Norepinephrine regulates hepatic innate immune system in leptin-deficient mice with nonalcoholic steatohepatitis

Abstract

It is not known why natural killer T (NKT) cells, which modulate liver injury by regulating local cytokine production, are reduced in leptin-deficient ob/ob mice. NKT cells express adrenoceptors. Thus, we hypothesize that the low norepinephrine (NE) activity of ob/ob mice promotes depletion of liver NKT cells, thereby sensitizing ob/ob livers to lipopolysaccharide (LPS) toxicity. To evaluate this hypothesis, hepatic NKT cells were quantified in wild-type mice before and after treatment with NE inhibitors, and in dopamine beta-hydroxylase knockout mice (which cannot synthesize NE) and ob/ob mice before and after 4 weeks of NE supplementation. Decreasing NE activity consistently reduces liver NKT cells, while increasing NE has the opposite effect. Analysis of hepatic and thymic NKT cells in mice of different ages demonstrate an age-related accumulation of hepatic NKT cells in normal mice, while liver NKT cells become depleted after birth in ob/ob mice, which have increased apoptosis of hepatic NKT cells. NE treatment inhibits apoptosis and restores hepatic NKT cells. In ob/ob mice with reduced hepatic NKT cells, hepatic T and NKT cells produce excessive T helper (Th)-1 proinflammatory cytokines and the liver is sensitized to LPS toxicity. NE treatment decreases Th-1 cytokines, increases production of Th-2 cytokines, and reduces hepatotoxicity. Studies of CD1d-deficient mice, which lack the receptor required for NKT cell development, demonstrate that they are also unusually sensitive to LPS hepatotoxicity. In conclusion, low NE activity increases hepatic NKT cell apoptosis and depletes liver NKT cells, promoting proinflammatory polarization of hepatic cytokine production that sensitizes the liver to LPS toxicity.

Full text

Norepinephrine Regulates Hepatic Innate Immune
System in Leptin-Deficient Mice With Nonalcoholic
Steatohepatitis
Zhiping Li,1Jude A. Oben,1Shiqi Yang,1Huizhi Lin,1Elizabeth A. Stafford,1Mark J. Soloski,1
Steven A. Thomas,2and Anna Mae Diehl3
It is not known why natural killer T (NKT) cells, which modulate liver injury by regulating
local cytokine production, are reduced in leptin-deficient ob/ob mice. NKT cells express
adrenoceptors. Thus, we hypothesize that the low norepinephrine (NE) activity of ob/ob
mice promotes depletion of liver NKT cells, thereby sensitizing ob/ob livers to lipopolysac-
charide (LPS) toxicity. To evaluate this hypothesis, hepatic NKT cells were quantified in
wild-type mice before and after treatment with NE inhibitors, and in dopamine
␤
-hydrox-
ylase knockout mice (which cannot synthesize NE) and ob/ob mice before and after 4 weeks
of NE supplementation. Decreasing NE activity consistently reduces liver NKT cells, while
increasing NE has the opposite effect. Analysis of hepatic and thymic NKT cells in mice of
different ages demonstrate an age-related accumulation of hepatic NKT cells in normal mice,
while liver NKT cells become depleted after birth in ob/ob mice, which have increased
apoptosis of hepatic NKT cells. NE treatment inhibits apoptosis and restores hepatic NKT
cells. In ob/ob mice with reduced hepatic NKT cells, hepatic T and NKT cells produce
excessive T helper (Th)-1 proinflammatory cytokines and the liver is sensitized to LPS
toxicity. NE treatment decreases Th-1 cytokines, increases production of Th-2 cytokines,
and reduces hepatotoxicity. Studies of CD1d-deficient mice, which lack the receptor re-
quired for NKT cell development, demonstrate that they are also unusually sensitive to LPS
hepatotoxicity. In conclusion, low NE activity increases hepatic NKT cell apoptosis and
depletes liver NKT cells, promoting proinflammatory polarization of hepatic cytokine pro-
duction that sensitizes the liver to LPS toxicity. (HEPATOLOGY 2004;40:434–441.)
