydroxylase immunohistochemistry
was performed on portal vein tissue, as described previously (10).
Briefly, the portal tissue was washed with 0.015 M PBS and dehydrated with a series of increasing concentrations of ethanol. After
dehydration, the tissue was incubated in xylene [2 ⫻ 60 min and then
infiltrated with paraffin, a paraplast-embedding medium (Sigma
Chemical)] at 60°C overnight and then embedded with fresh paraffin.
Each pad was sliced across its extent at 5 ␮m using a rotary
microtome (American Optical Instrument, Buffalo, NY). Sections
were placed in a tissue-floating water bath (37°C) immediately after
being sliced and were mounted on glass slides. Slides were left on a
slide-warming table (37°C) to air-dry overnight. Slides were placed in
a series of xylene to remove paraffin and then a series of ethanol. After
a wash in tap water, the slides were placed in 3% hydrogen peroxidemethanol solution for 10 min. After a wash the slides were incubated
for 25 min in citrate buffer (pH 6) at 95°C. The slides were brought
to room temperature, rinsed in PBS containing 0.05% Tween-20
(PBST), incubated with anti-tyrosine hydroxylase (AB152; Millipore,
Chicago, IL) at a dilution of 1:100 at 4°C overnight, and rinsed in
PBST. The slides were rinsed with PBST and incubated with Dako
EnVision⫹ System-HRP Labeled Polymer anti-rabbit (K4003; Dako)
for 30 min at room temperature. After a rinse with PBST, the slides
were incubated with 3,3=-diaminobenzidine for visualization. Subsequently, the slides were washed in tap water, counterstained with
Harris’ hematoxylin, dehydrated in ethanol, and mounted with media.
Two sections per animal were used to assess the lesion.
Calculations
IVGTT-derived indices: calculation of minimal model parameters.
Insulin sensitivity (SI) and glucose effectiveness (Sg) were calculated
by the application of the minimal model of glucose kinetics to the time
course of glucose and insulin from the IVGTT using MINMOD
Millennium software (version 6.02, 2004). The acute insulin response
to glucose (AIRg) was calculated as the area under the curve (AUC)
of the insulin concentrations above the average of the basal values
from 0 to 10 min (2). The disposition index was calculated as the
product SI ⫻ AIRg for each experiment. Kg (glucose tolerance) was
calculated as the slope of the linear regression of the natural log of
plasma glucose concentration vs. time from 10 to 19 min (1).
Insulin clearance. Fractional transfer rate reflecting insulin clearance was estimated using frequently sampled intravenous glucose
tolerance data by fitting the insulin values after the insulin bolus to a
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00244.2014 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (068.004.044.057) on November 11, 2020.
E646
PORTAL VEIN DENERVATION AND EXENATIDE’S EFFECT
epinephrine change (pmol/l)
800
pre DEN
post DEN
600
400
**
200
0
Fig. 1. Epinephrine response to hypoglycemia before (PRE-DEN; open bar)
and after (POST-DEN; black bar) portal vein denervation.
single exponential insulin clearance model (26). Metabolic clearance
of insulin (ml·kg⫺1·min⫺1) was calculated as the ratio of the dose of
insulin in (␮U/kg) and the area under the fitted decay curve of the
single exponential insulin clearance model extended to infinity (22).
Exenatide’s effect on glycemia was calculated for the same condition (before or after denervation) as glucose AUC during OGTT
minus glucose AUC during OGTT ⫹ exenatide.
Tyrosine hydroxylase quantification. Scanning and analyses were
performed through the Translational Pathology Core Laboratory,
Department of Pathology and Laboratory Medicine, University of
California Los Angeles. The technician who performed the analyses
was blind to the sample identity. Images were acquired and analyzed
using Ariol SL-50 (Applied Imaging, San Jose, CA). The Ariol
scanner is based on an Olympus BX61 microscope video camera (Jai
CVM2CL). Ariol can perform automatic quantification of protein
levels by measuring staining intensity and the pattern and location of
staining in each cell. Slides were scanned at ⫻20 objective magnification with the DAPI and FITC filters. Optimal exposure times were
determined before automated scanning. After scanning, threshold
levels of the individual signals were optimized before final analysis.
The readouts included area of signal (area of signal above the signal
threshold) and total area analyzed and were calculated by the software. The results are reported as percent marker staining (TH fluorescence) relative to the area of tissue analyzed.
Statistical Analysis and Power Calculation
Statistical analysis. Results are presented as means ⫾ SE unless
indicated otherwise. Repeated-measurements ANOVA was used for
timeline comparisons; paired t-tests with Bonferroni correction were
used for time point comparisons before and after denervation. AUC
was calculated using the trapezoid rule. All differences were considered statistically significant when P ⬍ 0.05.
