Liver International ISSN 1478-3223

CIRRHOSIS AND LIVER FAILURE

Lipopolysaccharide precipitates hepatic encephalopathy and increases

blood–brain barrier permeability in mice with acute liver failure
langer1, Bich N. Nguyen2 and Roger F. Butterworth1
Anne Chastre1, Mireille Be
1 Neuroscience Research Unit, Saint-Luc Hospital, CRCHUM, Montreal, QC, Canada
2 Department of pathology, Saint-Luc Hospital, CHUM, Montreal, QC, Canada

Keywords Abstract
acute liver failure – azoxymethane – blood– Background & Aims: Acute liver failure (ALF) is frequently complicated by
brain barrier – endotoxemia – infection leading to precipitation of central nervous system complications
immunoglobulin G extravasation – such as hepatic encephalopathy (HE) and increased mortality. There is evi-
lipopolysaccharide – matrix dence to suggest that when infection occurs in ALF patients, the resulting
metalloproteinase-9 – pro-inflammatory pro-inflammatory mechanisms may be amplified that could, in turn, have a
cytokines – systemic inflammatory response major impact on blood–brain barrier (BBB) function. The aim of this study
was to investigate the role of endotoxemia on the progression of encephalop-
Correspondence
athy in relation to BBB permeability during ALF. Methods: Adult male C57-
Roger F. Butterworth, PhD, DSc, BL6 mice with ALF resulting from azoxymethane-induced toxic liver injury
Neuroscience Research Unit, CRCHUM, were administered trace amounts of the endotoxin component lipopolysac-
Campus Saint-Luc, Universit
e de Montr
eal, charide (LPS). Effects on the magnitude of the systemic inflammatory
1058 St-Denis Street, H2X 3J4 Montreal, response, liver pathology and BBB integrity were measured as a function of
QC, Canada progression of HE, defined as time to loss of corneal reflex (coma). Results:
Tel: (514) 890-8000 (ext 35759) Lipopolysaccharide caused additional two- to seven-fold (P < 0.001)
Fax: (514) 412-7253 increases in circulating pro-inflammatory cytokines (TNF-a, IL-1b, IL-6),
e-mail: roger.butterworth@umontreal.ca worsening liver pathology and associated increases of circulating transamin-
ases as well as increased hyperammonaemia consistent with a further loss of
Received 3 April 2013 viable hepatocytes. LPS treatment of ALF mice led to a rapid precipitation of
Revised 16 May 2013 hepatic coma and the BBB became permeable to the 25-kDa protein immu-
Accepted 31 May 2013 noglobulin G (IgG). This extravasation of IgG was accompanied by ignificant
up-regulation of matrix metalloproteinase-9 (MMP-9), an endopeptidase
DOI:10.1111/liv.12252
known to modulate opening of the BBB in a wide range of neurological dis-
orders. Conclusions: These findings represent the first direct evidence of
inflammation-related BBB permeability changes in ALF.

Hepatic encephalopathy (HE) is the hallmark neurologi-
cal feature and a determinant of clinical outcome in
patients with acute liver failure (ALF). On account of
the presence of multiple immunological defects, over
90% of ALF patients have clinical or bacteriological
evidence of infection and there is a clear temporal asso-
ciation among the acquisition of the infection, the
appearance of a systemic inflammatory response and
progression of HE in these patients (1, 2). However, the
mechanisms responsible for the precipitation of HE by
infection in ALF remain poorly defined.
Gram-negative bacteria are frequently implicated as
the cause of infection/sepsis in ALF and increased circu-
lating levels of endotoxin, a lipopolysaccharide (LPS)
complex of the outer cell wall of gram-negative bacteria,
A portion of the results described in this manuscript was presented
at the annual meeting of the AASLD, Boston, October 2012, and
published in abstract form Hepatology, 956A, 2012.
are associated with an increased incidence of HE in ALF
(2, 3). Endotoxin treatment leads to the release of
pro-inflammatory cytokines such as tumour necrosis
factor-alpha (TNF-a) and the interleukins (IL-1b and
IL-6), which have been shown to enhance liver injury in
experimental animal models of ALF (4, 5).
To start to elucidate the basic mechanisms that
underlie the precipitation of HE by systemic inflamma-
tion caused by infection in ALF, this study made use of
a single low dose of LPS administered to mice with ALF
as a result of azoxymethane-induced hepatotoxicity.
This is a well-characterized mouse model of ALF that
recapitulates the nature and extent of liver injury as well
as systemic inflammation and neurological complica-
tions seen in the human condition (6). In view of the
recent resurgence of interest in changes in blood–brain
barrier (BBB) in relation to HE in ALF (7) and the role
of inflammation in the pathogenesis of BBB disruption
in a wide range of neurological disorders (8–10), BBB

Liver International (2013)