Obesity is strongly associated with nonalcoholic
fatty liver disease (NAFLD). Fatty livers are un-
usually susceptible to injury induced by a sec-
ondary inflammatory stress, including that evoked by
exposure to endogenous, intestine-derived lipopolysac-
carhide (LPS).
1
More serious liver injury (steatohepatitis)
results, and this eventually leads to cirrhosis in some in-
dividuals. Because the transition from steatosis to steato-
hepatitis dramatically increases the risk of developing
liver-related morbidity and mortality, it is important to
understand why fatty livers are so vulnerable to inflam-
matory stress.
To address this issue, our laboratory has been studying
genetically obese, leptin-deficient ob/ob mice.
2
Similar to
obese humans who are usually hyperleptinemic, ob/ob
mice are insulin-resistant, have fatty livers, and are exquis-
itely sensitive to LPS-induced liver injury.
3
Previously, we
noted that certain populations of liver lymphocytes—spe-
cifically natural killer T (NKT) cells—are selectively re-
duced in the livers of ob/ob mice and suggested that
hepatic NKT cell depletion may underlie fatty liver vul-
nerability to LPS in these animals.
4
NKT cells are com-
ponents of the innate immune system. These cells
originate in the thymus but predominately accumulate in
Abbreviations: NKT cells, natural killer T cells; NE, norepinephrine; LPS,
lipopolysaccharide; Th, T helper; NAFLD, nonalcoholic fatty liver disease; DBH,
dopamine
␤
-hydroxylase; HMNC, hepatic mononuclear cell; IL, interleukin;
6-OHDA, 6-hydroxydopamine; TNF-
␣
, tumor necrosis factor
␣
; IFN-
␥
, inter-
feron
␥
; ALT, alanine aminotransferase.
From the 1Department of Medicine, Johns Hopkins University, Baltimore, MD;
the 2Department of Pharmacology, University of Pennsylvania, Philadelphia, PA;
and the 3Department of Medicine, Duke University, Durham, NC.
Received February 2, 2004; accepted April 28, 2004.
This work was partially supported by RO1 DK 53792 from National Institute of
Diabetes and Digestive and Kidney Diseases (NIDDK) (A. M. D.), National
Institutes of Health training grant T32DK07632 (Z. L.), and the American Gas-
troenterological Association (AGA) Fellowship/Faculty Transition Award (Z. L.).
Address reprint requests to: Anna Mae Diehl, M.D., Duke University, Genome
Science Research Building-1, Suite 1073, Box 3256, Research Drive, Durham, NC
27710. E-mail: diehl004@mc.duke.edu; fax: (919) 684-4183.
Copyright © 2004 by the American Association for the Study of Liver Diseases.
Published online in Wiley InterScience (www.interscience.wiley.com).
DOI 10.1002/hep.20320
434

the liver, where they regulate local proinflammatory (T
helper [Th]-1) and anti-inflammatory (Th-2) cytokine
production by other mononuclear cells.
5
Certain types of
liver injury may result from selective reductions in NKT
cell populations. For example, infection with Propi-
onibacterium acnes is thought to sensitize normal livers to
subsequent LPS-induced injury by reducing hepatic
NKT cells.
6
Leptin is now known to have potent immunomodula-
tory actions.
7
Interactions between leptin and its recep-
tors on immune cells regulate immune cell functions,
8,9
including phagocytosis
10
and cytokine production.
11
Leptin deficiency also sensitizes T cells to corticosteroid-
induced apoptosis, which promotes thymic atrophy in
ob/ob mice.
12
In addition, leptin deficiency profoundly
alters the hypothalamic–pituitary–adrenal axis, increasing
certain stress-related factors (e.g., corticosteroids)
13
while
decreasing others (e.g., norepinephrine [NE]).
14
It is be-
coming increasingly apparent that changes in such leptin-
regulated, neurohumoral factors directly mediate many
features of leptin deficiency. For example, reduced NE
causes the hyperostosis and inhibited adipocyte lipolysis
that occur in leptin-deficient mice.
15
Moreover, targeted
disruption of leptin receptor genes in neurons reproduces
many aspects of the ob/ob phenotype, strongly suggesting
that neuronal factors are necessary to relay leptin-initiated
signals to many other cells.