Power calculation. The primary endpoint of the study was to detect
a change in exenatide effect (calculated as glucose AUC during
OGTT ⫺ glucose AUC during OGTT ⫹ exenatide). Exenatide effect
mean ⫾ SD, as determined from preliminary data, was 55 ⫾ 13
mmol/min. To detect a clinically relevant 30% decrease, for a twosided t-test, ␣ ⫽ 0.05 and power ⫽ 0.9, with a sample size of n ⫽ 6.
The secondary end point was to detect a change in glucose AUC after
denervation. Our database of OGTT in the canine indicates that
glucose AUC mean ⫾ SD ⫽ 1,902 ⫾ 1,177; for a clinically relevant
50% change, ␣ ⫽ 0.05 and power ⫽ 0.8; the sample size is 12. The
post hoc power calculation for IVGTT shows that (for insulin secretion) m1 ⫽ 438, m2 ⫽ 300, and SD ⫽ 100; n ⫽ 6, power ⫽ 0.92.
RESULTS
Validation of Portal Vein Denervation
Portal vein denervation was validated via physiological and
immunohistochemical methods. In a previous study, we have
shown that denervation of portal vein results in a blunting of
the counterregulatory response to hypoglycemia during a hyperinsulinemic hypoglycemic clamp (10). Therefore, we have
performed paired hypoglycemic clamps before and after denervation and used the blunted response to hypoglycemia as a
method of in vivo validation of portal vein denervation.
Hypoglycemic hyperinsulinemic clamps. Insulin infusion
during the hypoglycemic clamp resulted in an increase in
plasma insulin to 2,967 ⫾ 304 (PRE-DEN) and 3,083 ⫾ 242
pmol/l (POST-DEN), and we measured a slow decrease in
plasma glucose to 2.9 ⫾ 0.2 (PRE-DEN) and 3.0 ⫾ 0.2 mmol/l
(POST-DEN). There was a significant, 71% decrease in the
epinephrine response to hypoglycemia (Fig. 1).
Tyrosine hydroxylase immunohistochemistry. After portal
vein tissue was embedded in paraffin and stained for tyrosine
hydroxylase, the staining was quantified using the Ariol scanner. We found a reduction in tyrosine hydroxylase content
from 1.73 ⫾ 0.52 to 0.37 ⫾ 0.02% marker stain (P ⫽ 0.03).
Effect of Portal Vein Denervation on Body Weight and
Fasting Parameters
There were no differences in body weight and fasting parameters (glucose, insulin, glucagon, intact GLP-1, PYY) before and after denervation (Table 1). However, we measured a
significant reduction in total GLP-1.
Effect of Portal Vein Denervation on Oral Glucose
Tolerance
To compare the effect of portal vein denervation on oral
glucose tolerance, we performed OGTTs (without exenatide)
before and after portal denervation.
After portal denervation, plasma glucose levels were significantly
higher compared with after denervation (P ⫽ 0.02) (Fig. 2A). To
explore the possible mechanism of this difference in oral
glucose tolerance, we measured gastric emptying via acetaminophen concentration, pancreatic hormones (insulin and glucagon) and gut hormones GLP-1 (total and intact), and PYY
during the OGTT. Additionally, we performed IVGTTs to
assess insulin sensitivity, insulin secretion, and insulin clearance.
Table 1. Body weight and fasted-state plasma variables
before and after portal vein denervation
Body weight
Fasted-state variable
Glucose, mmol/l
Insulin, pmol/l
Glucagon, ng/l
Lactate
Intact GLP-1, pmol/l
Total GLP-1, pmol/l
PYY
PRE-DEN
POST-DEN
P Value
28.3 ⫾ 0.9
28.2 ⫾ 0.9
0.88
5.3 ⫾ 0.1
53 ⫾ 8
82 ⫾ 5
0.83 ⫾ 0.06
6⫾0
12 ⫾ 1
161 ⫾ 16
5.2 ⫾ 0.1
49 ⫾ 6
68 ⫾ 6
0.81 ⫾ 0.07
6⫾0
9 ⫾ 1**
151 ⫾ 17
0.22
0.61
0.08
0.75
0.70
0.002
0.48
Values are means ⫾ SD. PRE-DEN, before portal vein denervation; POST-DEN,
after portal vein denervation; GLP-1, glucagon-like peptide-1; PYY, peptide YY.
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00244.2014 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (068.004.044.057) on November 11, 2020.
