Portosystemic collaterals are not prerequisites for the development
of hepatic encephalopathy in cirrhotic rats
I-Fang Hsina, Sun-Sang Wanga,c, Hui-Chun Huanga,c, Fa-Yauh Leea,c, Cho-Yu Chanc,d,*,
Ching-Chih Changb,c, Chia-Yang Hsuc,e, Han-Chieh Lina,c, Shou-Dong Leea,c
aDivision of Gastroenterology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan, ROC
bGeneral Medicine, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan, ROC
cFaculty of Medicine, National Yang-Ming University School of Medicine, Taipei, Taiwan, ROC
dTaipei Medical University Hospital, Taipei, Taiwan, ROC
eNational Yang-Ming University Hospital, Yilan, Taiwan, ROC
Received May 30, 2011; accepted July 26, 2011
Background: Liver functions and portosystemic collaterals influence the development and severity of hepatic encephalopathy (HE) in cirrhosis.
However, it has not been examined which factor has a greater influence or if shunts can be used to determine the presence and severity of HE.
The expression of tumor necrosis factor-a (TNF-a) is increased in cirrhosis, and its role in HE deserves further evaluation.
Methods: Portal hypertension was induced by portal vein ligation (PVL; a model of high-degree portosystemic shunting without significant liver
damage) and liver cirrhosis was induced by bile duct ligation (BDL; a model of low-degree shunting with liver cirrhosis) in male Spraque-
Dawley rats. Sham-operated rats were used as controls. Motor activity counts, hemodynamic parameters, plasma levels, liver biochemistry
parameters, TNF-a, and a flow-pressure curve study of portosystemic collaterals (where a higher slope indicates fewer portosystemic collaterals)
were performed on Day 7 after PVL and Week 5 after BDL.
Results: Portal pressure was significantly higher in the PVL and BDL groups than in controls. The liver biochemistry parameters, TNF-a, and
motor activities were not significantly different between the PVL and PVL-control groups. In the BDL group, TNF-a, AST, and total bilirubin
levels were significantly higher and the motor activity counts were lower than in the BDL-control group. Moreover, in the BDL rats, TNF-
a (p ¼0.037, R ¼ -0.490), AST (p ¼0.007, R ¼ -0.595) and total bilirubin (P¼ 0.001, R ¼?0.692) levels, but not the slopes of the flow-
pressure curves, were significantly and negatively correlated with the motor activity counts.
Conclusion: The presence of a high degree of portosystemic shunting without significant liver damage may not be adequate for the development
Copyright ? 2011 Elsevier Taiwan LLC and the Chinese Medical Association. All rights reserved.
Keywords: hepatic encephalopathy; liver cirrhosis; portal hypertension; portal-systemic collaterals; tumor necrosis factor-a
Hepatic encephalopathy (HE) is a neuropsychiatric disorder
the encompasses a broad spectrum of symptoms that may vary
from subtle mental changes in the early stages to the complete
loss of consciousness (i.e., hepatic coma) in the late stages.1
Numerous substances have been shown to be involved in the
pathogenesis of HE.2Among these toxic substances, the most
widely studied is ammonia, which is mainly synthesized by
bacteria in the intestines and carried by portal venous blood to
the liver for further metabolism.
The presence of portosystemic collaterals is a well-known
feature of portal hypertension.3
a response to increased portal pressure; they divert the portal
tributary blood flow into systemic circulation in order to
Collaterals develop as
* Corresponding author. Dr. Cho-Yu Chan, Taipei Medical University
Hospital, 252, Wu Hsing Street, Taipei 110, Taiwan, ROC.
E-mail address: email@example.com (C.-Y. Chan).
Journal of the Chinese Medical Association 75 (2012) 3e9
1726-4901/$ - see front matter Copyright ? 2011 Elsevier Taiwan LLC and the Chinese Medical Association. All rights reserved.
decompress the pressurized portal system. Previous studies
have reported that increased blood and brain ammonia
concentrations can induce encephalopathy in portacaval-
shunted rats, but not in sham-operated rats.4Moreover,
portacaval-shunted rats have higher concentrations of brain
glutamine, aromatic amino acids, cerebrospinal fluid gluta-
mine, aromatic amino acids, glutamate, and aspartate
compared with sham-operated rats.4These results indicate that
hyperammonemia alone does not induce encephalopathy,
whereas portosystemic collaterals are essential to the patho-
genesis of encephalopathy. In a portal hypertensive state,
ammonia may bypass the liver directly into systemic circula-
tion in the form of collaterals and consequently cause neuro-
behavior changes. Furthermore, it is possible that the higher
the degree of portosystemic shunting, the greater the amount
of ammonia that is shunted directly into systemic circulation.