16
Previous studies have found that the sympathetic ner-
vous system regulates NKT cells.
17
We hypothesize that
reduced NE inhibits the hepatic accumulation of NKT
cells in leptin-deficient mice. If NE is indeed a major
proximal regulator of hepatic NKT cell populations, then
changes in NE activity may alter hepatic NKT cell num-
bers and influence hepatic cytokine production indepen-
dently of leptin. This has important implications, because
most obese humans with NAFLD are not leptin-deficient;
consequently, it has been unclear if mechanisms that me-
diate NAFLD pathogenesis in ob/ob mice have more gen-
eral relevance. In the present study, we address the
importance of NE in hepatic NKT cell regulation and
attempt to determine if hepatic NKT cell depletion causes
Th-1 polarization of hepatic cytokine-producing cells and
enhances sensitivity to liver injury.
Materials and Methods
Animal Experiments. Adult (8- to 10-week-old) and
young (3- and 6-week-old) male C57BL6 ob/ob mice,
their lean litter mates, and wild-type C57BL6 mice were
purchased from Jackson Laboratories (Bar Harbor, ME).
CD1d⫺/⫺C57BL6 mice were a gift from Dr. Albert Ben-
delac (University of Chicago, Chicago, IL). Dopamine
␤
-hydroxylase–deficient C57BL6 mice (DBH⫺/⫺) and
their heterozygous litter mates (DBH⫹/⫺) were from the
colony that has been generated by Steven Thomas’ labo-
ratory.
18
All mice were maintained in a temperature- and
light-controlled facility and were permitted ad libitum
consumption of water and pellet chow.
Young (3- and 6-week-old) ob/ob mice and their lean
litter mates were used to evaluate developmental changes
in thymic and hepatic mononuclear cell (HMNC) popu-
lations. Older (8- to 10-week-old) ob/ob mice and their
lean litter mates were treated with vehicle (pyrogen-free
saline), NE (Sigma, St. Louis, MO, 2.5 mg/kg/d via Alzet
minipumps [Durect, Cupertino, CA] for 3 weeks), or
interleukin (IL) 15 (a gift from Immunex, Seattle, WA,
10
␮
g/d intraperitoneally for 1 week). In parallel studies,
some DBH⫺/⫺mice were also treated with vehicle and
NE. Wild-type C57BL6 mice (10 to 12 weeks old) were
treated with prazosin (Sigma, 0.05 mg/mL in drinking
water) or 6-hydroxydopamine (6-OHDA) (Sigma, 100
mg/kg) for 4 weeks. 6-OHDA was injected intraperitone-
ally daily for 5 consecutive days to induce chemical sympa-
thectomy. Thereafter, mice received 6-OHDA injections
three times per week for the remainder of the 4-week
treatment period to ensure continued sympathectomy.
19
Control mice received either vehicle in drinking water
(for the prazosin studies) or via intraperitoneal injection
(for the 6-OHDA studies). CD1d⫺/⫺C57BL6 mice and
wild-type C57BL6 mice were injected intraperitoneally
with a single dose of Eschericia coli LPS (Sigma, 50
␮
g/
mouse) and then killed after 0, 1.5, or 6 hours to obtain
serum and liver tissue.
All animal experiments fulfilled National Institutes of
Health and Johns Hopkins University criteria for the hu-
mane treatment of laboratory animals.
Liver Mononuclear Cell Isolation and Cell Surface
Labeling. HMNCs were isolated and labeled using a mi-
nor modification of the method we described previously.
4
Anti-mouse NK1.1-PE, CD3-FITC, CD4-APC, CD8-
PerCP, Annexin-V-PE, and 7-AAD were obtained from
Pharmingen (San Diego, CA). After surface labeling,
HMNCs were evaluated using flow cytometry (Becton
Dickenson, Palo Alto, CA). Data were analyzed using
Cell Quest software (Becton Dickenson).
