Am J Physiol Endocrinol Metab 304: E1089–E1096, 2013.
First published March 26, 2013; doi:10.1152/ajpendo.00614.2012.
The autonomic nervous system regulates postprandial hepatic lipid
metabolism
Eveline Bruinstroop,1 Susanne E. la Fleur,1 Mariette T. Ackermans,2 Ewout Foppen,1,3 Joke Wortel,3
Sander Kooijman,4 Jimmy F. P. Berbée,4 Patrick C. N. Rensen,4 Eric Fliers,1 and Andries Kalsbeek1,3
1
Department of Endocrinology and Metabolism, Academic Medical Center (AMC), University of Amsterdam, Amsterdam, The
Netherlands; 2Department of Clinical Chemistry, Laboratory of Endocrinology, AMC, University of Amsterdam, Amsterdam,
The Netherlands; 3Department of Hypothalamic Integration Mechanisms, Netherlands Institute for Neuroscience, an Institute
of the Royal Netherlands Academy of Arts and Sciences, Amsterdam, The Netherlands; and 4Department of Endocrinology
and Metabolic Diseases and Einthoven Laboratory for Experimental Vascular Medicine, Leiden University Medical Center,
Leiden, The Netherlands
Bruinstroop E, la Fleur SE, Ackermans MT, Foppen E, Wortel
J, Kooijman S, Berbée JF, Rensen PC, Fliers E, Kalsbeek A. The
autonomic nervous system regulates postprandial hepatic lipid metabolism. Am J Physiol Endocrinol Metab 304: E1089 –E1096, 2013.
First published March 26, 2013; doi:10.1152/ajpendo.00614.2012.—
The liver is a key organ in controlling glucose and lipid metabolism
during feeding and fasting. In addition to hormones and nutrients,
inputs from the autonomic nervous system are also involved in
fine-tuning hepatic metabolic regulation. Previously, we have shown
in rats that during fasting an intact sympathetic innervation of the liver
is essential to maintain the secretion of triglycerides by the liver. In
the current study, we hypothesized that in the postprandial condition
the parasympathetic input to the liver inhibits hepatic VLDL-TG
secretion. To test our hypothesis, we determined the effect of selective
surgical hepatic denervations on triglyceride metabolism after a meal
in male Wistar rats. We report that postprandial plasma triglyceride
concentrations were significantly elevated in parasympathetically denervated rats compared with control rats (P ⫽ 0.008), and VLDL-TG
production tended to be increased (P ⫽ 0.066). Sympathetically
denervated rats also showed a small rise in postprandial triglyceride
concentrations (P ⫽ 0.045). On the other hand, in rats fed on a
six-meals-a-day schedule for several weeks, a parasympathetic denervation resulted in ⬎70% higher plasma triglycerides during the day
(P ⫽ 0.001), whereas a sympathetic denervation had no effect. Our
results show that abolishing the parasympathetic input to the liver
results in increased plasma triglyceride levels during postprandial
conditions.
parasympathetic; denervation; liver; triglycerides; feeding
THE LIVER IS ESSENTIAL for maintaining metabolic equilibrium in
the body by storing and releasing large quantities of energy
according to changing demands. In the postprandial state the
liver plays a central role in storing energy for later use, whereas
in the fasted state the liver increases fuel availability for
immediate use. These processes are controlled by circulating
hormones, such as insulin, and by the availability of different
nutrients. It has long been known that the autonomic nervous
input to the liver can also alter hepatic function, especially
glucose metabolism (27, 28). Next to being a key player in
glucose metabolism, the liver is also important for the regula-
Address for reprint requests and other correspondence: E. Bruinstroop,
Academic Medical Center, Univ. of Amsterdam, Dept. of Endocrinology and
Metabolism, F5-169, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands
(e-mail: [email protected]).
http://www.ajpendo.org
tion of lipid metabolism, as it synthesizes and degrades lipids.
Recently, various investigators have revealed a role for the
autonomic nervous system in the regulation of hepatic lipid
metabolism as well (6, 18, 32).
The autonomic nervous system consists of a sympathetic and
a parasympathetic branch. Tracing studies have shown that
both branches are anatomically connected to the liver and
originate from distinct neuronal populations within the hypothalamus (7, 17, 35). The two autonomic branches are hypothesized to have opposing or complimentary effects on target
organs, which has indeed been shown for hepatic glucose
metabolism. Studies involving electrical stimulation of the
hepatic nerves in vivo or pharmacological interventions in
perfused livers show that sympathetic hepatic activation increases hepatic glucose output by stimulating glycogen breakdown, whereas parasympathetic hepatic activation decreases
blood glucose levels by promoting glycogen synthesis (25).
Nevertheless, the role of hepatic autonomic nerves under
physiological conditions has often been debated since no gross
abnormality of metabolic homeostasis was observed after interruption of the nervous supply to the liver. However, effects
of a selective parasympathetic denervation on hepatic glucose
production, glycogen synthesis, and hepatic insulin sensitivity
were reported, pointing to a possible fast route to fine-tune
metabolism in addition to the direct regulation by nutrients and
hormones (19, 22, 37, 38). Interestingly, the effects of parasympathetic hepatic denervation were most obvious in the
postprandial state, indicating that feeding activates the parasympathetic nervous system. This is in line with earlier studies
from Niijima (23), who showed that, after intravenous administration of D-glucose at levels observed postprandially, the
firing rate in efferent vagal nerves innervating the liver was
increased.
Previously, we showed that the sympathetic nervous input to
the liver is necessary to maintain VLDL-triglyceride (TG)
secretion during fasting, whereas we found no evidence for the
involvement of the parasympathetic nervous input (6). We
hypothesized that the parasympathetic nervous system is important in the postprandial state to attenuate VLDL-TG secretion and store TG in the liver. Therefore, we determined the
effect of selective hepatic denervations on TG metabolism
acutely after a meal, in a basal and VLDL-TG secretion
experiment, and chronically in animals adapted to a six-mealsa-day feeding schedule.