E647
PORTAL VEIN DENERVATION AND EXENATIDE’S EFFECT
A
Glucose AUC (mmol/l x min)
100
Glucose (mmol/l)
7.5
pre
7.0
post
6.5
6.0
5.5
5.0
4.5
4.0
-30
0
30
60
90
120
150
180
time (min)
40
20
pre
post
Insulin delta AUC (pM x min)
17500
pre
400
post
300
insulin (pM)
60
0
B
*
80
200
100
0
-30
0
30
60
90
120
150
15000
10000
7500
5000
2500
0
Plasma insulin was significantly higher after denervation (P ⫽
0.01; Fig. 2B), and there was a highly significant correlation
between glucose and insulin levels (P ⬍ 0.0001). Moreover,
there was a significant difference in the slope of the glucose/
insulin regression line before (65.92 ⫾ 6.83) and after denervation (92. 14 ⫾ 8.29, P ⬍ 0.05) (Fig. 3).
There was significantly higher rise to peak in acetaminophen
concentration before (62 ⫾ 8 min) than after denervation (36 ⫾ 6
min, P ⫽ 0.01), although the total 0 –180 AUC was not signifi-
insulin (pM)
pre DEN
post DEN
p<0.05
600
500
400
300
200
100
0
1
2
3
4
5
6
7
8
9
**
12500
180
time (min)
700
Fig. 2. Effect of portal denervation on oral
glucose tolerance (n ⫽ 12). Left: glucose (A)
and insulin (B) plasma levels during oral glucose tolerance tests before (pre; Œ) or after
(post; ) portal vein denervation. Right: area
under the curve (AUC) 0 –180 for the timelines at left.
10
glucose (mM)
Fig. 3. Relationship between plasma glucose and insulin before (black lines and Œ)
and after portal vein denervation (gray lines and gray circles). The correlation
between glucose and insulin was highly significant (P ⬍ 0.0001) both before
(slope: 65.93 ⫾ 6.833, r ⫽ 0.72) and after denervation (slope: 92.14 ⫾ 8.291, r ⫽
0.77). The slope difference was statistically significant; P ⬍ 0.05.
pre
post
cantly different (AUC0 –180 298 ⫾ 102 ␮g·ml⫺1·min before vs.
282 ⫾ 54 ␮g·ml⫺1·min, P ⫽ 0.81) (Fig. 4A). No statistically
significant difference in intact GLP-1 total GLP-1 or PYY before
and after denervation was detected (Fig. 4, B–D).
Effect of Portal Vein Denervation on Insulin Sensitivity, Insulin
Secretion, and Insulin Clearance Measured by IVGTT
Hepatic portal vein denervation did not result in any changes
in insulin sensitivity (SI: 4.53 ⫾ 0.83 ␮U·ml⫺1·min⫺1 PREDEN vs. 4.23 ⫾ 0.79 ␮U·ml⫺1·min⫺1 POST-DEN, P ⫽ 0.69),
insulin secretion (AIRg 438 ⫾ 41 ␮U·ml⫺1·min⫺1 PRE-DEN
vs. 459 ⫾ 44 ␮U·ml⫺1·min⫺1 POST-DEN, P ⫽ 0.42), disposition index (1,965 ⫾ 461 PRE-DEN vs. 1,881 ⫾ 341 POSTDEN, P ⫽ 0.84), intravenous glucose tolerance Kg (3.24 ⫾
0.41 min PRE-DEN vs. 2.74 ⫾ 0.51 min POST-DEN, P ⫽
0.29), Sg (3.89 ⫾ 0.41 min⫺1/10⫺2 PRE-DEN vs. 3.82 ⫾ 0.55
min⫺1/10⫺2 POST-DEN, P ⫽ 0.89), or insulin clearance:
(MCRinsulin; 27.44 ⫾ 3.68 ml/kg PRE-DEN vs. 26.85 ⫾ 3.32
ml/kg POST-DEN, P ⫽ 0.81) (Fig. 5).
Effect of Portal Vein Denervation on Exenatide’s Action to
Improve Glucose Homeostasis
Before denervation surgery, administration of an oral glucose challenge resulted in significant increases in plasma glucose (P ⬍ 0.05) and insulin (P ⫽ 0.004) compared with basal
(Fig. 6 A and B).
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00244.2014 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (068.004.044.057) on November 11, 2020.
E648
PORTAL VEIN DENERVATION AND EXENATIDE’S EFFECT
B
4
pre DEN
post DEN
300
pre DEN
post DEN
250
PYY (pg/ml)
3
2
1
0
-30
200
150
100
0
30
60
90
120
150
50
-30
180
0
30
60
time (min)
D
30
25
40
180
pre DEN
post DEN
30
15
20
10
10
5
0
30
60
As expected, administration of exenatide at the beginning of
the OGTT resulted in significantly improved glucose tolerance,
(P ⫽ 0.009; Fig. 6A). The lower glucose was accompanied by
considerable slowing of gastric emptying (as shown by the
acetaminophen concentration; Fig. 6C), lower insulin (Fig.