Liver Mononuclear Cell Intracellular Cytokines
Labeling. After isolation, HMNCs were incubated with
phorbol 1,2-myristate 1,3-acetate (Sigma, 50 ng/mL),
ionomycin (Sigma, 500 ng/mL) and GolgiPlug (Pharm-
ingen, 1
␮
L/mL). Cells were labeled with surface anti-
body as decribed above and then permeablized with
Cytoperm/Cytofix (Pharmingen) according to the man-
ufacturer’s instructions. After permeablization, cells were
further labeled for intracellular cytokines such as anti–
HEPATOLOGY, Vol. 40, No. 2, 2004 LI ET AL. 435

mouse tumor necrosis factor
␣
(TNF-
␣
) and interferon
␥
(IFN-
␥
) (Pharmingen). After incubation cells were eval-
uated using flow cytometry. Data were analyzed as de-
scribed above.
Liver Mononuclear Cell IL-4 Assay. After isolation,
HMNCs (2 ⫻105/well) were incubated with anti–CD3
mAb (Pharmingen, 10
␮
g/mL) overnight at 37°C. IL-4
concentration in the medium was determined via en-
zyme-linked immunosorbent assay using mouse recombi-
nant IL-4 as standard according the manufacturer’s
instructions (Biosource, Camarillo, CA).
Serum Alanine Aminotransferase Levels. Alanine
aminotransferase (ALT) levels were measured with a mul-
tichannel autoanalyzer in the Clinical Chemistry Labora-
tory of the Johns Hopkins University Department of
Comparative Medicine.
Statistical Analysis. All values are expressed as the
mean ⫾SD. The group means were compared via
ANOVA using Microsoft Excel (Microsoft, Redmond,
WA).
Results
Selective Depletion of Hepatic, But Not Thymic,
NKT Cells During Leptin Deficiency. Because NKT
cells originate in the thymus and migrate to the liver dur-
ing development,
5
we compared thymic and hepatic
NKT cell populations in 3- and 6-week-old ob/ob mice
and their lean littermates to understand when NKT cells
are depleted in congenital leptin deficiency. As reported
by others,
12
we noted that ob/ob mice develop premature
thymic atrophy, which is reflected by dramatic decreases
in total thymic mononuclear cell numbers (Fig. 1A). Sur-
prisingly, the NKT cells are relatively spared during thy-
mic involution in ob/ob mice such that the remaining
thymic mononuclear cells in these mice are relatively en-
riched with NKT cells (Fig. 1B). In contrast, there is a
selective depletion of NKT cells in the livers of ob/ob
mice during the same period. Compared with lean control
mice, total hepatic mononuclear cell numbers are not de-
creased in ob/ob mice at either 3 or 6 weeks after birth
(Fig. 1C). However, in the ob/ob group, hepatic NKT
cell populations progressively decline during this period
(Fig. 1D). These findings suggested to us that extrathymic
factors are predominately responsible for the lower levels
of hepatic NKT cells in leptin-deficient mice.
Sympathetic Neurotransmitters Regulate Hepatic
NKT Cell Populations. As discussed earlier, NE activity
is reduced during leptin deficiency.
14
NKT cells express
adrenoceptors.
20
Furthermore, work by Minagawa and
colleagues suggests that sympathetic neurotransmitters
regulate the accumulation of NKT cells in the liver after
partial hepatectomy.
17
However, it is not known how
important sympathetic neurotransmitters are for the gen-
eral maintainence of hepatic NKT cell populations. To
address this question, we treated wild-type (leptin-suffi-
cient) mice with prazosin, an
␣
-adrenoceptor blocker, or
6-OHDA to induce chemical sympathectomy. Both
agents significantly decreased hepatic NKT cell popula-
tions in normal mice (Fig. 2A).
To determine whether or not the effects of
␣
-adreno-
ceptor blockade and chemical sympathectomy are due to
specific inhibition of NE activity, we studied DBH⫺/⫺
mice, which cannot produce NE.
18
Hepatic NKT cells are
also significantly decreased in DBH⫺/⫺mice (Fig. 2B)
compared with their heterozygous littermates. Hepatic
NKT cells also increase after supplementing NE in
DBH⫺/⫺mice (see Fig. 2B). Therefore, either acquired or
congenital deficiency of NE causes depletion of liver
NKT cells. These findings demonstrate that NE is criti-
cally important for normal maintenance of hepatic NKT
cell populations.