0193-1849/13 Copyright © 2013 the American Physiological Society
E1089
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.1 on June 18, 2017
Submitted 5 December 2012; accepted in final form 22 March 2013
E1090
ANS CONTROL OF POSTPRANDIAL HEPATIC LIPID METABOLISM
EXPERIMENTAL PROCEDURES
RESULTS
Acute meal experiments. In the acute meal experiments, all
groups received a meal after an overnight fast. In the basal
experiment, where we looked at the effects of the meal on
plasma TG concentrations in the different denervation groups,
all animals consumed on average 3.55 ⫾ 0.14 g of chow
(means ⫾ SE), with no differences between the groups (P ⫽
0.765) (Table 1). In the VLDL-TG secretion experiment, where
we used the same experimental setup and in addition injected
tyloxapol after the meal, all animals consumed on average 3.85 ⫾
0.15 g of chow (means ⫾ SE), with no significant differences
between the groups (P ⫽ 0.912) (Table 1). Table 1 shows the
absolute concentrations of TG and glucose in the fasted sample
taken directly before the meal and the baseline postprandial
sample (t ⫽ 0) collected directly after the meal of the basal and
VLDL-TG secretion experiment. In both experiments, consumption of the meal significantly decreased plasma TG concentrations and significantly increased blood glucose concentrations compared with fasted levels in all groups, but no
significant differences between groups were observed. In the
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00614.2012 • www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.1 on June 18, 2017
Animals. Male Wistar rats weighing 280 –310 g (Charles River
Breeding Laboratories, Sulzfeld, Germany, and Harlan Nederland,
Horst, The Netherlands) were housed in individual cages with a
12:12-h light-dark schedule (lights on at 7 AM). Chow (Teklad Global
18% protein rodent diet; Harlan) and water were available ad libitum,
unless stated otherwise. All procedures were approved by the Animal
Care Committee of the Royal Netherlands Academy of Arts and
Sciences.
Surgery. All surgeries were performed under anaesthesia with 0.8
ml/kg im Hypnorm (Janssen, High Wycombe, Buckinghamshire, UK)
and 0.4 ml/kg sc Dormicum (Roche, Almere, The Netherlands). After
1 wk in the facility, rats underwent denervation surgery according to
previous reports (14). In short, a laparotomy was performed in the
midline in all groups. For the hepatic sympathectomy (Sx), nerve
bundles running along the hepatic artery proper were removed using
microsurgical instruments. Any connective tissue attachments between the hepatic artery and portal vein were also dissected, eliminating any possible nerve crossings. For the hepatic parasympathectomy (Px) the neural tissue was transected between the ventral vagus
trunk and liver. A total liver denervation (Tx) was achieved by cutting
the sympathetic and parasympathetic branches to the liver. Rats with
sham denervation surgery, as described above except for cutting the
nerve, served as the control group. The effectiveness of the hepatic
sympathetic denervation was checked by measurement of hepatic
noradrenaline concentration using an in-house HPLC method with
fluorescent detection (15). A sympathetically or totally denervated rat
was included in the analysis only if noradrenaline concentration of the
liver was ⬍10% of the sham denervated noradrenaline concentration.
We have previously validated our method for selective hepatic parasympathectomy by using retrograde viral tracing (14). During the
same surgical procedure, an intra-arterial silicone catheter was implanted in the jugular vein according to the method of Steffens (31).
Basal experiment. After surgery, rats were allowed to recover to
presurgery body weight for ⱖ10 days. On the day before the experiment, the food was removed at 5 PM, and the rats were attached to
a metal collar for adaptation. The next day, at 9 AM, the rats were
connected to the blood sampling line, which was attached to the metal
collar and kept out of reach of the animals by means of a counterbalanced beam. This allowed all blood sampling to be performed outside
the cages without further handling of the animals. At 1 PM a
20-h-fasted blood sample was taken before the meal, and subsequently
8 –10 g of laboratory chow was presented. During a time frame of 20
min, the rats were allowed to eat. The remaining food was removed
for the further duration of the experiment and weighed. Immediately
after the meal, a baseline postprandial sample was taken (t ⫽ 0). After
this baseline sample, four more blood samples were taken 20 min
apart to create a postprandial curve. Out of the total of 34 rats tested,
eight rats that did not eat between 2.0 and 4.5 g during the 20-min
meal were excluded from the final analysis of this experiment.
VLDL-TG secretion experiment. After the basal experiment, the
rats were allowed to recover for 3–5 days before the experiment was
repeated to measure VLDL-TG secretion. VLDL-TG secretion was
measured by blocking VLDL-TG clearance through an intravenous
bolus of 0.7 ml of 15% tyloxapol (Sigma-Aldrich) directly after the
baseline postprandial sample (t ⫽ 0), which immediately inhibits
lipoprotein lipase-mediated TG hydrolysis. After this, blood samples
were taken once every 20 min to determine the slope in the rise of
plasma TG levels, indicating VLDL-TG secretion.
Six-meals-a-day feeding schedule. Rats were entrained to a feeding
schedule of six 10-min meals spaced equally over the light-dark cycle.
Food pellets were available in metal food hoppers at Zeitgeber time
(ZT)2, ZT6, ZT10, ZT14, ZT18, and ZT22 [ZT0 is defined as the
“lights-on” time point (7 AM)]. Access to the food could be prevented
by a sliding door situated in front of the food hopper activated by an
electrical motor and controlled by a clock. Water was available ad
libitum. Rats were given 2 wk to adapt to the feeding schedule and
subsequently underwent surgery. A sympathetic or parasympathetic
liver denervation was performed, and a jugular vein catheter was
placed. Rats with an intact hepatic innervation, but on the same
feeding schedule, served as the control group. After 2 wk of recovery,
0.2 ml of blood once every hour for 12 consecutive hours was taken
on two different occasions, starting at either ZT6.5 or ZT18.5. Between both 12-h sampling experiments, the rats were allowed to
recover for ⱖ10 days. All experiments were performed in the rat’s
home cage.
Analysis. Plasma TG levels were assayed using a kit from Roche
(Mannheim, Germany). Blood glucose concentrations were determined during the experiment in blood spots using a glucose meter
(Freestyle, Abbott, The Netherlands). By using a radioimmunoassay
kit, plasma insulin (Linco Research, St. Charles, MO) was measured.
TG concentration in liver was measured after a single-step lipid
extraction with methanol and chloroform (12). The pellets were finally
dissolved in 2% Triton X-100 (Sigma-Aldrich), and TGs were measured using the “Trig/GB” kit (Roche).