However, whether the degree of shunting per se determines the
presence and severity of encephalopathy is unknown.
Tumor necrosis factor-a (TNF-a) is a proinflammatory
cytokine that is released into the bloodstream upon the
recognition of bacterial products by the monocyte CD14
receptor and is essential for an adequate host response against
infection.5It has been shown that TNF-a participates in the
development of hyperdynamic circulatory syndrome and
hepatopulmonary syndrome in cirrhotic rats.6We have also
demonstrated that cirrhotic rats with ascites have increased
plasma levels of TNF-a.7Regarding the involvement of TNF-
a in the pathogenesis of HE, cerebral inflammatory responses
have been proposed to be the major cause of bile duct ligation
(BDL; i.e., an animal model of liver cirrhosis)-induced HE.
This is supported by a report that peripheral TNF-a-secreting
monocytes can infiltrate the brain on Day 10 after BDL.8
Although we have previously found that there is no correla-
tion between TNF-a and the severity of HE, as assessed by
motor activity counts and plasma levels of TNF-a in rats with
thioacetamide-induced acute liver failure,9the relationship
between TNF-a and HE in cirrhosis and portal hypertension
deserves further evaluation.
In this study, we used two animal models: partially portal
vein-ligated (PVL; high shunting model) rats,3in which more
than 90% of the portal venous blood is shunted directly into
systemic circulation without significant liver damage, and
BDL (low shunting model)-induced cirrhotic rats,10,11in
which both liver failure and collaterals are present, in order to
investigate the relationship between liver function, the circu-
lating concentration of TNF-a, degree of portosystemic
shunting, and changes in motor activities that occur in portal
hypertensive rats with or without liver injury.
2.1. PVL model
PVL rats were used as described in a previous report.12
Male Sprague-Dawley rats weighing 280e300 g at the time
of surgery were used in these experiments. Ketamine HCl
(100 mg/kg-body weight, intramuscularly administered) was
used as the anesthesia. In brief, the portal vein was isolated
and a 3-0 silk ligature was tied around both the portal vein and
an adjacent 20-gauge, blunt-tipped needle. The needle was
then removed and the vein was allowed to re-expand. A
second, loose ligature was left around the portal vein with
the two endings of the ligature were placed on each side of the
abdominal cavity. The abdomen was then closed and the
animal was allowed to recover. The body weights of rats were
measured on the day of the experiments.
2.2. BDL model
Male Sprague-Dawley rats weighing 240e270 g at the time
of surgery were used for experiments. A common bile duct
(CBD) division and occluded animal model was created, as
described in previous reports.10,11In brief, the rats were
anesthetized with ketamine (100 mg/kg-body weight, intra-
muscularly administered), and then the CBD was exposed and
ligated using two 3-0 silk sutures. The first ligature was made
below the junction of the intrahepatic bile ducts, and the
second ligature was made above the entrance to the pancreatic
ducts. Then, the CBD was catheterized by insertion of a PE-10
catheter, and 10% formalin (100 mL/100g) was slowly injected
into the biliary tree in order to prevent the subsequent dilatation
of the ligated bile ducts.13Finally, the CBD was transected
by the two ligatures. Benzathine benzylpenicillin was post-
administered) to prevent infection. Vitamin K (8 mg/kg-body
weight, intramuscularly administered) was administered after
surgery at weekly intervals. The animals were allowed to
recover and then studied 5 weeks after surgery.
All of the rats in each animal model were housed in plastic
cages and allowed free access to food and water. They were
fasted for 12 hours prior to surgery. The body weights of the
rats were measured on the day of the experiment. This study
was approved by Taipei Veterans General Hospital Animal
Committee. In addition, the principles of laboratory animal
care (Guide for the Care and Use of Laboratory Animals,
DHEW publication no. [NIH] 85-23, rev. 985, Office of
Science and Health Reports, DRR/NIH, Bethesda, MD 20205,
USA) were followed throughout all experiments.