© 2013 John Wiley & Sons A/S. Publishing by John Wiley & Sons Ltd. 1
Endotoxemia during acute liver failure in mice
integrity was assessed by measurement of the extravasa-
tion of immunoglobulin G (IgG) and by measurement
of metalloproteinase-9 activity (MMP-9), suggested
previously to contribute to altered BBB permeability to
small molecules in this model of ALF (11). In this study,
measurement of BBB integrity was made as a function
of the presence of the systemic inflammatory response
and the progression of encephalopathy.
Material and methods
Animals
Adult male C57BL6 mice (20–30 g) (Charles River,
Saint-Constant, QC, Canada) were maintained in a 12-h
light/dark cycle, and supplied with standard laboratory
chow and water ad libitum. All animals were free of
infection at the onset of the experiment. All procedures
were carried out in accordance to the Guidelines of the
Canadian Council of Animal Care and protocols were
approved by the Animal Research Committee at Saint-
Luc Hospital (CHUM).
Animal treatments and experimental design
Acute liver failure was induced by the injection of
azoxymethane (AOM) (100 lg/g, i.p.) (Sigma-Aldrich, St.
Louis, CO, USA) dissolved in 100 ll saline as previously
described (6). To mimic infection, mice received a single
dose of LPS (2 lg/kg, i.p.) (Escherichia coli 0111:B4;
Sigma-Aldrich), dissolved in 100 ll saline 1 h after
AOM injection. Five groups were constituted, namely
(i) Saline-treated mice (Control-Veh), (ii) LPS-treated
normal mice (Control-LPS), (iii) Saline-treated ALF
mice (ALF-Veh), (iv) LPS-treated ALF mice at coma
stages of HE (ALF+LPS) and (v) Saline-treated ALF
mice at coma stages of HE (ALF-Coma). CTRL-Veh,
Time (h) 0 1
Group 1 Saline
Group 2 Saline
Group 3 AOM
Group 4 AOM
Group 5 AOM
Fig. 1. Experimental design. Schematic representation of the procedures and timeline followed up for each group of this study. Mice from
CTRL-Veh, CTRL-LPS and ALF-Veh groups were sacrificed along with ALF+LPS mice, when ALF+LPS mice reached coma stages of hepatic
encephalopathy, on average 8.3 ± 0.5 h following AOM injection. ALF-Coma mice were sacrificed at coma stages of HE, which occur
16.4 ± 0.4 h following AOM injection.
2 © 2013 John Wiley & Sons A/S. Publishing by John Wiley & Sons Ltd.
Chastre et al.
CTRL-LPS and ALF-Veh mice were pair-sacrificed when
ALF+LPS mice reached coma stage of encephalopathy,
which occurred on average 8.3 h following AOM
administration (Fig. 1). LPS doses in the lg/kg range
were chosen to induce non-lethal endotoxemia, as much
higher doses (in the range of 1–25 mg/kg) are known to
mimic sepsis in mice resulting in death of approxi-
mately half of animals (12). Following AOM injection,
body temperature was carefully monitored with a rectal
probe and rigorously maintained in the range of 36.5–
37.5°C using heating pads and lamps. Glycaemia was
monitored and kept at 5–7 mM by subcutaneous injec-
tions of 10% dextrose. Time to coma was determined by
time to loss of corneal reflex. After anaesthesia with a
ketamine/xylazine cocktail (50 and 9 mg/kg, respec-
tively, i.p.), plasma samples were immediately collected
from the heart into heparinized tubes, centrifuged
(10 min, 12 000 g) and kept at 70°C. Mice were then
perfused transcardially with saline. Brains were rapidly
removed, flash frozen and kept at 70°C until use,
whereas livers were fixed overnight by immersion in
10% buffered formalin.
Histological assessment of liver damage
Paraffin-embedded specimens were prepared and sec-
tions (6 lm) were mounted on Superfrost-plus micro-
scope slides (Fisher Scientific, Pittsburg, PA, USA). HPS
(haematoxylin-phloxin-saffron) staining was performed
according to a standard protocol and liver pathology
was assessed by an investigator blinded to the experi-
mental treatment groups.
Biochemical assays
Plasma samples were diluted 10-fold in saline and alanine
aminotransferase (ALT) and aspartate aminotransferase
PAIR-SACRIFICE COMA
8.3 ± 0.5 16.4 ± 0.4
Saline Control–Veh
LPS Control–LPS
Saline ALF–Veh
LPS ALF+LPS
(Precipitated Coma)
Saline ALF–Coma
(Separate study)
Liver International (2013)
Chastre et al.
(AST) activities were assayed with an automated analy-
ser. Ammonia levels in plasma were determined using
a commercial ammonia assay kit (Sigma-Aldrich).
Samples were diluted 40-fold in assay diluent and absor-
bance was read at 340 nm.
Plasma IL-6, IL-1b and TNF-a measurements
Pro-inflammatory cytokine levels were measured in
plasma using enzyme-linked immunosorbent assay
(ELISA) kits specific for mouse IL-6 (eBioscience, San
Diego, CA, USA), IL-b and TNF-a (R&D Systems, Min-
neapolis, MN, USA). Detection limits were 4 pg/ml,
3 pg/ml and 5.1 pg/ml respectively. For IL-1b and TNF-