RNA isolation and real-time PCR. After the acute experiments, the
left lateral liver lobe was removed after an overdose of pentobarbital
IV. In addition to the livers from the acute meal experiment, livers
from a control experiment were also included to determine which
genes are regulated by the meal. The control experiment contained
four intact rats that continued overnight fasting and four intact rats
that received the meal using the same method as the final experiment.
RNA isolation and real-time PCR were performed as described
previously (6). The expression of the reference gene Hprt was not
regulated by the different conditions in the control experiment (P ⫽
0.751) or by the different denervations (P ⫽ 0.516).
Statistical analysis. Data are presented as means ⫾ SE. When a
curve was plotted for the different groups, a general linear model
(GLM) analysis with repeated measurements was used, with denervation as between-animal factor and time as within-animal factor.
Data were analyzed by one-way ANOVA when single outcome
measurements or separate time points between the groups were
compared. A significant (P ⱕ 0.05) global effect of the repeatedmeasurements GLM or one-way ANOVA was followed by post hoc
tests for individual group differences (Fisher’s protected least significant difference). To investigate the separate effects of time in the
basal experiment, a repeated-measurements analysis was done for all
groups separately, followed by a paired t-test to plot the significance
compared with baseline. Significance was defined at P ⬍ 0.05.
E1091
ANS CONTROL OF POSTPRANDIAL HEPATIC LIPID METABOLISM
Table 1. The effects of the meal in the basal and VLDL-TG secretion experiment
Control
Px
Sx
Tx
3.58 ⫾ 0.24
3.34 ⫾ 0.20
3.57 ⫾ 0.35
3.79 ⫾ 0.33
3.5 ⫾ 0.1
5.6 ⫾ 0.1*
3.8 ⫾ 0.2
5.6 ⫾ 0.2*
3.5 ⫾ 0.2
5.6 ⫾ 0.2*
3.6 ⫾ 0.1
5.8 ⫾ 0.3*
0.33 ⫾ 0.04
0.26 ⫾ 0.03*
0.30 ⫾ 0.03
0.20 ⫾ 0.02*
0.30 ⫾ 0.03
0.20 ⫾ 0.01*
0.32 ⫾ 0.04
0.23 ⫾ 0.03
0.64 ⫾ 0.19
4.97 ⫾ 0.55*
0.39 ⫾ 0.10
5.26 ⫾ 0.65*
0.43 ⫾ 0.07
3.93 ⫾ 0.43*
0.35 ⫾ 0.10
4.04 ⫾ 0.76*
Basal experiment
Food intake, g
Glucose, mmol/l
Fasted
Postprandial
TG, mmol/l
Fasted
Postprandial
Insulin, ng/ml
Fasted
Postprandial
VLDL-TG secretion experiment
3.71 ⫾ 0.16
3.90 ⫾ 0.38
3.92 ⫾ 0.34
4.02 ⫾ 0.53
3.8 ⫾ 0.1
5.8 ⫾ 0.2*
3.8 ⫾ 0.1
5.9 ⫾ 0.1*
3.6 ⫾ 0.2
5.4 ⫾ 0.1*
3.9 ⫾ 0.2
5.7 ⫾ 0.5*
0.28 ⫾ 0.05
0.21 ⫾ 0.03*
0.28 ⫾ 0.04
0.21 ⫾ 0.02
0.30 ⫾ 0.04
0.19 ⫾ 0.05*
0.27 ⫾ 0.05
0.20 ⫾ 0.04
Values are means ⫾ SE. TG, triglyceride; Px, parasympathectomy; Sx, sympathectomy; Tx, total liver denervation. Food intake and fasted and baseline
postprandial (t ⫽ 0) concentrations of blood glucose, plasma TGs, and plasma insulin are presented for both acute meal experiments. No significant differences
between groups were observed in amount of food consumed during the 20-min meal in either experiment. The overall repeated-measures analysis of the 4 groups,
including the effect of the meal and the denervation on plasma concentrations of TG, glucose, and insulin, showed a significant effect of the meal on all 3
parameters (post hoc paired-sample t-test: *P ⬍ 0.05) but no effect of the denervation or interaction effect.
basal experiment, we measured plasma insulin concentrations and
observed a significant increase in all groups after the meal, with no
difference between the groups (Table 1).
Basal experiment. The postprandial curves show the difference in plasma TGs from the baseline postprandial sample
taken directly after the meal (t ⫽ 0) and the subsequent
samples taken at t ⫽ 20, t ⫽ 40, t ⫽ 60, and t ⫽ 80 in the
control, parasympathetically, sympathetically, and total denervated rats (Fig. 1, A–C). After the initial decrease in TG
observed in all groups (Table 1), in the control group the
postprandial curve showed a further significant decline in
plasma TG 20 min after the end of the meal (Fig. 1, A–C).
Subsequently, a slow and steady increase was observed. In
both the parasympathetically (Fig. 1A) and sympathetically
denervated rats (Fig. 1B), we observed a significant postprandial rise in plasma TG levels, reaching significance at t ⫽ 40
and t ⫽ 60 min, respectively. When all postprandial curves
were compared, the curves of both the parasympathetically
Fig. 1. Postprandial triglyceride, glucose, and insulin curves during the basal experiment. A–C: postprandial plasma triglyceride curves of control () compared
with parasympathectomy (Px) (Œ; A), sympathectomy (Sx) (o; B), and total liver denervation rats (Tx) (䊐; C). For clarity, the denervation groups are compared
with the control group in separate graphs, plotted as the delta from the first postprandial sample (t ⫽ 0). The overall repeated-measures analysis of the 4 groups
showed significant effects of time (P ⬍ 0.001) and denervation (P ⫽ 0.046) on plasma triglycerides. The interaction effect of time ⫻ denervation just missed
significance (P ⫽ 0.062). Post hoc analysis revealed a significant difference between the Px and control groups (P ⫽ 0.008; A) as well as between the Sx and
control groups (P ⫽ 0.045; B). In all groups but the Tx group, plasma triglyceride concentrations changed significantly over time (control, P ⫽ 0.019; Px,
P ⬍ 0.001, Sx, P ⫽ 0.020; Tx, P ⫽ 0.254; *P ⬍ 0.05 compared with t ⫽ 0). When comparing the denervation groups at the different time points, only t ⫽
40 min shows a significant difference (1-way ANOVA, P ⫽ 0.043), with a significant difference between the control and Px group (#P ⫽ 0.006). D: postprandial
glucose curves of control (), Px (Œ), Sx (o), and Tx (䊐) rats plotted as the delta from the first postprandial sample (t ⫽ 0). The overall repeated-measures analysis
of the 4 groups showed that glucose remained stable throughout the experiment, with no significant effects of time (P ⫽ 0.353), denervation (P ⫽ 0.858), or
time ⫻ denervation (P ⫽ 0.756). E: postprandial insulin curves of control (), Px (Œ), Sx (o), and Tx (䊐) rats plotted as the delta from the first postprandial
sample (t ⫽ 0). The overall repeated-measurement analysis showed that postprandial insulin decreased significantly over time, with no significant differences
between the groups (time, P ⬍ 0.001; denervation, P ⫽ 0.645; time ⫻ denervation, P ⫽ 0.939). Values are means ⫾ SE.