2.3. Experimental design
Several parameters were measured, including spontaneous
motor activity counts (total, ambulatory, and vertical), mean
arterial pressure (MAP), portal pressure (PP), and heart rate
(HR), and flow-pressure curve studies were performed in the
PVL rats 7 days after surgery and in the BDL rats 5 weeks
after surgery. After measurement of the motor activities using
an Opto-Varimex animal activity meter (Columbus Instru-
ments, Inc., Columbus, OH, USA),14MAP, PP, and HR were
measured in each rat. Blood was collected before the flow-
pressure curve studies were performed. The blood samples
were separated in a refrigerated centrifuge at 4?C and 2935 g
for 10 minutes, then stored at ?70?C in pyrogen-free poly-
ethylene tubes for subsequent analysis within 6 weeks. TNF-a
4 I.-F. Hsin et al. / Journal of the Chinese Medical Association 75 (2012) 3e9
and liver biochemistry tests (alanine aminotransferase [ALT],
aspartate aminotransferase [AST], and total bilirubin) were
performed. The flow-pressure curve study was performed by
in situ collateral perfusion. The slope of the flow-pressure
curve represents the overall resistance of the portosystemic
collaterals to the perfusion flow and was used as an index of
the degree of portosystemic shunting. Correlations between
motor activity counts, liver biochemistry parameters, and the
degree of portosystemic shunting were determined. Rats that
received sham operation served as controls. Since no porto-
systemic collaterals developed in the sham-operated rats, they
were analyzed for all of the aforementioned parameters except
the flow-pressure curve study.
2.4. Measurement of spontaneous motor activities
In the open field, the total, ambulatory, and vertical motor
activity counts were measured using the Opto-Varimex animal
activity meter (Columbus Instruments, Inc., Columbus, OH,
USA)14for 30 minutes. The Opto-Varimex activity sensors
utilize high-intensity, modulated, infrared light beams to
detect animal motion. Animals were housed in transparent
cages (17 inches ? 17 inches ? 8 inches) through which 30
infrared beams were passed through the horizontal plane, 15
through each axis. This device is able to differentiate between
nonambulatory (e.g., scratching, gnawing, etc.) and ambula-
tory movements based on the number of consecutive inter-
ruptions of the infrared monitoring beams.
2.5. Hemodynamic measurements
Hemodynamic studies were performed on rats under ket-
amine anesthesia (100 mg/kg body weight, intramuscularly
administered). The right internal carotid artery was cannulated
using a polyethylene PE-50 catheter connected to a Spec-
tramed DTX transducer (Spectramed, Inc., Oxnard, CA,
USA), and continuous recordings of MAP were made using
a multichannel recorder (model RS 3400; Gould Inc., Cuper-
tino, CA, USA). An external zero reference limit was placed at
the midportion of each rat. The heart rate was determined from
the recording. The abdomen was then opened with a midline
incision, and the mesenteric vein was cannulated using
a saline-filled PE-50 catheter for the sham-operated rats or an
18-G catheter for the PVL and BDL rats, which was connected
to a Spectramed DTX transducer. The abdominal cavity was
then closed, and the portal pressure was recorded using
a Gould Model RS 3400 recorder.15,16
2.6. Flow-pressure curve study
The flow-pressure curve studies were performed using in
situ-perfused portosystemic collaterals. When determining the
flow-pressure curves, the perfusion flow rate increased by
3 mL/min, starting at 5 mL/min and climbing to 30 mL/min in
the PVL rats and starting at 6 mL/min and climbing to 18 mL/
min in the BDL rats. Each new flow rate was allowed to
stabilize for 3 minutes before the next higher flow rate was
adjusted. The perfusion pressures were continuously recorded,
and only one flow-pressure curve was performed for each
preparation. The slope of a flow-pressure curve represents the
overall resistance of the portosystemic collaterals and was
regarded as the severity index of portosystemic shunting.17,18
A higher resistance indicates a lower degree of portosyste-
2.7. In situ perfusion of portosystemic collaterals
The in situ perfusion system was used as described in
previous reports.17e19In brief, both jugular veins were can-
nulated using 16-gauge Teflon cannulas in order to ensure an
adequate outflow without any resistance, even at high flow
rates. Heparin (200 units/100g body weight) was injected
through one of the cannulas. The abdomen was opened and an
18-gauge Teflon cannula was inserted into the distal mesen-
teric vein and used as the perfusate inlet. To exclude the liver
from perfusion, a ligature was tied twice around the portal
vein. The animal was then transferred to a warm chamber
(37 ?0.5?C). The temperature around the perfusion area was
continuously monitored using a thermometer placed inside the
mesentery and maintained at approximately 37 ? 0.5?C using
a thermostatic pad and a temperature-controlled infrared lamp.