NE Treatment Increases Hepatic NKT Cell Popu-
lation in Leptin-Deficient ob/ob Mice. Leptin defi-
ciency is known to induce multiple hormonal, metabolic,
and immunological abnormalities. Therefore, although
NE is a major regulator of hepatic NKT cells in leptin-
sufficient mice, other factors may be more important in
maintaining these cells in the livers of leptin-deficient
mice. To address this issue, we treated ob/ob mice with
Fig. 1. Selective depletion of hepatic NKT cells during development.
(A) Thymic mononuclear cells were isolated from 3- and 6-week-old
ob/ob and lean mice. (B) Flow cytometry analysis quantified thymic
CD4⫹NKT cell subpopulations. (C) HMNCs were also isolated from the
livers of 3- and 6-week-old ob/ob and lean mice. (D) Hepatic CD4⫹NKT
cell subpopulations were quantified using flow cytometry. In each exper-
iment, cells were pooled from paired groups of mice (6 –9 mice per
group). Experiments were repeated twice. Mean ⫾SD results of dupli-
cate experiments are shown. *P⬍.01, **P⬍.05 indicate differences
between ob/ob and lean groups. TNMC, thymic mononuclear cells; NKT
cell, natural killer cells; HMNC, hepatic mononuclear cells.
436 LI ET AL. HEPATOLOGY, August 2004

NE or saline vehicle. NE significantly increased hepatic
NKT cells in leptin-deficient mice (Fig. 3), demonstrat-
ing that this sympathetic neurotransmitter regulates he-
patic NKT cell populations independently of leptin.
Further analysis of different NK cell–specific surface an-
tigens is planned to define the heterogeneity of the NKT cell
population in ob/ob livers before and after NE treatment.
Such information will be helpful in determining if various
NKT cell subsets differ in their requirements for NE.
NE Treatment Reduces Hepatic NKT Cell Apopto-
sis in Leptin-Deficient ob/ob Mice. To gain insight into
the mechanisms through which NE increases hepatic
NKT cells, we assessed the effects of NE on NKT cell
apoptosis. Hepatic expression of TNF-
␣
, a factor that
induces NKT cell apoptosis,
21
is known to be increased in
ob/ob mice.
22
Therefore, we suspected that NKT cell ap-
optosis might be increased in ob/ob livers. To assess this
possibility, we evaluated hepatic NKT cell apoptosis us-
ing Annexin V. We found that, as predicted, hepatic
NKT cell apoptosis is increased significantly in ob/ob
mice. Moreover, 3 weeks of NE treatment decreased he-
patic NKT cell apoptotic activity to normal levels (Fig. 4).
Compared with NE, IL-15, another factor that increases
NKT cells,
11
has much less of an inhibitory effect on
hepatic NKT cell apoptosis (see Fig. 4).
NE Treatment Reverses Hepatic Proinflammatory
Cytokine Production During Leptin Deficiency. The
livers of leptin-deficient mice are unusually sensitive to
LPS-induced injury, a process that is mediated by proin-
flammatory cytokines such as TNF-
␣
and IFN-
␥
. IFN-
␥
is known to sensitize hepatocytes to TNF-
␣
killing.
23
Studies with TNF-
␣
–neutralizing antibodies demon-
strate that TNF-
␣
is required for LPS liver injury.
24
How-
ever, IFN-
␥
sensitization to TNF-
␣
is also critically
important, because mice that are genetically deficient in
IFN-
␥
are completely protected from LPS hepatotoxicity
despite persistent TNF-
␣
expression.
25
NKT cells pro-
duce both IFN-
␥
and IL-4.
5
While the former exacerbates
TNF-
␣
toxicity, the latter is a key inducer of anti-inflam-
matory (Th-2) cytokines, which generally attenuate the
toxic effects of TNF-
␣
.
26
Therefore, it is difficult to pre-
dict the ultimate effects of hepatic NKT cell depletion on
hepatic cytokine production and LPS sensitivity.