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00614.2012 • www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.1 on June 18, 2017
Food intake, g
Glucose, mmol/l
Fasted
Postprandial
TG, mmol/l
Fasted
Postprandial
E1092
ANS CONTROL OF POSTPRANDIAL HEPATIC LIPID METABOLISM
Fig. 2. Plasma triglycerides during VLDL-triglyceride secretion experiment. A–C: plasma triglyceride accumulation
after blockade of the triacylglycerol hydrolase activity of
lipoprotein lipase in control () vs. Px (Œ; A), Sx (o; B), and
Tx rats (䊐; C). For clarity, the denervation groups are
compared with the control group in separate graphs. The
overall repeated-measures analysis of the 4 groups showed
a significant effect of time (P ⬍ 0.001) and a significant
interaction effect (time ⫻ denervation, P ⫽ 0.039). No
significant effect of denervation was found (P ⫽ 0.244). In
a post hoc analysis, only the difference between the parasympathetic denervated and the control group approached
significance (P ⫽ 0.066). Values are means ⫾ SE.
food was changed by the denervation. The overall wet weight
of the stomach, including content, was 5.54 ⫾ 0.32 g (means ⫾
SE), with no significant differences between the groups (1-way
ANOVA, P ⫽ 0.715).
Liver analysis acute meal experiments. Noradrenaline concentration in liver tissue was measured to ensure that the
denervation was successful. Two sympathetically denervated
rats were excluded from the analysis for the basal and
VLDL-TG secretion experiment based on a noradrenaline
concentration of ⬎10% of control noradrenaline liver concentration (86 ⫾ 7 ng/g; means ⫾ SE). For all other Sx and Tx
liver samples, noradrenaline values were ⬍10% of control
noradrenaline concentration (Fig. 3A). The TG concentration
of the livers was not different between the groups (Fig. 3B).
We performed RT-PCR analysis on the liver tissue to investigate possible pathways in liver lipid metabolism controlled by
liver innervation. In a separate set of animals, we first investigated the genes that were changed by the meal by comparing
four animals that received a meal in the same experimental
setup as above and four animals that did not receive the meal
but continued the overnight fast. Rats were euthanized ⬃3 h
after the start of the meal. We measured the expression of
acetyl-coenzyme A carboxylase-␣ (Acc1), acetyl-coenzyme A
carboxylase-␤ (Acc2), apolipoprotein B (ApoB), ADP ribosylation factor (Arf-1), carnitine palmitoyltransferase 1␣ (Cpt1a),
fatty acid synthase (Fas), microsomal triglyceride transfer
protein (Mttp), stearoyl-coenzyme A desaturase 1 (Scd1), sterol
regulatory element-binding transcription factor-1c (Srebp1c),
and peroxisome proliferator-activated receptor-␥ (Pparg).
Liver gene expression data showed that the meal increased
mainly expression of genes involved in lipogenesis (Fas, P ⫽
0.019; Srebp1c, P ⫽ 0.021; Pparg, P ⫽ 0.010). Subsequently,
Fig. 3. Liver noradrenaline and liver triglycerides. A: liver
noradrenaline concentrations were measured after the
VLDL-triglyceride secretion experiment. Only rats in the
Sx and Tx groups with liver noradrenaline concentrations
⬍10% of the control value were included in the analysis of
the 2 acute meal experiments displayed in Figs. 1– 4. Two
Sx animals with noradrenaline concentrations of 10 and 18
ng/g were excluded on the basis of this cutoff. B: liver
triglyceride levels were not different between the 4 groups
(1-way ANOVA, P ⫽ 0.887). Values are means ⫾ SE.
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00614.2012 • www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.1 on June 18, 2017
(P ⫽ 0.008; Fig. 1A) and sympathetically denervated rats (P ⫽
0.045; Fig. 1B) were significantly different from that of the
control group. However, the postprandial curve of rats with a
total denervation showed no significant difference compared
with the control group (P ⫽ 0.312; Fig. 1C). We performed gel
electrophoresis on the samples at t ⫽ 0, t ⫽ 40, and t ⫽ 80 min
to separate chylomicrons and VLDL particles and check
whether chylomicrons could contribute to this effect. In the t ⫽
40-min sample, no significant amounts of chylomicrons were
detected. After 80 min a small increase was observed, but this
was only a fraction of the chylomicrons found in an ad
libitum-fed sample. Postprandial blood glucose concentrations
showed no significant increase and were not different between
the four groups (Fig. 1D). Plasma insulin concentrations were
measured at three time points after the meal, starting at t ⫽ 0
min. Although a significant postprandial decrease at t ⫽ 40 and
t ⫽ 80 min was observed, no significant differences between
the groups were observed (Fig. 1E).