An open-circuit perfusion was then started using Krebs solu-
tion (composed of the following [in mM]: NaCl, 118; KCl,
4.7; KH2PO4, 1.2; MgSO4, 1.2; CaCl2, 2.5; NaHCO3, 25; and
dextrose 11.0; pH 7.4; 37 ? 0.5C) via a mesenteric cannula by
a roller pump (model 505S; Watson-Marlow Limited, Fal-
mouth, Cornwall, UK). The perfusate was equilibrated with
carbogen gas (95% O2,5% CO2) using a silastic membrane
lung.20Both jugular vein cannulas were simultaneously
opened in order to completely wash out the blood. Pneumo-
thorax was created by opening slits in the diaphragm in order
to increase resistance in the pulmonary arteries and prevent the
perfusate from entering the left chambers of the heart. The
portosystemic collaterals were then perfused with oxygenated
(95% O2, 5% CO2) Krebs solution containing 3% wt/vol
albumin (factor V bovine serum albumin; Sigma, St. Louis,
MO, USA). The effluent of the perfusion was collected in
a reservoir and was not recirculated. In order to monitor and
continuously record the pressure in the portosystemic collat-
erals, a Spectramed DTX transducer attached to a Gould
model RS 3400 recorder was connected to a side arm and
placed just proximal to the perfusion cannula with the zero
placed at the level of the right atrium. Because the temperature
and pressure of the system was stabilized within 20 minutes,
all of the experiments were performed within 25 minutes after
starting perfusion at a constant rate of 20 mL/minute (PVL rat)
or 12 mL/minute (BDL rat).
2.8. Determination of plasma TNF-a levels
Plasma TNF-a levels were measured using a commercially
available solid-phase sandwich enzyme-linked immunosorbent
assay (rat TNFa kits; R & D Systems, Minneapolis, MN,
USA) according to the protocol supplied by the manufacturer.
5 I.-F. Hsin et al. / Journal of the Chinese Medical Association 75 (2012) 3e9
The standards and samples were incubated in a TNF-a anti-
body-coated 96-well microtiter plate. The wells were washed
with buffer and then incubated with anti-TNF-a antibodies
that had been conjugated to horseradish peroxidase for
2 hours. This was washed away, and a yellow-brownish color
developed in the presence of the tetramethylbenzidine chro-
mogen substrate. The intensity of the color was measured
using a Bio-kinetics Reader (Bio-Tek Instruments, Inc.,
Winooski, VT, USA) in order to determine the absorbance at
450 nm with a corrected wavelength of 570 nm. The samples
were compared against a standard curve in order to determine
the amount of TNF-a present. All samples were run in
duplicate. The lower limit of the TNF-a sensitivity of this
assay was 5 pg/dL. The intra- and interassay coefficients of
variation were 5.1% and 9.7%, respectively.
2.9. Determination of liver biochemistry parameters
Serum levels of ALT, AST, and total bilirubin were
measured using a colorimetric assay (Daiichi Olympus AU
600; Olympus, Inc., Mizima, Japan).
The reagents used to prepare the Krebs solution were
purchased from Sigma (Sigma Chemical Co., St. Louis, MO,
USA). All of the solutions were freshly prepared on the days
2.11. Statistical analysis
The results are expressed as the means ? standard error of
the mean (SEM). The flow-pressure curves were analyzed by
linear regression. Statistical analyses were performed using
the independent Student’s t test and linear correlation analysis,
as appropriate. The results were considered statistically
significant if a p-value <0.05 was determined using the two-
3.1. Hemodynamic data of the study animals
Table 1 shows the MAP, HR, and PP of the PVL, BDL, and
corresponding control groups. Compared with the PVL-
control group, PP was significantly higher (p < 0.001) and
MAP was significantly lower in the PVL group (p ¼ 0.005).