To address the first issue, we treated ob/ob mice with
NE or vehicle, isolated hepatic mononuclear cells, and
measured intracellular cytokines. Results from both
ob/ob groups were also compared with those of lean con-
trol mice. The production of IFN-
␥
and TNF-
␣
, in-
creased significantly in total liver mononuclear cells from
ob/ob mice compared with lean controls. These differ-
ences reflect increases in Th-1 cytokine production by
several different cell populations, as demonstrated by in-
creased IFN-
␥
and/or TNF-
␣
expression in hepatic T
Fig. 3. Effect of NE on hepatic NKT cell content of ob/ob mice. ob/ob
mice were treated with NE or sterile, pyrogen-free saline. At the end of the
treatment period, liver mononuclear cells were harvested and analyzed
using flow cytometry. Representative flow cytometry data are displayed in
the upper panel. Mean ⫾SD results from duplicate experiments are
graphed. **P⬍.05, *P⬍.01 indicate differences between the
saline-treated ob/ob mice and lean controls. §§P⬍.05 indicate differ-
ences between NE-treated ob/ob mice and saline-treated ob/ob mice.
NE, norepinephrine; NKT cells, natural killer T cells.
Fig. 2. Effect of adrenoreceptor blocker, chemical sympathectomy,
and NE on hepatic NKT cells. (A) Wild-type C57BL6 mice were treated
with prazosin or 6-OHDA. Total liver mononuclear cells were harvested
and analyzed using flow cytometry. (B) Total liver mononuclear cells were
isolated from DBH⫺/⫺mice, which cannot produce NE, DBH⫺/⫺mice
supplemented with NE,and their heterozygous littermates (DBH⫹/⫺).
Cells were analyzed using flow cytometry. Representative flow cytometry
data are displayed in the upper panel. Mean ⫾SD results of triplicate
experiments are graphed. §P⬍.01 indicates difference between
DBH⫺/⫺and DBH⫹/⫺controls. §§P⬍.05 indicates difference between
DBH⫺/⫺mice treated with NE and vehicle. prz, prazosin; 6-OHDA,
6-hydroxydopamine; DBH, dopamine
␤
-hydroxylase; NE, norepinephrine;
NKT cells, natural killer T cells.
HEPATOLOGY, Vol. 40, No. 2, 2004 LI ET AL. 437

cells and NK cells (Fig. 5A). Treatment with doses of NE
that restores hepatic NKT cell numbers also reduces
proinflammatory cytokine production by all of the he-
patic mononuclear cell populations that we evaluated
(Fig. 5B). On the other hand, the production of IL-4, an
anti-inflammatory Th-2 cytokine, is decreased in total
liver mononuclear cells from ob/ob mice (Fig. 5C). Treat-
ment with NE significantly increases IL-4 production (see
Fig. 5C).
CD1dⴚ/ⴚMice Are More Susceptible to LPS-In-
duced Injury. The previous studies demonstrate that in
ob/ob mice, decreases in hepatic NKT cell populations
are accompanied by Th-1 polarization of other cytokine-
producing cells in the liver. Because proinflammatory cy-
tokines mediate LPS liver injury, the latter suggests a
mechanism through which hepatic NKT cell depletion
may promote vulnerability to LPS hepatotoxicity in lep-
tin-deficient mice. However, as discussed earlier, leptin-
deficient ob/ob mice have multiple immunological
abnormalities.
7
Therefore, to clarify the significance of
NKT cell depletion as a vulnerability factor for LPS hep-
atotoxicity, we evaluated otherwise normal mice that were
genetically deficient in CD1d, a Class I–like molecule
that is required for the positive selection of certain NKT
cell populations during development.