a assays, samples were diluted 20-fold and 5-fold,
respectively, in calibrator diluent and incubated 2 h at
room temperature. For IL-6 measurements, samples
were diluted 17-fold in assay diluent and incubated
overnight at 4°C. The plates were read at 450 nm and
values at 570 nm were subtracted. Absorbance was con-
verted to pg/ml for cytokine measurements in plasma
using standard curves prepared with respective recombi-
nant cytokines.
Western blot analysis
Samples of frontal cortex were homogenized in ice-cold
buffer (50 mM Tris pH 7.5, 150 mM NaCl, 1 mM
EDTA, 1% Triton X-100, 0.1% SDS, 0.5% NaDOC)
containing a protease inhibitor cocktail (Sigma-Aldrich)
and centrifuged at 12 000 g for 45 min. Proteins
(10–30 lg) were resolved on 8% denaturing SDS-poly-
acrylamide gels and transferred for 1 h to polyvinylid-
ene fluoride (PVDF) membranes (Bio-Rad Laboratories,
Hercules, CA, USA). The membranes were probed for
1 h with antibodies directed against ZO-1 (rabbit,
1/1000, #61-7300, Invitrogen, Carlsbad, CA, USA),
ZO-2 (rabbit, 1/1000, #38-9100, Invitrogen), occludin
(rabbit, 1/1000, #71-1500, Invitrogen) and b-actin
(mouse, 1/25 000, #A3853, Sigma-Aldrich), then
incubated for 1 h with respective horseradish peroxi-
dase-conjugated secondary antibodies (1/10 000, Per-
kin-Elmer Life Sciences, Boston, MA, USA). After
extensive washing, peroxidase activity was detected by
enhanced chemiluminescence (GE Healthcare, Arling-
ton Heights, IL, USA). The intensity of the bands was
measured by densitometry using Quantity One software
(Bio-Rad Laboratories).
IgG extravasation
IgG (25 kDa) extravasation was performed as previously
described, with modifications (13). Briefly, twenty
micrograms of proteins from samples of frontal cortex
were solubilized in Laemmli buffer, boiled for 5 min,
then resolved on 8% denaturing SDS–PAGE and trans-
ferred 1 h to PVDF membranes. Membranes were incu-
bated for 1 h with an anti-mouse horseradish
Liver International (2013)
© 2013 John Wiley & Sons A/S. Publishing by John Wiley & Sons Ltd.
Endotoxemia during acute liver failure in mice
peroxidase-conjugated secondary antibody (1/3000, Per-
kin-Elmer Life Sciences). Peroxidase activity was then
detected by enhanced chemiluminescence (GE Health-
care). Intensity of the bands was measured by densitome-
try using Quantity One software (Bio-Rad Laboratories).
SDS–PAGE gelatin zymography
Gelatin zymography was performed as previously
described, with modifications (13). Twenty micrograms
of protein fraction were electrophoresed onto 10% non-
denaturing SDS–PAGE containing 1 mg/ml of gelatin
(Bio-Rad Laboratories). Mixtures (1.5 ng) of recombi-
nant MMP-9 and MMP-2 (US Biological, Cleveland,
OH, USA) were used as standards. The gels were
processed, incubated in 2.5% Triton X-100 for 1 h,
incubated in a zymogen buffer (50 mM Tris–HCl, pH
7.4, 5 mM CaCl2, 200 mM NaCl, 0.02% Brij.35) at
37°C for 40 h, stained with 0.5% Coomassie Blue R-250
for 3 h, and destained with three changes of 30% meth-
anol, 10% acetic acid (for 15, 30 and 60 min). Gels were
scanned and the bands were quantified using Quantity
One Software (Bio-Rad Laboratories). Gelatinase activ-
ity appears as a clear band against a blue background.
Statistical analysis
All data are expressed as the mean ± SEM. Statistical
analyses were performed using one-way analysis of vari-
ance (ANOVA) followed by Tukey’s post hoc analysis.
P value < 0.05 was considered to indicate a significant
difference. Data were analysed using Prism 5.0 software
(GraphPad Prism 5.0, San Diego, CA, USA).
Results
LPS treatment worsens hepatic damage following AOM
administration
Azoxymethane administration rapidly induced diffuse
necrosis affecting primarily centrilobular and mediolob-
ular regions in ALF-Veh mice (Fig. 2C), which
expanded at coma stages of HE to massive haemorrhag-
ic necrosis and apoptosis with scarce neutrophil infiltra-
tion in ALF-Coma mice (Fig. 2D). In contrast,
ALF+LPS mice displayed peliosis and increased neutro-
phil infiltration in addition to the massive haemorrhagic
necrosis and apoptosis found in ALF-Coma mice
(Fig. 2E,F). LPS treatment alone induced hepatocyte
ballooning (Fig. 2B). Both plasma transaminase activi-
ties and ammonia levels were significantly increased at
the onset of coma in ALF+LPS mice. Compared with
CTRL-Veh mice, AST activity was increased 9.2-fold
(P < 0.001) in ALF+LPS mice, whereas ALT activity was
increased 29.6-fold (P < 0.001) (Table 1). Plasma
ammonia levels were increased 2.8-fold (P < 0.001) in
ALF+LPS mice compared with CTRL-Veh mice
(Table 1). No changes were observed in transaminase
3
Endotoxemia during acute liver failure in mice Chastre et al.