VLDL-TG secretion experiment. In a second experiment, the
hepatic VLDL-TG secretion was determined after the meal by
blocking the clearance of TG in peripheral tissues using tyloxapol. In Fig. 2, the absolute TG concentrations from the first
postprandial sample (t ⫽ 0) and the subsequent samples after
injection of tyloxapol are shown. TG accumulation in plasma
differed significantly between groups, as indicated by the
significant interaction effect of time ⫻ denervation (P ⫽ 0.039;
Fig. 2, A–C). Post hoc testing showed that, compared with the
control group, only the increased VLDL-TG secretion in the
parasympathetically denervated rats approached significance
(P ⫽ 0.066; Fig. 2A). At the end of the experiment the rats
were euthanized, and the stomach was removed and weighed,
including its content, to get an indication of whether passage of
ANS CONTROL OF POSTPRANDIAL HEPATIC LIPID METABOLISM
DISCUSSION
This study revealed a physiologically relevant role for the
parasympathetic nervous innervation of the liver in lipid metabolism after a meal. Denervation of the parasympathetic
nerves innervating the liver resulted in higher postprandial
plasma TG responses, indicating that after a meal the parasympathetic nervous system is necessary to inhibit the release of
plasma TGs from the liver. Since this effect is rather small
compared with the initial decrease of plasma TG levels during
the meal (Table 1), we postulate that the autonomic nervous
system has a role in fine-tuning postprandial liver lipid metabolism as opposed to the more robust and direct effects of
hormones and nutrients. However, when challenged chronically with a six-meals-a-day feeding schedule, these subtle
effects became more apparent, and the disinhibitory effect of
the parasympathetic denervation resulted in higher plasma TG
levels during the entire day. These data complement our
previous findings in which we showed that during fasting the
sympathetic nervous system is necessary to stimulate TG
secretion, whereas the parasympathetic nervous system did not
play a role (6). Together, these data indicate that the traditional
concept, i.e., that parasympathetic nerves favor fuel storage
postprandially and sympathetic nerves stimulate fuel availability during fasting, also holds true for hepatic lipid metabolism.
In the current experiments, a sympathetic hepatic denervation resulted in a small but significant increase in plasma TGs
after a meal. We believe that feeding shifts the autonomic
balance toward both increased parasympathetic activity and
decreased sympathetic activity. Because of the opposing
forces, losing either the parasympathetic or sympathetic innervation yields similar results compared with the control condition. Denervation of the sympathetic hepatic nerve will result
in a lesser inhibition of TG secretion after a meal, with a
resulting tendency to increase plasma TG. However, the less
pronounced effects and the absence of an effect in the sixmeals-a-day feeding schedule suggests that there is no major
role for the sympathetic nervous system after feeding. Interestingly, the disinhibitory effects of the sympathetic and parasympathetic denervations were lost in rats with a total dener-
Fig. 4. Plasma triglycerides in the 6-meals-a-day feeding schedule. A: 24-h plasma triglyceride profile of control (), Px (Œ), and Sx rats (o) on a 6-meals-a-day
feeding schedule [Zeitgeber time (ZT) ⫽ 0, i.e., defined as the “lights-on” time point (7 AM)]. The meals are indicated by the dotted lines. There was a significant
difference between the groups (time, P ⬍ 0.001; denervation, P ⫽ 0.002; time ⫻ denervation, P ⬍ 0.001). Post hoc analysis revealed that parasympathetic
denervated rats had higher plasma triglyceride levels over the 12:12-h light-dark cycle compared with control rats (P ⫽ 0.001) and sympathetically denervated
rats (P ⫽ 0.005). All separate time points were analyzed by 1-way ANOVA with post hoc least significant difference between all groups (*P ⬍ 0.05 compared
with control and Sx; #P ⬍ 0.05 compared with control; §P ⬍ 0.05 compared with Sx). B and C: mean triglyceride increments to a meal in control, Px and Sx
rats expressed as the difference between the first sample point after the meal and the last sample point before the meal in the 6-meals-a-day experiment. The results
are shown for all meals (B) and only the meals consumed during the dark phase (C). Only in the dark phase was a significant difference observed between the
groups (1-way ANOVA, P ⫽ 0.033; *P ⬍ 0.05), with a significant difference between the control and Px rats. Values are means ⫾ SE.
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00614.2012 • www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.1 on June 18, 2017
we measured gene expression in the denervation groups after
the meal to determine whether these genes are in part regulated
by innervation, but we found no significant differences in gene
expression between the denervation groups.
Six-meals-a-day feeding schedule. Finally, we investigated
the potential physiological relevance of the autonomic nervous
system for hepatic lipid metabolism in a chronic experiment by
challenging the rats with a six-meals-a-day schedule that consisted of 1 meal every 4 h for ⱖ4 wk. This model ensures that
the rat is maintained in the postprandial state during the entire
24-h cycle. Previous experiments have shown that this model is
very suitable to standardize the meals for all rats over the 24-h
cycle concerning the timing of meals and amount of chow
consumed. All rats readily adapted to the feeding schedule
within a couple of days, as evidenced by a continued increase
in their body weight, and consumed ⬃3.5 g during every single
meal, comparable with the previous experiments. No significant differences in meal size between time points or groups
were detected. After 4 wk, blood samples were taken every
hour for 12 consecutive hours on two different occasions to
measure plasma TG. All groups showed a significant effect
of time on plasma TG during the total 24-h sampling period
(Fig. 4A). However, parasympathetically denervated rats
had significantly higher plasma TG levels during the daynight cycle compared with control (P ⫽ 0.001) and sympathetically denervated rats (P ⫽ 0.005) (Fig. 4A). Twentyfour-hour plasma TG profiles of sympathetically denervated
rats were not significantly different from those of control
rats. The average concentration of TG during the entire 24-h
sampling period was also increased for parasympathetically
denervated rats (Px: 3.46 ⫾ 0.36 mmol/l; Sx: 2.20 ⫾ 0.09
mmol/l; control: 2.01 ⫾ 0.13 mmol/l; means ⫾ SE). Further
analysis of the difference between the premeal samples and the
first sample taken 30 min after the meal revealed again that
sympathetic and parasympathetic denervation tended to increase the meal-induced rise in plasma TG (Fig. 4B). However,
in these non-fasting-induced feeding conditions the effect of
the parasympathetic denervation reached significance only
when meals were consumed during the normal time of feeding,
i.e., the dark phase (Fig. 4C).