27
Because both
ob/ob mice and CD1d⫺/⫺mice are relatively depleted of
hepatic NKT cells,
4,28
we predicted that CD1d⫺/⫺mice
would behave like ob/ob mice and exhibit more sensitivity
to LPS toxicity than control mice that have normal levels
of hepatic NKT cells. To test this hypothesis, we treated
CD1d⫺/⫺mice with low doses of LPS that are generally
well-tolerated by normal mice. After receiving LPS,
CD1d⫺/⫺mice exhibited significantly more liver injury
than control mice, as reflected by two- to threefold higher
serum levels of ALT (Fig. 6A). However, CD1d⫺/⫺mice
appear to be less sensitive to LPS toxicity than ob/ob mice,
which have 10-fold greater ALT levels than LPS-treated
controls at the same time point (Fig. 6B). This suggests
that ob/ob mice have additional factors that enhance their
vulnerability to LPS hepatoxicity. Our preliminary com-
parison of other liver mononuclear cell populations in
Fig. 4. Effect of NE and IL-15 on hepatic NKT cell apoptosis. ob/ob
mice were treated with vehicle, NE, or human recombinant IL-15. Hepatic
mononuclear cells were isolated from these mice and from their lean
littermates. Flow cytometry was used to identify different mononuclear
cell subpopulations, and NKT cell apoptosis was measured using An-
nexin-V staining with concurrent incubation of 7-AAD to assess cell
necrosis. Apoptotic NKT cells are defined as Annexin-V⫹/7-AAD⫺of NKT
cells. Representative flow cytometry data are displayed in the upper
panel. Mean ⫾SD results of duplicate experiments are graphed. *P⬍
.01 indicates difference between lean and ob/ob mice. §§P⬍.05, §P⬍
.01 indicates difference between NE- and IL-15–treated ob/ob mice and
control ob/ob mice. IL-15, interleukin 15; NE, norepinephrine; NKT cells,
natural killer T cells.
Fig. 5. Effect of NE on hepatic cytokine profiles. IFN-
␥
and TNF-
␣
were
measured as intracellular cytokines using flow cytometry. IL-4 was
measured in the media via enzyme-linked immunosorbent assay. The
results were compared between lean mice, ob/ob mice treated with
vehicle, and ob/ob mice treated with NE. Mean ⫾SD results of duplicate
experiments are graphed. (A) IFN-
␥
. (B) TNF-
␣
. (C) IL-4. *P⬍.01
indicates difference between lean and ob/ob mice. §§P⬍.05, §P⬍.01
indicates difference between NE-treated ob/ob mice and control ob/ob
mice. NE, norepinephrine; IFN-
␥
, interferon
␥
; TNF-
␣
, tumor necrosis
factor
␣
; IL-4, inerleukin 4; TNMC, thymic mononuclear cells; NK, natural
killer cells.
Fig. 6. Effect of hepatic NKT cells on liver injury. (A) Serum ALT levels
in CD1d⫺/⫺mice and wild-type mice after LPS. **P⬍.05 for CD1d⫺/⫺
group versus wild-type controls. (B) Serum ALT levels were compared in
lean, NE-treated ob/ob mice and vehicle-treated ob/ob mice. *P⬍.01
for lean versus ob/ob mice. §P⬍.01 for ob/ob versus ob/ob ⫹NE
mice. wt, wild-type; ALT, alanine aminotransferase; LPS, lipopolysaccha-
ride; NE, norepinephrine.
438 LI ET AL. HEPATOLOGY, August 2004

ob/ob and control mice failed to demonstrate quantitative
differences in resident CD4⫹or CD8⫹T cells or NK cell
populations, but we cannot exclude the possibility that
strain-related differences in other types of immune cells
(e.g., gamma-delta T cells) or hepatocytes themselves
might contribute.
NE Treatment Reduces Hepatic Inflammation in
ob/ob Mice. With age, obese ob/ob mice develop intes-
tinal bacterial overgrowth,
29
which increases portal endo-
toxemia and hepatic exposure to endogenous LPS.
30
This,
in turn, promotes inflammatory cytokine production and
steatohepatitis, because treating ob/ob mice with either
probiotic (to modify their exposure to endogenous intes-
tinal bacterial products) or anti–TNF-
␣
antibodies (to
inhibit TNF-
␣
activity) significantly improves their ste-
atohepatitis.