Similarly, PP in the BDL rats was significantly higher
(p < 0.001) and MAP was significantly lower in the BDL
group (p ¼ 0.048) than in the BDL-control group. There were
no statistical differences in terms of HR between the PVL,
BDL, and their corresponding control groups.
3.2. Plasma TNF-a and liver biochemistry parameters
There were no significant differences between the PVL and
PVL-control groups in terms of TNF-a (1.3 ? 0.7 vs.
1.2 ? 0.8 pg/mL, p ¼ 0.868), ALT (14.5 ?3.7 vs. 14.8 ?1.9 m/
L, p ¼ 0.943), and AST (96.6 ?10.4 vs. 78.5 ? 6.1 m/L,
p ¼ 0.161).
Table 2 shows comparisons between the BDL and BDL-
control groups. Plasma TNF-a, AST, and total bilirubin
concentrations of the BDL group were significantly higher
than those of the BDL-control group. In contrast, ALT levels
were not significantly different between the BDL and BDL-
3.3. Motor activities
Fig. 1 depicts the motor activity counts of the PVL and
2498.2 ? 316.4 counts/30 minutes, p ¼ 0.727), ambulatory
(1368.9 ? 188.6
p ¼ 0.500), and vertical (529.9 ? 67.7 vs. 712.0 ? 129.4
counts/30 minutes, p ¼ 0.200) movements of the PVL group
were not significantly different from those measured in the
PVL-control group. Compared with the BDL-control group,
thetotal(1030.0 ? 109.5
30 minutes, p < 0.001) and ambulatory (550.7 ?79.9 vs.
1045.4 ? 129.4 counts/30 minutes, p ¼ 0.005) movements
were significantly lower in the BDL group. The vertical
total (2351.9 ?263.4
vs. 1568.7 ? 215.9 counts/30 minutes,
vs.1990.0 ? 203.1counts/
Hemodynamic parameters of PVL, BDL and corresponding control groups.
Tested itemsPVL (n ¼8)
PVL-control (n¼ 6)
PVL¼partial portal vein ligation; BDL ¼common bile duct ligation; MAP¼mean arterial pressure; PP¼portal pressure.
Data are expressed as the mean ?SEM.
ap<0.05 between PVL and PVL-control groups.
bp<0.05 between BDL and BDL-control groups.
Blood test results of the BDL and BDL-control groups.
Tested itemsBDL BDL-controlp
Total bilirubin (mg/dL)
BDL¼bile-duct ligation; ALT¼alanine aminotransferase; AST¼aspartate
aminotransferase; TNF-a¼tumor necrosis factor-a. Data are expressed as
6 I.-F. Hsin et al. / Journal of the Chinese Medical Association 75 (2012) 3e9
30 minutes) was also lower than that of the BDL-control
(532.2 ? 73.6 counts/30 minutes), but it did not reach a level
of statistical significance (p ¼ 0.136).
of theBDL group (328.9 ? 72.2 counts/
3.4. Degree of correlation between the extent of HE,
flow-pressure curves, liver biochemistry parameters, and
In both the PVL and BDL rats, there was no significant
correlation between the motor activity counts and the slope of
the flow-pressure curve (Fig. 2). In BDL rats with significant
HE, as indicated by decreased motor activities, a negative
correlation was found between the motor activity count and
the concentrations of various liver injury parameters, including
AST (p ¼ 0.007, R¼ ?0.595) and total bilirubin (p ¼ 0.001,
R ¼?0.692, Fig. 3). Furthermore, a negative correlation was
found between the TNF-a level and total movements in the
BDL rats (Fig. 4). No correlations were observed between
motor activity counts, the concentrations of various liver
injury parameters, and TNF-a levels in PVL rats (p > 0.05).
The pathogenesis of HE is complicated and not fully
understood.21,22However, insufficient liver functions and
the presence of portosystemic collaterals, which result in the
failure to metabolize toxic substances derived from the
gastrointestinal tract, are considered fundamental to HE.