(A) (B)

(C) (D)

(E) (F)

Fig. 2. LPS worsens hepatic pathology in ALF mice. Representative liver sections from (A) CTRL-Veh mice, (B) CTRL-LPS mice, (C) ALF-Veh
mice, (D) ALF-Coma mice and (E and F) ALF+LPS mice. Mice from CTRL-Veh, CTRL-LPS and ALF-Veh groups were sacrificed along with
ALF+LPS mice, when ALF+LPS mice reached coma stages of HE, on average 8.3 ± 0.5 h following AOM injection. ALF-Coma mice were
sacrificed at coma stages of HE, which occur 16.4 ± 0.4 h following AOM injection. ALF-Veh mice show early centrolobular necrosis (C) that
progresses into extensive haemorrhagic liver necrosis at coma stages in ALF-Coma mice (D). LPS challenge during ALF results in extensive
coagulative liver necrosis with peliosis and increased neutrophil infiltration in comatose ALF+LPS mice (E and F). Black arrow shows
neutrophil infiltration. Arrow heads indicate areas of apoptosis. Representative pictures from n = 5 in each group. Scale bar: 50 lm.

Table 1. LPS increases plasma transaminase activities and ammonia levels in ALF mice
CTRL-Veh CTRL-LPS ALF-Veh ALF+LPS ALF-Coma
AST (U/L) 310 ± 46 860 ± 180 1193 ± 299 2843 ± 422* 4086 ± 484*
ALT (U/L) 96 ± 23 215 ± 87 535 ± 351 2842 ± 511* 4057 ± 2176*
Ammonia (lM) 131.8 ± 26.3 118.0 ± 11.3 144.9 ± 24.7 370.2 ± 53.0* 535.1 ± 63.0*

Mice were administered LPS 1 h after azoxymethane (AOM) or saline (vehicle). CTRL-Veh mice received only saline. CTRL-LPS mice received only LPS.
CTRL-Veh mice, CTRL-LPS mice and ALF-Veh mice were sacrificed along with ALF+LPS mice, when ALF+LPS mice reached coma stages of HE, on aver-
age 8.3 ± 0.5 h following AOM injection. ALF-Coma mice were sacrificed at coma stages of HE, which occur 16.4 ± 0.4 h following AOM injection.
Data represent mean ± SEM of n = 10 in each group.
*P < 0.01 vs. all groups.

activities or in ammonia levels in CTRL-LPS mice or in ALF-Coma mice, whereas plasma levels of ammonia
ALF-Veh (8 h) mice. Compared with CTRL-Veh mice, were increased 4.1-fold (P < 0.001). Plasma levels of
AST and ALT activities were, respectively, increased AST, ALT and ammonia in ALF-Coma mice were not
13.2-fold (P < 0.001) and 42.2-fold (P < 0.001) in significantly different from those ALF+LPS mice.

Liver International (2013)

4 © 2013 John Wiley & Sons A/S. Publishing by John Wiley & Sons Ltd.
Chastre et al. Endotoxemia during acute liver failure in mice

LPS treatment results in further increases in plasma †

pro-inflammatory cytokines in ALF mice 400 (A)
Compared with CTRL-Veh mice, significant increases in
TNF-a levels were observed in ALF-Veh mice (3.6-fold;
**
P < 0.05) and in ALF+LPS mice (7.1-fold; P < 0.001) 300

Plasma TNF-α (pg/mL)

(Fig. 3A). Interestingly, TNF-a levels in ALF+LPS mice
were 2-fold higher than those observed in ALF-Veh
mice (P < 0.01) (Fig. 3A). Similarly, IL-6 levels were
200
significantly increased in ALF-Veh mice (153.5-fold; *
P < 0.05) and in ALF+LPS mice (332.0-fold; P < 0.001)
compared with CTRL-Veh mice (Fig. 4B). IL-6 levels
were significantly higher in ALF+LPS compared with 100
ALF-Veh mice (2.2-fold; P < 0.01) (Fig. 3B). IL-1b lev-
els were significantly increased in ALF+LPS mice com-
pared with CTRL-Veh mice (48.9-fold; P < 0.001)
0
(Fig. 3C). LPS alone did not significantly increase CTRL–Veh CTRL–LPS ALF–Veh ALF+LPS
plasma levels of TNF-a, IL-1b or IL-6.