E1093
E1094
ANS CONTROL OF POSTPRANDIAL HEPATIC LIPID METABOLISM
VLDL-TG secretion. In our study, using a mixed meal, glucose
levels do not peak but remain stable during the experiment. In
the control rats, we do find that plasma TGs decrease after
feeding. In our denervated rats, a smaller decrease in TGs is
observed, from which we conclude that the autonomic nerves
also act to inhibit TG secretion after a meal. We have sought to
elucidate the mechanism via which the autonomic nerves exert
their fine-tuning on lipid metabolism by measuring VLDL-TG
secretion, liver TG concentrations, and expression of various
genes involved in hepatic lipid metabolism. We indeed found
a trend toward an increase in VLDL-TG secretion in the
parasympathetic denervated group. No effects on liver TG
were observed, although this study could be underpowered to
show the expected small differences when taking into account
the variation within the groups. Although we did observe the
known effects of feeding on the expression of lipogenic genes,
we found no effect of a denervation on the expression level of
these genes. Perhaps the major effects of nutrients and hormones on the liver after a meal mask the more subtle effects of
a denervation. Next to a direct effect of the autonomic nervous
system on lipid metabolism, it is possible that parasympathetic
activity increases insulin sensitivity of the liver. We did not
find an effect of a parasympathetic denervation on plasma
concentrations of glucose or insulin, but previous studies using
several methods to measure insulin sensitivity found decreased
insulin sensitivity after parasympathetic denervation (19, 22,
37, 38). Despite stimulating hepatic fatty acid and TG synthesis, which are substrates for VLDL-TG secretion, acute elevations of insulin are known to inhibit VLDL-TG secretion (20).
By affecting insulin sensitivity, parasympathetic activity could
enhance this insulin-induced inhibition of VLDL-TG secretion,
and a denervation would then cause the opposite effect. Alternatively, the parasympathetic nervous system may regulate
lipid metabolism by other mechanisms, e.g., by altering hepatic
blood flow. Clearly, further studies are necessary to investigate
how the parasympathetic nervous system fine-tunes lipid metabolism after a meal.
Next to the mechanism within the liver, several previous
studies have provided evidence for peripheral signals that
could activate the parasympathetic nervous system innervating
the liver. Van den Hoek et al. (34) showed that the effect of
peripheral hyperinsulinemia on VLDL-TG secretion is prevented by intracerebroventricular (icv) infusion of NPY, suggesting that insulin could possibly act via both a peripheral and
central route to inhibit VLDL-TG secretion. In addition to
insulin, Lam et al. (18) showed that also an icv infusion of
glucose inhibits VLDL-TG secretion. This effect was dependent on an intact vagal nerve, thereby supporting the results
from our experiments; i.e., vagal nervous activity inhibits
VLDL-TG secretion. In addition and in line with our previous
results, they found no effects of a selective parasympathetic
denervation in fasted rats, emphasizing that a meal or icv
glucose should be given to observe an effect of a parasympathetic denervation.
Finally, the results presented here are of interest for human
pathophysiology. Similar to our rats under a six-meals-a-day
schedule, many humans spend the entire day and most of the
night in the postprandial state (36). Despite this, most studies
and clinical guidelines of cardiovascular risk to date have
focused on fasting conditions. Interestingly, a few studies have
demonstrated that the postprandial TG level is a better inde-
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00614.2012 • www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.1 on June 18, 2017
vation of the liver. Previous studies have also indicated that the
sympathetic and parasympathetic nervous system work together to cause an effect on the liver. Gardemann and Jungermann (13) showed that the effects of parasympathetic stimulation on glucose metabolism occurred only during concurrent
␣- and ␤-adrenergic blockade in perfused rat liver. In a previous experiment, we also observed that the effects of a parasympathetic or sympathetic denervation, i.e., a loss of the daily
glucose rhythm, were not observed in animals with a total
denervation (10). Together, these observations lead us to propose that it is not the absence but rather the disbalance of the
autonomic nervous system that induces an abnormal metabolism in the liver. This is in concordance with the observation
that no major metabolic abnormalities occur after liver transplantation. When the liver has no autonomic innervation, it
responds mainly to important signals from the periphery such
as insulin and free fatty acids.
By using a surgical technique that enabled us to selectively
cut the autonomic nerves innervating the liver, we can conclude that the changes observed are most likely due to a hepatic
mechanism. Because the ingested chow also contains 6% fat
(soybean oil), the possibility that the effects observed were
caused by a change in the uptake of TG from the gut as a
consequence of the surgery could not be ruled out. However,
based on the following arguments, we think an indirect effect
via the intestinal system is highly unlikely. First of all, the
uptake of dietary lipids and the subsequent formation of chylomicrons is a time-consuming process compared with glucose
uptake. Based on the literature, we have estimated the contribution of TGs packed in chylomicrons to total plasma TGs to
be negligible within the first 80 min after consumption of chow
(1, 2, 4). With gel electrophoresis we confirmed that after 80
min only a minor fraction of lipids is packed in chylomicrons,
whereas the effects of the denervation were already apparent
after 40 min. However, for the VLDL-TG secretion experiment, we had to take samples over a longer period of time (160
min) to be able to estimate VLDL-TG secretion rate. Therefore, in these experiments we cannot exclude a contribution of
chylomicrons after 80 min. Because the increase of TG remained linear during the experiment, no major increases in
chylomicrons during these 160 min are to be expected. Because of the presence of tyloxapol in the samples, this could
not be confirmed by gel electrophoresis. We have no reason to
believe that the uptake of TGs between the denervation groups
was different. The rats were fasted overnight to ensure that no
chylomicrons were present in plasma at the start of the experiment and that all rats were motivated to eat. There were no
differences in food intake or wet weight of the stomach
between the groups, which was indicative of similar gastric
motility after denervation. Therefore, it is highly unlikely that
our results are based on a differential intestinal uptake of TG
between the groups. Nevertheless, it would be interesting to
confirm the origin of TGs after a meal with tracer methods.
In the fed state, the liver and adipose tissue are set to store
the newly available nutrients. High levels of insulin acutely
inhibit VLDL-TG secretion in humans and human hepatocytes
in culture (8, 20). Bülow et al. (8) have measured VLDL-TG
secretion across the splanchnic bed in humans receiving a
glucose meal and during a hyperinsulinemic euglycemic clamp
and concluded that only the latter inhibits VLDL-TG secretion.