Common animal models used to study HE include models of
fulminant hepatic failure23,24and portacaval shunts.25,26To
investigate the pathogenesis of HE, the early detection of
neurological and behavioral disturbances in these animals
is crucial. Our previous studies27,28have demonstrated that
the Opto-Varimex animal activity meter is a valid tool for the
detection of motor activities in study animals. Based on the
consecutive number of interruptions of the infrared monitoring
beams, we were able to differentiate nonambulatory (e.g.,
scratching, gnawing, etc.) from ambulatory movements. In
addition to horizontal movements, an additional row of
infrared cells placed above this plane could provide more
information on vertical movements. The results of the motor
activity studies correlated with neurobehavioral scores.27
Fig. 1. Spontaneous, total, ambulatory, and vertical activity counts in portal vein-ligated (PVL) and PVL-control groups (A) and bile duct-ligated (BDL) and BDL-
control groups (B). The BDL rats, but not PVL rats, demonstrated lower total and ambulatory movements than their corresponding controls.
0.6 0.91.21.5 1.8
Fig. 2. Correlation between the slope of the flow-pressure curve and the total motor activity counts of the PVL (A) and BDL (B) rats. There was no significant
correlation between these two parameters in either group.
7 I.-F. Hsin et al. / Journal of the Chinese Medical Association 75 (2012) 3e9
The portosystemic collateral perfusion model used in this
study provides information on the degree of shunting.17,18This
is a highly sensitive and reproducible technique that can be
used to evaluate the collateral vasculature. According to
Poiseuille’s Law, the amount of resistance in the collateral
vascular bed can be used as an index of portosystemic
shunting. This technique has demonstrated that the slope of the
flow-pressure relationship (i.e., the collateral vascular resis-
tance) decreases significantly and progressively 2e7 days after
induction of portal hypertension in rats,18which is comparable
to that measured using the microsphere method.29In addition,
the nonrecirculating perfusion system excludes the influence
of vasoactive substances that are released from anoxic organs
In the present study, we used two animal modelsdBDL
and PVL rats (representing animals that suffer from porto-
systemic collaterals, with
respectively)dto investigate the relationship between degrees
of portosystemic shunting and changes in motor activities. We
demonstrated that BDL rats have significantly more severe
liver injuries and a higher level of TNF-a. In addition, their
motor activities were significantly lower compared with those
of the BDL-control group, which is compatible with the
findings of previous studies.27,28In contrast, the motor activ-
ities of the PVL group were not significantly different from
those of the PVL-control group. These findings indicate that
HE might not even occur in animals with a high degree of
portosystemic collaterals and normal liver functions. There-
fore, PVL model could not supersede other animal models for
The current study reveals that the slope of the flow-pressure
curve does not correlate with motor activity counts. Rather, in
BDL rats with chronic liver injuries, AST and total bilirubin
levels are correlated with motor activity counts. These findings
suggest that portosystemic collaterals play a smaller role than
liver functions when determining the severity of HE. However,
we could not completely exclude the presence of collaterals as
an additional factor that contributes to HE in cirrhotic rats.
Collaterals divert toxic substances that elicit HE from the
portal tributary blood flow into systemic circulation, and the
study animals may have had a wide variation in their tolerance
to the implicated substances.
It has been reported in recent years that the increased
expression of TNF-a in the frontal cortex of rats following
hepatic devascularization is correlated with the clinical
progression of encephalopathy and brain edema.30In this
study, there was a significant correlation between TNF-
a levels and motor activities in BDL and BDL-control rats. In
other words, the higher the TNF-a concentration, the lower the
or withoutliver failure,
0 1020 3040 50 6070
0 300 600900 1200 15001800
Total bilirubin (mg/dl)
Fig. 3. Correlation between circulating liver biochemistry concentrations and
motor activity counts in BDL rats. Negative correlations were found between
AST, total bilirubin, and total motor activity counts.
0 20 406080 100 120140
TNF- ( g/mL)
Fig. 4. Correlation between plasma TNF-a levels and motor activity counts in
BDL rats. A negative correlation was found between the TNF-a concentration
and the motor activity count.