90
120
150
0
-30
180
0
30
60
AIRg
600
5
4
3
2
150
DI
5000
4000
500
3000
400
DI
AIRg (µU/ml*min)
700
7
6
120
300
2000
200
1000
100
1
0
0
pre
post
0
pre
Kg
post
pre
Sg
5
6
3
2
1
post
Insulin MCR
50
MCR (ml/kg/min-1)
7
Sg (min-1 x 10-2)
6
4
5
4
3
2
40
30
20
10
1
0
0
pre
post
180
6B), and higher glucagon suppression (OGTT-exenatide
AUC0 –180: ⫺2,744 ⫾ 637 ng·l⫺1·min vs. OGTT-control
AUC0 –180: ⫺667 ⫾ 532 ng·l⫺1·min, P ⫽ 0.04).
Denervation of hepatic portal vein did not result in an
attenuation of exenatide’s effect on glycemia. The glycemic
8
0
90
time (min)
SI
SI (µU/ml-1*min-1)
150
post DEN
time (min)
Kg (min-1)
120
pre DEN
20
0
-30
Fig. 5. Effect of portal denervation on intravenous glucose tolerance test-derived variables: insulin sensitivity (SI), acute insulin
secretion (AIRg), disposition index (DI), glucose effectiveness (Sg), intravenous glucose
tolerance (Kg), and insulin clearance (MCR).
90
time (min)
GLP-1
C
intact GLP-1 (pmol/l)
Fig. 4. Effect of portal denervation on plasma
concentration of acetaminophen (A) and gut
hormones peptide YY (PYY; B), intact glucagon-like peptide-1 (GLP-1; C), and total
GLP-1 (D). Variables are presented before
(Œ) and after portal vein denervation ().
acetaminophen (µg/ml)
A
pre
post
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00244.2014 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (068.004.044.057) on November 11, 2020.
pre
post
E649
PORTAL VEIN DENERVATION AND EXENATIDE’S EFFECT
5.0
4.5
-30
0
30
60
90
120
150
180
time (min)
B
pre
ex pre
400
300
*
200
*
100
0
-30
0
30
60
90
120
150
180
time (min)
1050
1000
***
6.0
5.5
5.0
4.5
-30
pre
0
30
EX pre
60
90
120
150
180
time (min)
E
30000
post
tEx pos
500
*
20000
10000
pre
*
400
*
300
200
100
0
-30
0
EX pre
C
0
30
60
90
120
150
180
time (min)
1200
1150
1100
*
1050
1000
post
EX post
30000
**
20000
10000
0
post
EX post
F
5
pre
ex pre
4
3
**
2
1
0
-30
0
30
60
90
time (min)
120
150
180
300
acetaminophen AUC
(µg/ml x min)
acetaminophen (µg/ml)
6.5
post
tEx pos
200
**
100
0
pre
EX pre
5
400
post
tEx pos
4
3
2
1
0
-30
0
30
60
90
120
150
180
acetaminophen AUC
(µg/ml x min)
Insulin (pmol/l)
500
**
1100
7.0
Glucose AUC (mmol/l x min)
5.5
*
Insulin AUC (pmol/l x min)
6.0
1150
Glucose (mmol/l)
*
6.5
Insulin (pmol/l)
7.0
POST-DEN
7.5
1200
acetaminophen (µg/ml)
Glucose (mmol/l)
pre
ex pre
D
Glucose AUC (mmol/l x min)
PRE-DEN
7.5
Insulin AUC (pmol/l x min)
A
time (min)
300
200
*
100
0
post
EX post
Fig. 6. Exenatide effect on oral glucose tolerance before (PRE-DEN) and after (POST-DEN) portal denervation (n ⫽ 6). Glucose (A and D), insulin (B and E)
and acetaminophen (C and F) during oral glucose tolerance test (OGTT) without (Œ) or with exenatide (). *P ⬍ 0.05; **P ⬍ 0.01.
excursion in the presence of exenatide was similar before and
after surgery (fig. 6D), and there were no significant differences in insulin change, (Fig. 6E), change in acetaminophen
plasma concentration reflecting gastric emptying (fig. 6F), or
change in glucagon suppression (exenatide effect ⌬AUC0 –180:
1,654 ⫾ 200 ngl·l⫺1·min PRE-DEN vs. 1,148 ⫾ 200
ngl·l⫺1·min POST-DEN, P ⫽ 0.41).
Overall, denervation of the portal vein did not impact the
effect of exenatide on glucose tolerance (exenatide effect
⌬AUC0 –180: 48 ⫾ 12 mmol·l⫺1·min PRE-DEN vs. 64 ⫾ 26
mmol·l⫺1·min POST-DEN, P ⫽ 0.67; Fig. 7).