15 000
†
(B)
LPS treatment results in a further increase in MMP-9
activity in AOM-induced ALF mice
**
No changes of protein expression were observed for
ZO-1, ZO-2 or occludin in either group (data not
Plasma IL-6 (pg/mL)

shown). Gelatin zymography analysis showed increased 10 000

activity of brain MMP-9 in ALF-Veh mice (5.7-fold;
P < 0.05) and in ALF+LPS mice (11.3-fold; P < 0.001)
compared with CTRL-Veh mice (Fig. 4). MMP-9 activ- *
ity in brains of ALF+LPS mice was significantly higher 5000
than that in the brains of ALF-Veh mice (2-fold;
P < 0.01). LPS treatment alone had no effect on brain
MMP-9 activity. No changes in MMP-2 activity were
demonstrated in any of the treatment groups (Fig. 4).
0
CTRL–Veh CTRL–LPS ALF–Veh ALF+LPS
LPS treatment results in IgG extravasation in brain of
AOM-induced ALF mice 2000
(C)
A significant increase in brain IgG extravasation was
observed in ALF+LPS mice compared with CTRL-Veh ¥
mice (16.0-fold; P < 0.001) (Fig. 5). No significant IgG 1500
Plasma IL-1β (pg/mL)

extravasation was observed in LPS-Veh mice or in ALF-

Veh mice.

1000
Discussion
The AOM mouse model has been widely used for the
study of basic mechanisms implicated in the pathogene- 500
sis of the central nervous system (CNS) complications
of ALF (6, 7, 14–16). The model recapitulates the major
components of human ALF resulting from toxic liver
injury including characteristic liver pathology (17), 0
CTRL–Veh CTRL–LPS ALF–Veh ALF+LPS
raised serum transaminases, hyperammonemia, sys-
temic inflammation (15) and severe encephalopathy (2). Fig. 3. Circulating pro-inflammatory cytokines are further
In this study, the superposition of LPS in AOM-treated increased at coma stages of HE in LPS-treated ALF mice. Plasma
mice resulted in rapid worsening of liver function as TNF-a (A), IL-6 (B) and IL-1b (C) were quantified by ELISA. Data
assessed by histopathology, further increases in serum represent mean ± SEM of n = 8 in each group. *P < 0.05 vs. CTRL-
transaminases and accelerated progression to coma Veh; **P < 0.001 vs. CTRL-Veh; †P < 0.01 vs. ALF-Veh; ¥P < 0.001
stages of encephalopathy. Ammonia concentrations vs. all other groups.

Liver International (2013)