8 I.-F. Hsin et al. / Journal of the Chinese Medical Association 75 (2012) 3e9
motor activity counts. Our data are compatible with the results Download full-text
of a previous study that reported that cirrhotic patients with
abnormal psychometric hepatic encephalopathy scores had
higher TNF-a levels than those with normal scores. Further-
more, there was a significant correlation between psycho-
metric hepatic encephalopathy
concentration.31This is further demonstrated by an animal
study that reported that the onset of severe encephalopathy
(e.g., coma) and brain edema in mice with azoxymethane-
induced acute liver failure were significantly delayed in TNF
receptor-1 knockout mice.32Overall, therapies that target
TNF-a may contribute to the amelioration of HE, but further
clarification is required.
In conclusion, in rats with chronic liver injury and HE, liver
function parameters, rather than the degree of shunting, were
shown to correlate with motor activities. Portosystemic
collaterals seem to play a less prominent role in determining
the severity of HE. The significant correlation between
circulating TNF-a levels and motor activity counts suggest the
significant role that TNF-a plays in the development of HE.
The authors gratefully acknowledge Yun-Ni Hsieh and Jia-
Yi Liao for their excellent technical assistance. This work was
supported by grants (VGH 92-20 and VGH92-50) from Taipei
Veterans General Hospital, Taiwan.
1. Gammal SH, Jones EA. Hepatic encephalopathy. Med Clin North Am
2. Mousseau DD, Butterworth RF. Current theories on the pathogenesis of
hepatic encephalopathy. Soc Exp Biol Med 1994;206:329e44.
3. Geraghty JG, Angerson WJ, Carter DC. Portal venous pressure and por-
tasystemic shunting in experimental portal hypertension. Am J Physiol
4. Vogels BA, van Steynen B, Maas MA, Jorning GG, Chamuleau RA. The
effect of ammonia and portal-systemic shunting on brain metabolism,
neurotransmission and intracranial hypertension in hyperammonaemia-
induced encephalopathy. J Hepatol 1997;26:387e95.
5. Borrelli E, Roux-Lombard P, Grau GE, Ricou B, Dayer J, Suter PM.
Plasma concentrations of cytokines, their soluble receptors, and antioxi-
dant vitamins can predict the development of multiple organ failure in
patients at risk. Crit Care Med 1996;24:392e7.
6. Sztrymf B, Rabiller A, Nunes H, Savale L, Lebrec D, Le Pape A, et al.
Prevention of hepatopulmonary syndrome and hyperdynamic state by
pentoxifylline in cirrhotic rats. Eur Respir J 2004;23:752e8.
7. Chu CJ, Lee FY, Wang SS, Lu RH, Tsai YT, Lin HC, et al. Hyperdynamic
circulation of cirrhotic rats with ascites: role of endotoxin, tumour
necrosis factor-alpha and nitric oxide. Clin Sci (Lond) 1997;93:219e25.
8. Kerfoot SM, D’Mello C, Nguyen H, Ajuebor MN, Kubes P, Le T, et al.
TNF-a- secreting monocytes are recruited into the brain of cholestatic
mice. Hepatology 2006;43:154e62.
9. Chu CJ, Chen CT, Wang SS, Lee FY, Chang FY, Lin HC, et al. Hepatic
encephalopathy in rats with thioacetamide-induced fulminant hepatic
failure: role of endotoxin and tumor necrosis factor-alpha. J Chin Med
10. Kountouras J, Billing BH, Scheuer PJ. Prolonged bile duct obstruction:
a new experimental model for cirrhosis in rat. Br J Exp Pathol 1984;65:
11. Lebrec D, Blanchet L. Effect of two models of portal hypertension on
splanchnic organ blood flow in the rat. Clin Sci 1985;68:23e8.
12. Chojkier M, Groszmann RJ. Measurement of portal-systemic shunting in
the rat by using g-labeled microspheres. Am J Physiol 1981;240:G371e5.
13. Fernandez M, Pizcueta P, Garcia-Pagan JC, Feu F, Cirera I, Bosch J, et al.
Effects of ritanserin, a selective and specific S2-serotoninergic antagonist,
on portal pressure and splanchnic hemodynamics in rats with long-term
bile duct ligation. Hepatology 1993;18:389e93.