DISCUSSION
The abdominal hepatic portal vein occupies a central locus
in the architecture of the intestinal/hepatic/brain communication pathway. It is richly innervated (27) and receptive to
metabolites (such as glucose and lactate) (5, 17) and hormones
(such as GLP-1) (20). The central location and the ability to
respond to nutrient, hormonal, and neural input has made the
portal vein a true nexus of interorgan communication. However, many studies exploring the role of portal vein (and of the
liver) in glucose homeostasis have been done using intraportal
infusions, which inevitably results in both the liver and the
portal vein being exposed to the stimulus. In a previous study,
we have shown that infusion of glucose and GLP-1 in the
portal vein results in a lowering of systemic glucose, an effect
that 1) is independent of corresponding changes in insulin and
glucagon (11), 2) is not replicated when glucose and GLP-1 are
infused peripherally (12), and 3) is partially blocked by the
intraportal infusion of the GLP-1 receptor antagonist exendin(9 –39) (13). In a subsequent study that used tracer to determine
the site of glucose disappearance, we found that intraportal
presence of glucose and GLP-1 receptor activator exenatide
results in important changes in liver glucose uptake/release,
with a favorable impact on glycemia (28). However, due to the
experimental setup, none of these studies was able to identify
whether the locus of signal initiation is the portal vein, the
liver, or both.
To determine whether the portal vein glucose and GLP-1
sensor activation plays a major role in the exenatide’s role in
glycemic control in the habitual subcutaneous administration,
we performed OGTTs (with or without administration of exenatide) before and after portal vein denervation in a large
animal model, the canine.
Confirming previous results (18), denervation of the portal
vein impaired oral glucose tolerance. However, portal vein
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00244.2014 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (068.004.044.057) on November 11, 2020.
E650
PORTAL VEIN DENERVATION AND EXENATIDE’S EFFECT
100
p=0.67
pre
Exenatide effect (mmol/l x min)
post
75
50
25
0
Fig. 7. No impact of portal vein denervation on exenatide’s hypoglcemic
effect. Exenatide effect before (pre) and after denervation (post) is calculated
as the difference (absolute nos.) between AUC0 –180 glucose during OGTT
without exenatide and AUC0 –180 glucose during OGTT with exenatide.
denervation did not impact exenatide’s effect on glucose homeostasis. Despite unequivocal evidence for portal denervation, and contrary to our hypothesis, we found that the effect of
GLP-1 receptor activation on glucose homeostasis is not adversely affected by disconnecting the portal vein sensors from
their neural connections.
There are several explanations for this finding. One is that
the site of glucose and GLP-1 sensing is not the portal vein but
the liver, despite GLP-1 receptor presence (as protein, but not
mRNA) in the portal vein in rats (25). This would be concordant with the identification of both GLP-1 receptor mRNA and
protein on primary human hepatocytes by Gupta et al. (6). The
effect of GLP-1 on the liver could be direct by binding to the
GLP-1 receptor and activation of pathways involved in glucose
and lipid metabolism. Indeed, it has been shown that GLP-1
activates phosphoinositide 3-kinase/PKB in rat hepatocytes
and stimulates the activity of glycogen synthase a (23). Alternatively, the effect of GLP-1 on the liver could be mediated by
binding to the GLP-1 and glucose receptors on nerve terminals
in hepatic hilum area (as well as in the fibers extending into the
connective tissue and hepatic lobules and ending on liver
parenchymal cells in humans) (14). This in turn would initiate
a neural signal to the brain and from the brain via efferent
fibers to other organs (adipose tissue, muscle, liver itself)
involved in a coordinated response to a glucose challenge.
Another possible explanation for our results is that the
experimental design affected the magnitude of the total GLP-1
receptor activation effect, rendering a portally initiated effect
undetectable. It is known that in physiological situations, a
majority of the GLP-1 effect during a meal is due to the
decrease in gastric emptying, with the insulinotropic effect
playing a less important role (18); the insulin-sensitizing effect
(decrease in glycemia in the absence of corresponding changes
in pancreatic hormones) contributes an even lower proportion
(⬃10 –30%) to the glycemic decrease (28). Thus, the possible
lack of effect of GLP-1 receptor activation with portal dener-
vation might have been obscured by the strong “ileal brake”
effect. Moreover, the presence of exenatide, a longer-lasting
GLP-1 receptor activator, and subcutaneous administration
(rather than directly into the portal vein) could have activated
receptors in areas other than the portal vein and liver, with
stronger effects than the hepatoportal initiated signaling. Unlike hypoglycemia, where a single signal (low plasma glucose)
is detected by glucose regulation systems, an oral glucose load
(or a meal) initiates multiple signals coming from the intestinal
area. These signals (glucose, GLP-1, gastric inhibitory polypeptide, PYY, etc.) are sensed by several areas: gut and
afferent nerves and vasculature, wall of the portal vein, liver,
and other organs in the general circulation (14). It is possible
that the putative effects of exenatide or GLP-1 mediated by
portal sensors were mitigated by compensatory changes such
that portal denervation by comparison had a minimal impact on
the overall response to a glycemic load and oral glucose
tolerance.