It seems that high concentrations of glucose in turn stimulate
ANS CONTROL OF POSTPRANDIAL HEPATIC LIPID METABOLISM
ACKNOWLEDGMENTS
We thank Mariem Bouali and Marja Neeleman from the Department of
Clinical Chemistry, Laboratory of Endocrinology, Academic Medical Center,
University of Amsterdam, for technical assistance. Part of the data has been
presented as an abstract at ENDO 2012 (OR22-2).
GRANTS
This work is supported by NWO-ZonMw (TOP 91207036). P. C. N. Rensen
is an Established Investigator of the Netherlands Heart Foundation (2007B081,
2009T038).
DISCLOSURES
There are no conflicts of interest, financial or otherwise, for any of the
authors.
AUTHOR CONTRIBUTIONS
E.B., S.E.l.F., E. Fliers, and A.K. contributed to the conception and design
of the research; E.B., S.E.l.F., M.T.A., E. Foppen, J.W., and S.K. performed
the experiments; E.B., S.E.l.F., M.T.A., and S.K. analyzed the data; E.B.,
S.E.l.F., M.T.A., S.K., J.F.B., P.C.N.R., E. Fliers, and A.K. interpreted the
results of the experiments; E.B. prepared the figures; E.B. drafted the manuscript; E.B., S.E.l.F., M.T.A., E. Foppen, J.W., S.K., J.F.B., P.C.N.R., E. Fliers,
and A.K. edited and revised the manuscript; E.B., S.E.l.F., M.T.A., E. Foppen,
J.W., S.K., J.F.B., P.C.N.R., E. Fliers, and A.K. approved the final version of
the manuscript.
REFERENCES
1. Aoki T, Yoshinaka Y, Yamazaki H, Suzuki H, Tamaki T, Sato F,
Kitahara M, Saito Y. Triglyceride-lowering effect of pitavastatin [corrected] in a rat model of postprandial lipemia. Eur J Pharmacol 444:
107–113, 2002.
2. Argiles J, Herrera E. Appearance of circulating and tissue 14C-lipids
after oral 14C-tripalmitate administration in the late pregnant rat. Metabolism 38: 104 –108, 1989.
3. Bansal S, Buring JE, Rifai N, Mora S, Sacks FM, Ridker PM. Fasting
compared with nonfasting triglycerides and risk of cardiovascular events
in women. JAMA 298: 309 –316, 2007.
4. Berbée JF, van der Hoogt CC, Sundararaman D, Havekes LM,
Rensen PC. Severe hypertriglyceridemia in human APOC1 transgenic
mice is caused by apoC-I-induced inhibition of LPL. J Lipid Res 46:
297–306, 2005.
5. Berglund L, Brunzell JD, Goldberg AC, Goldberg IJ, Sacks F, Murad
MH, Stalenhoef AF; Endocrine society. Evaluation and treatment of
hypertriglyceridemia: an Endocrine Society clinical practice guideline. J
Clin Endocrinol Metab 97: 2969 –2989, 2012.
6. Bruinstroop E, Pei L, Ackermans MT, Foppen E, Borgers AJ, Kwakkel J, Alkemade A, Fliers E, Kalsbeek A. Hypothalamic neuropeptide Y
(NPY) controls hepatic VLDL-triglyceride secretion in rats via the sympathetic nervous system. Diabetes 61: 1043–1050, 2012.
7. Buijs RM, la Fleur SE, Wortel J, van Heijningen C, Zuiddam L,
Mettenleiter TC, Kalsbeek A, Nagai K, Niijima A. The suprachiasmatic
nucleus balances sympathetic and parasympathetic output to peripheral
organs through separate preautonomic neurons. J Comp Neurol 464:
36 –48, 2003.
8. Bülow J, Simonsen L, Wiggins D, Humphreys SM, Frayn KN, Powell
D, Gibbons GF. Co-ordination of hepatic and adipose tissue lipid metabolism after oral glucose. J Lipid Res 40: 2034 –2043, 1999.
9. Burneo JG, Faught E, Knowlton R, Morawetz R, Kuzniecky R.
Weight loss associated with vagus nerve stimulation. Neurology 59:
463–464, 2002.
10. Cailotto C, van Heijningen C, van der Vliet J, van der Plasse G,
Habold C, Kalsbeek A, Pevet P, Buijs RM. Daily rhythms in metabolic
liver enzymes and plasma glucose require a balance in the autonomic
output to the liver. Endocrinology 149: 1914 –1925, 2008.
11. Das UN. Vagus nerve stimulation as a strategy to prevent and manage
metabolic syndrome. Med Hypotheses 76: 429 –433, 2011.
12. Folch J, Lees M, Sloane Stanley GH. A simple method for the isolation
and purification of total lipides from animal tissues. J Biol Chem 226:
497–509, 1957.
13. Gardemann A, Jungermann K. Control of glucose balance in the
perfused rat liver by the parasympathetic innervation. Biol Chem Hoppe
Seyler 367: 559 –566, 1986.
14. Kalsbeek A, La Fleur S, van Heijningen C, Buijs RM. Suprachiasmatic
GABAergic inputs to the paraventricular nucleus control plasma glucose
concentrations in the rat via sympathetic innervation of the liver. J
Neurosci 24: 7604 –7613, 2004.
15. Klieverik LP, Sauerwein HP, Ackermans MT, Boelen A, Kalsbeek A,
Fliers E. Effects of thyrotoxicosis and selective hepatic autonomic denervation on hepatic glucose metabolism in rats. Am J Physiol Endocrinol
Metab 294: E513–E520, 2008.
16. Koren MS, Holmes MD. Vagus nerve stimulation does not lead to
significant changes in body weight in patients with epilepsy. Epilepsy
Behav 8: 246 –249, 2006.
17. la Fleur SE, Kalsbeek A, Wortel J, Buijs RM. Polysynaptic neural
pathways between the hypothalamus, including the suprachiasmatic nucleus, and the liver. Brain Res 871: 50 –56, 2000.
18. Lam TK, Gutierrez-Juarez R, Pocai A, Bhanot S, Tso P, Schwartz GJ,
Rossetti L. Brain glucose metabolism controls the hepatic secretion of
triglyceride-rich lipoproteins. Nat Med 13: 171–180, 2007.