14. Bengtsson F, Gage FH, Jeppsson B, Nobin A, Rosengren E. Brain
monoamine metabolism and behavior in portacaval shunted rats. Exp
15. Wang SS, Chan CC, Lee FY, Chang FY, Lin HC, Chen CT, et al. Effects of
long-term octreotide treatment on the response of portal-systemic collat-
erals to vasopressin in portal hypertensive rats. Eur J Clin Invest
16. Lee FY, Wang SS, Tsai YT, Lin HJ, Lin HC, Chu CJ, et al. Amino-
guanidine corrects hyperdynamic circulation without ameliorating portal
hypertension and portal hypertensive gastropathy in anesthetized portal
hypertensive rats. J Hepatol 1997;26:687e93.
17. Mosca P, Lee FY, Kaumann AJ, Groszmann RJ. Pharmacology of portal-
systemic collaterals in portal hypertensive rats: role of endothelium. Am J
18. Lee FY, Colombato LA, Albillos A, Groszmann RJ. Administration of N-
omega-nitro-L-arginine ameliorates portal-systemic shunting in portal-
hypertensive rats. Gastroenterology 1993;105:1464e70.
19. Chan CC, Lee FY, Wang SS, Chang FY, Lin HC, Chu CJ, et al. Effects of
vasopressin on portal-systemic collaterals in portal hypertensive rats: role
of nitric oxide and prostaglandin. Hepatology 1999;30:630e5.
20. Hamilton RL, Berry MN, Williams MC, Severinghaus EM. A simple and
inexpensive membrane “lung” for small organ perfusions. J Lipid Res
21. Vaquero J, Chung C, Cahill ME, Blei AT. Pathogenesis of hepatic
encephalopathy in acute liver failure. Semin Liver Dis 2003;23:259e69.
22. Haussinger D, Kircheis G, Fischer R, Schliess F, vom Dahl S. Hepatic
encephalopathy in chronic liver disease: a clinical manifestation of
23. Chu CJ, Wang SS, Lee FY, Chang FY, Lin HC, Hou MC, et al. Detri-
mental effects of nitric oxide inhibition on hepatic encephalopathy in rats
with thioacetamide-induced fulminant hepatic failure. Eur J Clin Invest
24. Rahman TM, Hodgson HJ. Animal models of acute hepatic failure. Int J
Exp Pathol 2000;81:145e57.
25. Rigotti P, Jonung T, James JH, Edwards LL, Peters JC, Fischer JE.
Infusion of branched-chain amino acids and ammonium salts in rats with
portacaval shunts. Arch Surg 1985;120:1290e5.
26. Hawkins PA, DeJoseph MR, Vina JR, Hawkins RA. Comparison of the
metabolic disturbances caused by end-to-side and side-to-side portacaval
shunts. J Appl Physiol 1996;80:885e91.
27. Chu CJ, Lee FY, Wang SS, Chang FY, Lin HC, Wu SL, et al. Establish-
ment of an animal model of hepatic encephalopathy. J Chin Med Assoc
28. Chan CY, Huang SW, Wang TF, Lu RH, Lee FY, Chang FY, et al. Lack of
detrimental effects of nitric oxide inhibition in bile duct-ligated rats with
hepatic encephalopathy. Eur J Clin Invest 2004;34:122e8.
29. Sikuler E, Kravetz ED, Groszmann RJ. Evolution of portal hypertension
and mechanisms involved in its maintenance in a rat model. Am J Physiol
30. Jiang W, Desjardins P, Butterworth RF. Direct evidence for central
proinflammatory mechanisms in rats with experimental acute liver failure:
protective effect of hypothermia. J. Cereb Blood Flow Metab 2009;29:
31. Montagnese S, Biancardi A, Schiff S, Carraro P, Carla V, Mannaioni G,
et al. Different biochemical correlates for different neuropsychiatric
abnormalities in patients with cirrhosis. Hepatology 2011;53:558e66.
32. Be ´meur C, Qu H, Desjardins P, Butterworth RF. IL-1 or TNF receptor
gene deletion delays onset of encephalopathy and attenuates brain edema
in experimental acute liver failure. Neurochem Int 2010;56:213e5.
cerebral edema.J Hepatol
9I.-F. Hsin et al. / Journal of the Chinese Medical Association 75 (2012) 3e9