© 2013 John Wiley & Sons A/S. Publishing by John Wiley & Sons Ltd. 5
Endotoxemia during acute liver failure in mice
Fig. 4. ALF mice challenged with LPS manifest further increases in MMP-9 activity in brain compared with ALF mice. Activity of matrix
metalloproteinases 9 (MMP-9) and 2 (MMP-2) was quantified by zymography and expressed as a percentage of control. Data represent
mean ± SEM of n = 8 in each group. *P < 0.05 vs. CTRL-Veh; **P < 0.001 vs. CTRL-Veh; †P < 0.01 vs. ALF-Veh.
Fig. 5. LPS challenge results in IgG extravasation in brain of ALF
mice at coma stages of encephalopathy. IgG (light chain, 25 kDa)
extravasation was quantified by western blot analysis, normalized
vs. b-actin and expressed as a percentage of control. Data represent
mean ± SEM of n = 10 in each group. *P < 0.001 vs. all groups.
were also increased in these animals, consistent with
additional decreases in the capacity for urea synthesis
by injured hepatocytes leading to decreased hepatic
6 © 2013 John Wiley & Sons A/S. Publishing by John Wiley & Sons Ltd.
Chastre et al.
clearance of ammonia. No changes in any of the above
parameters were observed in LPS-treated control mice
indicating that the low dose of LPS used in this study
had no harmful effects per se on the liver in the absence
of AOM.
Treatment of mice with AOM alone resulted in
increased concentrations of the circulating pro-inflam-
matory cytokines TNF-a and IL-6 consistent with a
systemic inflammatory response in these animals that
was comparable to that reported in patients with ALF
owing to toxic liver injury (18) and in other experimen-
tal models of ALF (19) where the precise nature and
magnitude of the cytokine increases may vary depend-
ing upon the nature of the toxin (20).
It has been proposed that liver-derived pro-inflam-
matory cytokines are involved in the pathogenesis of
intracranial hypertension in ALF. This proposal was
based on a report of a patient with ALF resulting from
toxic liver injury awaiting liver transplantation who, on
account of severe uncontrolled intracranial hyperten-
sion, underwent an emergency hepatectomy (21); liver
removal led to a rapid sustained reduction in circulating
pro-inflammatory cytokines and a marked lowering of
intracranial pressure resulting in bridging to successful
liver transplantation.
Results of this study demonstrate that LPS adminis-
tration to AOM mice caused a further two- to three-fold
increase in serum TNF-a and IL-6 and a seven-fold
increase in IL-1b compared with vehicle-treated ALF
mice. In parallel with this increased systemic inflamma-
tory response, LPS treatment of AOM mice resulted in
shortening of the time to appearance of coma defined
as loss of corneal reflex. These findings of increases in
concentration of circulating cytokines in relation to
the precipitation of encephalopathy in ALF add to a
Liver International (2013)
Chastre et al.
growing body of evidence in support of a fundamental
role of systemic inflammation in the pathogenesis of
HE in ALF. For example, previous studies demonstrate
that TNF-a gene polymorphism significantly influences
ALF outcome in both humans and experimental ani-
mals (18, 22), and TNF or IL-1 receptor gene deletion
or plasma TNF-a inhibition delay the progression of
HE in AOM mice (14, 23).
Effects of ALF on BBB integrity and function have
been the subject of considerable debate over the last sev-
eral decades (24). Studies in autopsied brain tissue from
patients with ALF who died in hepatic coma as a result
of acetaminophen hepatotoxicity revealed swelling of as-
trocytes but no clear evidence of structural damage to
the BBB (25), and similar negative findings were
reported in animal models of ALF resulting from galac-
tosamine-induced toxic liver injury (26, 27). On the
other hand, one study in the AOM mouse suggested that
the BBB becomes increasingly permeable to small mole-
cules such as water, as a result of discrete changes in
expression of proteins related to the tight junction (TJ)
assembly of the barrier (7). However, a subsequent
study could not confirm these findings (28). Results of
this study likewise could find no evidence of altered per-
meability to IgG or alterations on expression of TJ-asso-
ciated proteins in AOM-induced ALF even at coma
stages of encephalopathy. On the other hand, when
trace amounts of LPS were administered to mice with
AOM-induced ALF, unequivocal evidence of IgG
extravasation became immediately apparent as mice
became comatose.
Thus, neither AOM-induced ALF alone nor LPS
treatment alone was sufficient to cause increased BBB
permeability to IgG, but the two insults together did so
unequivocally, suggestive of a synergistic mechanism.
These findings suggest the possibility that factors that
occur in ALF may be acting in concert with pro-inflam-
matory mechanisms. One such factor is ammonia. Arte-
rial ammonia concentrations are excellent predictors of
encephalopathy and intracranial hypertension in
patients with ALF resulting from toxic liver injury (29,
30). In this study, LPS treatment of ALF mice resulted
in a three-fold increase in circulating ammonia. Con-
centrations of ammonia comparable to those reported
in these studies and in LPS-treated ALF mice in this
study are known to facilitate ammonia diffusion into
cerebrovascular endothelial cells (31), an action that is
consistent with increased BBB permeability. Moreover,
millimolar concentrations of ammonia inhibit glucose
oxidation in brain mitochondria (32) resulting in accu-
mulation of lactate. Brain lactate concentrations as well
as lactate synthesis rates correlate significantly with HE
grade in experimental ALF (33). Moreover, exposure of