19. Lautt WW, Macedo MP, Sadri P, Takayama S, Duarte Ramos F,
Legare DJ. Hepatic parasympathetic (HISS) control of insulin sensitivity
determined by feeding and fasting. Am J Physiol Gastrointest Liver
Physiol 281: G29 –G36, 2001.
20. Lewis GF, Uffelman KD, Szeto LW, Weller B, Steiner G. Interaction
between free fatty acids and insulin in the acute control of very low density
lipoprotein production in humans. J Clin Invest 95: 158 –166, 1995.
21. Licht CM, Vreeburg SA, van Reedt Dortland AK, Giltay EJ,
Hoogendijk WJ, DeRijk RH, Vogelzangs N, Zitman FG, de Geus EJ,
Penninx BW. Increased sympathetic and decreased parasympathetic activity rather than changes in hypothalamic-pituitary-adrenal axis activity is
associated with metabolic abnormalities. J Clin Endocrinol Metab 95:
2458 –2466, 2010.
22. Matsuhisa M, Yamasaki Y, Shiba Y, Nakahara I, Kuroda A, Tomita
T, Iida M, Ikeda M, Kajimoto Y, Kubota M, Hori M. Important role of
the hepatic vagus nerve in glucose uptake and production by the liver.
Metabolism 49: 11–16, 2000.
23. Niijima A. Blood glucose levels modulate efferent activity in the vagal
supply to the rat liver. J Physiol 364: 105–112, 1985.
24. Nordestgaard BG, Benn M, Schnohr P, Tybjaerg-Hansen A. Nonfasting triglycerides and risk of myocardial infarction, ischemic heart disease,
and death in men and women. JAMA 298: 299 –308, 2007.
25. Püschel GP. Control of hepatocyte metabolism by sympathetic and
parasympathetic hepatic nerves. Anat Rec A Discov Mol Cell Evol Biol
280: 854 –867, 2004.
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00614.2012 • www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.1 on June 18, 2017
pendent predictor than the fasting TG level for assessing
present and future cardiovascular disease events (3, 24). Therefore, the need for routine screening tests of postprandial TG in
humans is acknowledged using either nonfasting levels or TGs
measured after an oral TG tolerance test (5, 26). In addition, a
small number of studies have addressed the possible relations
between autonomic nervous activity and the metabolic syndrome in humans. Interestingly, Licht et al. (21) showed that
decreased parasympathetic activity correlates with higher TG
levels in patients with the metabolic syndrome. Although their
study has been performed with cardiac measurements of autonomic activity, and no causal effect could be shown, it is
interesting to note that these results are in accord with the data
found in our study. One may speculate whether stimulating
vagal nerve activity would induce beneficial changes in the
lipid spectrum in these patients. Several animal studies found
associations between decreased body weight gain and vagal
nerve stimulation, although small retrospective studies in humans with a vagal nerve stimulator for the treatment of epilepsy are inconclusive on this matter (9, 16, 29, 33). Larger
prospective studies, including different components of the
metabolic syndrome, are necessary to show that this could
indeed be a novel treatment, as hypothesized previously (11).
E1095
E1096
ANS CONTROL OF POSTPRANDIAL HEPATIC LIPID METABOLISM
33. Val-Laillet D, Biraben A, Randuineau G, Malbert CH. Chronic vagus
nerve stimulation decreased weight gain, food consumption and sweet
craving in adult obese minipigs. Appetite 55: 245–252, 2010.
34. van den Hoek AM, Voshol PJ, Karnekamp BN, Buijs RM, Romijn JA,
Havekes LM, Pijl H. Intracerebroventricular neuropeptide Y infusion
precludes inhibition of glucose and VLDL production by insulin. Diabetes
53: 2529 –2534, 2004.
35. Vinuela MC, Larsen PJ. Identification of NPY-induced c-Fos expression
in hypothalamic neurones projecting to the dorsal vagal complex and the
lower thoracic spinal cord. J Comp Neurol 438: 286 –299, 2001.
36. Williams CM, Moore F, Morgan L, Wright J. Effects of n-3 fatty acids
on postprandial triacylglycerol and hormone concentrations in normal
subjects. Br J Nutr 68: 655–666, 1992.
37. Xie H, Lautt WW. Insulin resistance caused by hepatic cholinergic
interruption and reversed by acetylcholine administration. Am J Physiol
Endocrinol Metab 271: E587–E592, 1996.
38. Xue C, Aspelund G, Sritharan KC, Wang JP, Slezak LA, Andersen
DK. Isolated hepatic cholinergic denervation impairs glucose and glycogen metabolism. J Surg Res 90: 19 –25, 2000.
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00614.2012 • www.ajpendo.org
Downloaded from http://ajpendo.physiology.org/ by 10.220.33.1 on June 18, 2017
26. Ridker PM. Fasting versus nonfasting triglycerides and the prediction of
cardiovascular risk: do we need to revisit the oral triglyceride tolerance
test? Clin Chem 54: 11–13, 2008.
27. Shimazu T, Fukuda A. Increased activities of glycogenolytic enzymes in
liver after splanchnic-nerve stimulation. Science 150: 1607–1608, 1965.
28. Shimazu T, Fujimoto T. Regulation of glycogen metabolism in liver by
the autonomic nervous system. IV. Neural control of glycogen biosynthesis. Biochim Biophys Acta 252: 18 –27, 1971.
29. Sobocki J, Fourtanier G, Estany J, Otal P. Does vagal nerve stimulation
affect body composition and metabolism? Experimental study of a new
potential technique in bariatric surgery. Surgery 139: 209 –216, 2006.
30. Stanley S, Pinto S, Segal J, Perez CA, Viale A, DeFalco J, Cai X,
Heisler LK, Friedman JM. Identification of neuronal subpopulations that
project from hypothalamus to both liver and adipose tissue polysynaptically. Proc Natl Acad Sci USA 107: 7024 –7029, 2010.
31. Steffens AB. A method for frequent sampling blood and continuous
infusion of fluids in the rat without disturbing the animal. Physiol Behav
4: 833–836, 1969.
32. Tavares FL, Seelaender MC. Hepatic denervation impairs the assembly
and secretion of VLDL-TAG. Cell Biochem Funct 26: 557–565, 2008.