One of the limitations of our study is that exenatide was used
only acutely in a single dose. The study does not address the
issue of chronic exenatide administration and does not assess
whether portal denervation could impact the long-term effect
of exenatide. We cannot exclude that portal denervation might
have a detrimental effect on exenatide’s glucoregulatory function. Given that chronic exenatide has been shown to decrease
liver steatosis (4), improve ␤-cell function (3), and decrease
body weight in type 2 diabetic patients (16), it might be of
interest to investigate in future studies whether portal GLP-1
receptors are involved in mediating these changes.
Interestingly, portal denervation appeared to worsen oral
glucose tolerance, as indicated by higher glycemia and higher
insulinemia after denervation. This finding is concordant with
data from other investigators. Pagliassotti et al. (21) showed
that hepatic denervation modified the disposition of an enteral
glucose load; in their study, not only was the portal vein
denervated but also the liver, the hepatic artery, and the bile
duct near the hilum (in contrast, in our model portal vein alone
was denervated, distal from the hepatic hilum). The investigators found a 29% increase in glucose incremental area and a
22% increase in insulin incremental area in the denervated
animals. Gut hormones were not measured. In our study, the
incremental increase in glucose was ⬃50% and in insulin
⬃59%. This suggests that, in fact, the worsening of glucose
tolerance in the study of Pagliassotti et al. (21) was due entirely
to portal vein denervation.
Higher plasma glucose and insulin postdenervation could
have occurred through several mechanisms: changes in glucose
absorption due to modifications in gastrointestinal motility
with increased gastric emptying, leading to hyperglycemia and
consequent hyperinsulinemia; hepatic and/or peripheral insulin
resistance, resulting in higher levels of glucose and insulin; a
combination of insulin resistance and compensatory (but inadequate) insulin secretion due to effects on the portohepaticinsular axis; and changes in hepatic glucose uptake due to
effects on the portohepatic-brain axis. To explore some of
these mechanisms, we quantified acetaminophen absorption
(an indicator of gastric emptying) and gut hormones. In addition, we tested the correlation between glucose and insulin
levels, and we investigated whether there was a significant
difference in the glucose/insulin response slope, a possible
indicator of changes in ␤-cell responsiveness to insulin. For
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00244.2014 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (068.004.044.057) on November 11, 2020.
PORTAL VEIN DENERVATION AND EXENATIDE’S EFFECT
both before and after denervation, there was a highly significant correlation between the glucose and insulin levels (P ⬍
0.0001). Moreover, there was a statistically significant difference in the slope of the glucose/insulin response, suggesting
that the higher glycemia drives the insulin response and that a
mechanism originating from the portohepatic/gut area contributes to higher glycemic levels. This hypothesis is supported by
the finding that the higher glycemia postdenervation was associated with accelerated gastric emptying. Peak plasma acetaminophen was faster after denervation. The total acetaminophen AUC was not significantly different, but values after 60
min tend to reflect clearance as well as appearance. We can
speculate that portal denervation interrupted a positive feedback loop related to enhancement of GLP-1 secretion and that
reduction on GLP-1 secretion/action might have resulted in
accelerated stomach emptying (19) and a corresponding increase in plasma glucose, followed by a corresponding increase
in insulin secretion. Postdenervation total GLP-1 was significantly decreased in the fasting state but, it was not significantly
decreased during the oral glucose tolerance test. However, the
study was not powered for this outcome, and it is possible that
a bigger sample size would settle this question. Clearly, further
studies are required to explore this possibility.
To test whether the apparent worsening in oral glucose
tolerance postdenervation was due to changes in insulin sensitivity or insulin secretion, we performed IVGTTs, which
have a better ability to detect changes in insulin sensitivity or
secretion, without interference from factors related to the gut
(gastric emptying, absorption, intestinal hormones). We found
no changes in insulin sensitivity or secretion or in liver insulin
clearance. However, it is worth mentioning that IVGTT-derived insulin sensitivity reflects whole body insulin sensitivity,
and not necessarily hepatic glucose uptake/release, which constitutes only a small part of it. Thus, the current working
hypothesis is that the worsening of glycemia with denervation
might be due to changes in gastric emptying or possibly to
glucose/uptake release from the liver, phenomena that are
harder to detect in the current experimental conditions.