both astrocytes and microglial cells to lactate leads to
the release of pro-inflammatory cytokines (34) provid-
ing a possible explanation of synergism between ammo-
nia and neuroinflammation in relation to the
pathogenesis of encephalopathy in ALF.
Liver International (2013)
© 2013 John Wiley & Sons A/S. Publishing by John Wiley & Sons Ltd.
Endotoxemia during acute liver failure in mice
Matrix metalloproteinase-9 has been implicated in
the pathogenesis of BBB disruption in a range of neuro-
logical disorders including infectious encephalitis (35),
stroke (36) and traumatic brain injury (37). Evidence
for a role of MMP-9 in the pathogenesis of BBB perme-
ability changes in ALF has, more recently, been reported
(11) where it was suggested that the injured liver, not
the brain, was the source of the increased MMP-9. A
subsequent report, however, disputed the interpretation
of these findings (38). Results of this study clearly dem-
onstrate significant and selective increases in the activity
of MMP-9 in the brains of AOM mice, an effect that
started early in the progression of HE (prior to the onset
of coma). Moreover, in the presence of LPS, a further
two- to three-fold increase in MMP-9 activity was
observed in brain as the animals became comatose.
Together with the findings of extravasation of IgG, these
results represent the first direct evidence implicating
alterations of the BBB in the pathogenesis of HE precip-
itated by infection/inflammation in ALF and suggest
that up-regulation of brain MMP-9 activity is a possible
mechanism involved.
The findings of a lack of significant alterations in
expression of the tight junction-related proteins occlu-
din and ZO-1 in either group confirm, on one hand,
those of an earlier study in the same animal model
(7) and suggest, on the other hand, that BBB disrup-
tion in ALF+LPS mice did not involve tight junction
digestion. Other mechanisms, such as the degradation
of the basal lamina by MMP-9, might participate in
the increase in BBB permeability observed in
ALF+LPS mice. Interestingly, cerebral MMP-9 activity
was specifically induced, the expression and activation
of which is known to be modulated by inflammatory
stimuli such as TNF-a and IL-1b (39, 40), which are
also involved in the CNS complications of ALF (14,
23, 41, 42).
In conclusion, the aim of this study was the establish-
ment of an in vivo model preparation for the study of
basic mechanisms implicated in the pathogenesis of
encephalopathy precipitated by infection/inflammation
in ALF. Time to severe encephalopathy, in particular,
was chosen as a robust clinical endpoint for the studies.
Results demonstrate that the superposition of ALF
resulting from AOM-induced toxic liver injury together
with endotoxemia (LPS) results in worsening of both
hepatic function and hyperammonaemia as well as
breakdown of the BBB and precipitation of HE. These
findings represent the first direct evidence for BBB per-
meability changes in ALF owing to toxic liver injury and
suggest a role for these changes in the pathogenesis of
encephalopathy in this condition.
Acknowledgement
Financial support: Studies from the authors’ laboratory
were funded by operating grants from The Canadian
Institutes of Health Research.
7
Endotoxemia during acute liver failure in mice
References
1. Rolando N, Wade J, Davalos M, et al. The systemic
inflammatory response syndrome in acute liver failure.
Hepatology 2000; 32: 734–9.
2. Vaquero J, Polson J, Chung C, et al. Infection and the pro-
gression of hepatic encephalopathy in acute liver failure.
Gastroenterology 2003; 125: 755–64.
3. Iwai H, Nagaki M, Naito T, et al. Removal of endotoxin
and cytokines by plasma exchange in patients with acute
hepatic failure. Crit Care Med 1998; 26: 873–6.
4. Galanos C, Freudenberg MA, Reutter W. Galactosamine-
induced sensitization to the lethal effects of endotoxin.
Proc Natl Acad Sci USA 1979; 76: 5939–43.
5. Takada Y, Ishiguro S, Fukunaga K, et al. Increased intra-
cranial pressure in a porcine model of fulminant hepatic
failure using amatoxin and endotoxin. J Hepatol 2001; 34:
825–31.
6. Belanger M, Cote J, Butterworth RF. Neurobiological
characterization of an azoxymethane mouse model of
acute liver failure. Neurochem Int 2006; 48: 434–40.
7. Shimojima N, Eckman CB, McKinney M, et al. Altered
expression of zonula occludens-2 precedes increased
blood-brain barrier permeability in a murine model of ful-
minant hepatic failure. J Invest Surg 2008; 21: 101–8.
8. Wispelwey B, Lesse AJ, Hansen EJ, Scheld WM. Haemo-
philus influenzae lipopolysaccharide-induced blood brain
barrier permeability during experimental meningitis in the
rat. J Clin Invest 1988; 82: 1339–46.
9. Moor AC, de Vries HE, de Boer AG, Breimer DD. The
blood-brain barrier and multiple sclerosis. Biochem Phar-
macol 1994; 47: 1717–24.
10. de Vries HE, Blom-Roosemalen MC, van Oosten M,
et al. The influence of cytokines on the integrity of the
blood-brain barrier in vitro. J Neuroimmunol 1996; 64:
37–43.
11. Nguyen JH, Yamamoto S, Steers J, et al. Matrix metallo-
proteinase-9 contributes to brain extravasation and edema
in fulminant hepatic failure mice. J Hepatol 2006; 44:
1105–14.
12. Warren HS. Editorial: mouse models to study sepsis syn-
drome in humans. J Leukoc Biol 2009; 86: 199–201.
13. Beauchesne E, Desjardins P, Butterworth RF, Hazell AS.
Up-regulation of caveolin-1 and blood-brain barrier
breakdown are attenuated by N-acetylcysteine in thiamine
deficiency. Neurochem Int 2010; 57: 830–7.
14. Bemeur C, Qu H, Desjardins P, Butterworth RF. IL-1 or
TNF receptor gene deletion delays onset of encephalopa-
thy and attenuates brain edema in experimental acute liver
failure. Neurochem Int 2010; 56: 213–5.
15. Bemeur C, Vaquero J, Desjardins P, Butterworth RF.
N-acetylcysteine attenuates cerebral complications of non-
acetaminophen-induced acute liver failure in mice: antiox-
idant and anti-inflammatory mechanisms. Metab Brain
Dis 2010; 25: 241–9.
16. Rangroo TV, Thrane AS, Chanag J, et al. Real-time analy-
sis of microglial activation and motility in hepatic and
hyperammonemic encephalopathy. Neuroscience 2012;
220: 247–55.
17. Matkowskyj KA, Marrero JA, Carroll RE, et al. Azoxyme-
thane-induced fulminant hepatic failure in C57BL/6J mice:
characterization of a new animal model. Am J Physiol
1999; 277: G455–62.
8 © 2013 John Wiley & Sons A/S. Publishing by John Wiley & Sons Ltd.
Chastre et al.
18. Bernal W, Donaldson P, Underhill J, Wendon J, Williams
R. Tumor necrosis factor genomic polymorphism and out-
come of acetaminophen (paracetamol)-induced acute liver
failure. J Hepatol 1998; 29: 53–9.
19. Streetz K, Leifeld L, Grundmann D, et al. Tumor necrosis
factor alpha in the pathogenesis of human and murine ful-
minant hepatic failure. Gastroenterology 2000; 119: 446–
60.
20. Bemeur C, Butterworth R F. Liver-brain proinflammatory
signalling in acute liver failure: Role in the pathogenesis of
hepatic encephalopathy and brain edema. Metab Brain Dis
2013; 28: 145–50.
21. Jalan R, Pollok A, Shah SH, Madhavan K, Simpson KJ.
Liver derived pro-inflammatory cytokines may be impor-
tant in producing intracranial hypertension in acute liver
failure. J Hepatol 2002; 37: 536–8.
22. Tsuchiya N, Tokushige K, Yamaguchi N, et al. Influence
of TNF gene polymorphism in patients with acute and ful-
minant hepatitis. J Gastroenterol 2004; 39: 859–66.
23. Chastre A, Belanger M, Beauchesne E, et al. Inflammatory
cascades driven by tumor necrosis factor-alpha play a
major role in the progression of acute liver failure and its
neurological complications. PLoS ONE 2012; 7: e49670.
24. Blei AT, Larsen FS. Pathophysiology of cerebral edema in
fulminant hepatic failure. J Hepatol 1999; 31: 771–6.
25. Kato M, Hughes RD, Keays RT, Williams R. Electron
microscopic study of brain capillaries in cerebral edema
from fulminant hepatic failure. Hepatology 1992; 15:
1060–6.
26. Gove CD, Hughes RD, Ede RJ, Williams R. Regional cere-
bral edema and chloride space in galactosamine-induced
liver failure in rats. Hepatology 1997; 25: 295–301.
27. Traber PG, Dal Canto M, Ganger DR, Blei AT. Electron
microscopic evaluation of brain edema in rabbits with
galactosamine-induced fulminant hepatic failure:
ultrastructure and integrity of the blood-brain barrier.
Hepatology 1987; 7: 1272–7.
28. Bemeur C, Chastre A, Desjardins P, Butterworth RF. No
changes in expression of tight junction proteins or blood-
brain barrier permeability in azoxymethane-induced
experimental acute liver failure. Neurochem Int 2010; 56:
3.
29. Clemmesen JO, Larsen FS, Kondrup J, Hansen BA, Ott P.
Cerebral herniation in patients with acute liver failure is
correlated with arterial ammonia concentration. Hepatolo-
gy 1999; 29: 648–53.
30. Tofteng F, Hauerberg J, Hansen BA, et al. Persistent
arterial hyperammonemia increases the concentration of
glutamine and alanine in the brain and correlates with
intracranial pressure in patients with fulminant hepatic
failure. J Cereb Blood Flow Metab 2006; 26: 21–7.
31. Duchini A, Govindarajan S, Santucci M, Zampi G, Hof-
man FM. Effects of tumor necrosis factor-alpha and inter-
leukin-6 on fluid-phase permeability and ammonia
diffusion in CNS-derived endothelial cells. J Investig Med
1996; 44: 474–82.
32. Lai JC, Cooper AJ. Brain alpha-ketoglutarate dehydroge-
nase complex: kinetic properties, regional distribution,
and effects of inhibitors. J Neurochem 1986; 47: 1376–86.
33. Zwingmann C, Chatauret N, Leibfritz D, Butterworth RF.
Selective increase of brain lactate synthesis in experimental
acute liver failure: results of a [H-C] nuclear magnetic
resonance study. Hepatology 2003; 37: 420–8.
Liver International (2013)
Chastre et al.
34. Andersson AK, Ronnback L, Hansson E. Lactate induces
tumour necrosis factor-alpha, interleukin-6 and interleu-
kin-1beta release in microglial- and astroglial-enriched
primary cultures. J Neurochem 2005; 93: 1327–33.
35. Sporer B, Koedel U, Paul R, et al. Human immunodefi-
ciency virus type-1 Nef protein induces blood-brain
barrier disruption in the rat: role of matrix metallopro-
teinase-9. J Neuroimmunol 2000; 102: 125–30.
36. Fujimura M, Gasche Y, Morita-Fujimura Y, et al. Early
appearance of activated matrix metalloproteinase-9 and
blood-brain barrier disruption in mice after focal cerebral
ischemia and reperfusion. Brain Res 1999; 842: 92–100.
37. Rosenberg GA. Matrix metalloproteinases in brain injury.
J Neurotrauma 1995; 12: 833–42.
38. Palenzuela L, Mas A, Montaner J, Cordoba J. Matrix
metalloproteinase-9 in fulminant hepatic failure.
Hepatology 2010; 51: 1475–6; author reply 76.
Liver International (2013)
© 2013 John Wiley & Sons A/S. Publishing by John Wiley & Sons Ltd.
Endotoxemia during acute liver failure in mice
39. Quagliarello V, Scheld WM. Bacterial meningitis: patho-
genesis, pathophysiology, and progress. N Engl J Med
1992; 327: 864–72.
40. Sharief MK, Thompson EJ. In vivo relationship of tumor
necrosis factor-alpha to blood-brain barrier damage in
patients with active multiple sclerosis. J Neuroimmunol
1992; 38: 27–33.
41. Wright G, Shawcross D, Olde Damink SW, Jalan R. Brain
cytokine flux in acute liver failure and its relationship with
intracranial hypertension. Metab Brain Dis 2007; 22: 375–
88.
42. Jiang W, Desjardins P, Butterworth RF. Cerebral inflam-
mation contributes to encephalopathy and brain edema in
acute liver failure: protective effect of minocycline. J Neu-
rochem 2009; 109: 485–93.
9