In conclusion, we confirmed that hepatic portal vein denervation impairs glucose tolerance. However, under our current
experimental conditions, hepatic portal vein denervation did
not impact exenatide’s effect on glycemia. It is possible that
the receptors for glucose and GLP-1 reside in the liver or in the
intestine; alternatively, the pharmacological effect of exenatide
on peripheral receptors might have obscured a portal effect of
relatively small magnitude. Portal vein denervation appeared to
negatively impact oral glucose tolerance, although the exact
mechanism for this effect remains unknown.
ACKNOWLEDGMENTS
This study was previously presented in abstract form at the American
Diabetes Association meeting, June 21–25 2013, Chicago, IL (1989-P).
GRANTS
This work was supported by National Institute of Diabetes and Digestive
and Kidney Diseases Grants R37-DK-027619 and 5R01-DK-029867 to R. N.
Bergman.
DISCLOSURES
The authors disclose no conflicts of interest. Dr. Viorica Ionut is the
guarantor of this work and, as such, had full access to all the data in the study
E651
and takes responsibility for the integrity of the data and the accuracy of the data
analysis.
AUTHOR CONTRIBUTIONS
V.I. and R.N.B. conception and design of research; V.I., A.V.B.C., O.O.W.,
D.S., M.S.I., J.L.B., M.B., R.E., H.M., and I.A.B. performed experiments; V.I.
and D.S. analyzed data; V.I., A.V.B.C., C.M.K., and R.N.B. interpreted results
of experiments; V.I. prepared figures; V.I. drafted manuscript; V.I., O.O.W.,
and R.N.B. edited and revised manuscript; V.I. and R.N.B. approved final
version of manuscript.
REFERENCES
1. Bergman RN, Ader M, Huecking K, Van Citters G. Accurate assessment of beta-cell function: the hyperbolic correction. Diabetes 51, Suppl
1: S212–S220, 2002.
2. Boston RC, Stefanovski D, Moate PJ, Sumner AE, Watanabe RM,
Bergman RN. MINMOD Millennium: a computer program to calculate
glucose effectiveness and insulin sensitivity from the frequently sampled
intravenous glucose tolerance test. Diabetes Technol Ther 5: 1003–1015,
2003.
3. Bunck MC, Diamant M, Cornér A, Eliasson B, Malloy JL, Shaginian
RM, Deng W, Kendall DM, Taskinen MR, Smith U, Yki-Järvinen H,
Heine RJ. One-year treatment with exenatide improves beta-cell function,
compared with insulin glargine, in metformin-treated type 2 diabetic
patients: a randomized, controlled trial. Diabetes Care 32: 762–768, 2009.
4. Cuthbertson DJ, Irwin A, Gardner CJ, Daousi C, Purewal T, Furlong
N, Goenka N, Thomas EL, Adams VL, Pushpakom SP, Pirmohamed
M, Kemp GJ. Improved glycaemia correlates with liver fat reduction in
obese, type 2 diabetes, patients given glucagon-like peptide-1 (GLP-1)
receptor agonists. PLoS One 7: e50117, 2012.
5. Donovan CM, Halter JB, Bergman RN. Importance of hepatic glucoreceptors in sympathoadrenal response to hypoglycemia. Diabetes 40:
155–158, 1991.
6. Gupta NA, Mells J, Dunham RM, Grakoui A, Handy J, Saxena NK,
Anania FA. Glucagon-like peptide-1 receptor is present on human hepatocytes and has a direct role in decreasing hepatic steatosis in vitro by
modulating elements of the insulin signaling pathway. Hepatology 51:
1584 –1592, 2010.
7. Hansen L, Deacon CF, Orskov C, Holst JJ. Glucagon-like peptide-1(7–36)amide is transformed to glucagon-like peptide-1-(9 –36)amide by
dipeptidyl peptidase IV in the capillaries supplying the L cells of the
porcine intestine. Endocrinology 140: 5356 –5363, 1999.
8. Hevener AL, Bergman RN, Donovan CM. Hypoglycemic detection
does not occur in the hepatic artery or liver: findings consistent with a
portal vein glucosensor locus. Diabetes 50: 399 –403, 2001.
9. Hevener AL, Bergman RN, Donovan CM. Portal vein afferents are
critical for the sympathoadrenal response to hypoglycemia. Diabetes 49:
8 –12, 2000.
10. Ionut V, Castro AV, Woolcott OO, Stefanovski D, Iyer MS, Broussard
JL, Mkrtchyan H, Burch M, Elazary R, Kirkman E, Bergman RN.
Essentiality of portal vein receptors in hypoglycemic counterregulation:
direct proof via denervation in male canines. Endocrinology 155: 1247–
1254, 2014.
11. Ionut V, Hucking K, Liberty IF, Bergman RN. Synergistic effect of
portal glucose and glucagon-like peptide-1 to lower systemic glucose and
stimulate counter-regulatory hormones. Diabetologia 48: 967–975, 2005.
12. Ionut V, Liberty IF, Hucking K, Lottati M, Stefanovski D, Zheng D,
Bergman RN. Exogenously imposed postprandial-like rises in systemic
glucose and GLP-1 do not produce an incretin effect, suggesting an
indirect mechanism of GLP-1 action. Am J Physiol Endocrinol Metab 291:
E779 –E785, 2006.
13. Ionut V, Zheng D, Stefanovski D, Bergman RN. Exenatide can reduce
glucose independent of islet hormones or gastric emptying. Am J Physiol
Endocrinol Metab 295: E269 –E277, 2008.
14. Jensen KJ, Alpini G, Glaser S. Hepatic nervous system and neurobiology of the liver. Compr Physiol 3: 655–665, 2013.
15. Johnson KM, Edgerton DS, Rodewald T, Scott M, Farmer B, Neal D,
Cherrington AD. Intraportal GLP-1 infusion increases nonhepatic glucose utilization without changing pancreatic hormone levels. Am J Physiol
Endocrinol Metab 293: E1085–E1091, 2007.
16. Klonoff DC, Buse JB, Nielsen LL, Guan X, Bowlus CL, Holcombe JH,
Wintle ME, Maggs DG. Exenatide effects on diabetes, obesity, cardiovascular risk factors and hepatic biomarkers in patients with type 2
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00244.2014 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (068.004.044.057) on November 11, 2020.
E652
17.
18.
19.
20.
21.
22.
PORTAL VEIN DENERVATION AND EXENATIDE’S EFFECT
diabetes treated for at least 3 years. Curr Med Res Opin 24: 275–286,
2008.
Matveyenko AV, Donovan CM. Metabolic sensors mediate hypoglycemic detection at the portal vein. Diabetes 55: 1276 –1282, 2006.
Nauck MA. Unraveling the science of incretin biology. Eur J Intern Med
20, Suppl 2: S303–S308, 2009.
Nauck MA, Niedereichholz U, Ettler R, Holst JJ, Orskov C, Ritzel R,
Schmiegel WH. Glucagon-like peptide 1 inhibition of gastric emptying
outweighs its insulinotropic effects in healthy humans. Am J Physiol
Endocrinol Metab 273: E981–E988, 1997.
Nishizawa M, Nakabayashi H, Uchida K, Nakagawa A, Niijima A. The
hepatic vagal nerve is receptive to incretin hormone glucagon-like peptide-1, but not to glucose-dependent insulinotropic polypeptide, in the
portal vein. J Auton Nerv Syst 61: 149 –154, 1996.
Pagliassotti MJ, Tarumi C, Neal DW, Cherrington AD. Hepatic denervation alters the disposition of an enteral glucose load in conscious
dogs. J Nutr 121: 1255–1261, 1991.
Polonsky KS, Pugh W, Jaspan JB, Cohen DM, Karrison T, Tager HS,
Rubenstein AH. C-peptide and insulin secretion. Relationship between
peripheral concentrations of C-peptide and insulin and their secretion rates
in the dog. J Clin Invest 74: 1821–1829, 1984.
23. Redondo A, Trigo MV, Acitores A, Valverde I, Villanueva-Peñacarrillo ML. Cell signalling of the GLP-1 action in rat liver. Mol Cell
Endocrinol 204: 43–50, 2003.
24. Sanaka M, Kuyama Y, Yamanaka M. Guide for judicious use of the
paracetamol absorption technique in a study of gastric emptying rate of
liquids. J Gastroenterol 33: 785–791, 1998.
25. Vahl TP, Tauchi M, Durler TS, Elfers EE, Fernandes TM, Bitner RD,
Ellis KS, Woods SC, Seeley RJ, Herman JP, D’Alessio DA. Glucagonlike peptide-1 (GLP-1) receptors expressed on nerve terminals in the portal
vein mediate the effects of endogenous GLP-1 on glucose tolerance in rats.
Endocrinology 148: 4965–4973, 2007.
26. Watanabe RM, Bergman RN. Accurate measurement of endogenous
insulin secretion does not require separate assessment of C-peptide kinetics. Diabetes 49: 373–382, 2000.
27. Yi CX, la Fleur SE, Fliers E, Kalsbeek A. The role of the autonomic
nervous liver innervation in the control of energy metabolism. Biochim
Biophys Acta 1802: 416 –431, 2010.
28. Zheng D, Ionut V, Mooradian V, Stefanovski D, Bergman RN. Exenatide sensitizes insulin-mediated whole-body glucose disposal and promotes uptake of exogenous glucose by the liver. Diabetes 58: 352–359,
2009.
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00244.2014 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (068.004.044.057) on November 11, 2020.