24
Evidence that CD1d⫺/⫺mice, which have
reduced hepatic NKT cells, exhibit increased sensitivity to
LPS-induced hepatotoxicity, suggested to us that reduced
hepatic NKT cells might be the underlying cause of LPS
sensitivity in ob/ob livers. If so, then NE treatment
(which increases hepatic NKT cells and decreases proin-
flammatory cytokine production in ob/ob mice) is pre-
dicted to enhance resistance to LPS hepatoxicity. To
evaluate this possibility, we measured serum levels of ALT
in NE-treated ob/ob mice, as well as vehicle-treated ob/ob
control mice and lean mice. As expected, baseline serum
ALT levels were greater in control ob/ob mice, which have
spontaneous steatohepatitis, than in lean mice, which
have normal livers. Simply treating ob/ob mice with NE
returns serum ALT levels to near normal values (see Fig.
6B). Thus, in adult leptin-deficient mice, replenishing
NE restores hepatic NKT cell populations, reverses Th-1
polarization of hepatic cytokine-producing cells, and re-
duces hepatic sensitivity to endogenous, intestinally de-
rived LPS, despite persistent obesity and leptin deficiency.
Discussion
Chronic inflammation is now recognized as a key me-
diator of obesity-related diseases, including type 2 diabe-
tes and NAFLD.
31
In ob/ob mice, an animal model for
obesity-associated diseases, expression of the proinflam-
matory cytokine TNF-
␣
is increased in white adipose
tissue and liver, and both insulin resistance and fatty liver
disease are improved significantly by inhibiting TNF-
␣
activity.
24
However, the mechanisms that induce and
maintain TNF-
␣
activity during obesity are poorly un-
derstood. In leptin-deficent ob/ob mice, leptin replace-
ment reverses obesity and obesity-related inflammatory
diseases; thus there is no doubt that leptin deficiency pro-
motes chronic inflammation. Paradoxically, however,
TNF-
␣
expression and activity are also increased in other
animal models of obesity and in obese humans with hy-
perleptinemia, prompting speculation that excessive
(rather than deficient) leptin drives obesity-related in-
flammation.
32
However, the latter concept has been dif-
ficult to validate, because some degree of leptin resistance
typically develops during chronic hyperleptinemia.
33
The poor correlation between serum leptin levels and
chronic inflammatory activity suggests that factors other
than leptin may be predominately responsible for control-
ling obesity-related inflammation. This possibility is cer-
tainly plausible, because leptin is merely one of the many
factors that regulate the hypothalamus and the pituitary
and adrenal glands. These tissues produce multiple im-
munomodulatory hormones and neurotransmitters,
34
and recent evidence suggests that such neurohumoral fac-
tors directly mediate many aspects of the ob/ob (leptin-
deficient) phenotype. For example, leptin deficiency
decreases NE,
14
and reduced NE is predominately re-
sponsible for the hypotension, hyperostosis, and periph-
eral adiposity that develops in ob/ob mice.
35
The present
study provides additional support for the concept that
reduced NE mediates key features of leptin deficiency by
demonstrating that supplemental NE reverses both the
depletion of hepatic NKT cell populations and the proin-
flammatory polarization of hepatic cytokine production
that develop during leptin deficiency.
Circulating levels of NE are variable in human obesity.
NE levels are increased in some obese populations, par-
ticularly those who have overt hypertension.
36
On the
other hand, other studies show that although many obese
individuals have normal “resting” levels of NE, they ex-
hibit significant differences in NE induction following
stress.
37
In addition, obesity has been associated with cel-
lular resistance to NE and other adrenergic agonists.
38
In
some tissues, this is due to decreased adrenoceptor expres-
sion.
39
Indeed, relative resistance to NE may contribute to
the pathogenesis of obesity, because sibutramine—a
pharmacological agent that prolongs adrenergic activity
by inhibiting catecholamine reuptake—is a reasonably ef-
fective treatment for obesity.
40
When NE activity is decreased, hepatic NKT cell ap-
optosis is increased and restoring NE reduces apoptosis,
even when leptin is absent. Thus, although leptin clearly
has immunomodulatory actions,
7
NE is a more important
viability factor for hepatic NKT cells. Parallel studies in
mice with intact genes for leptin and its receptors demon-
strate that NE is generally important for regulating
hepatic NKT cell populations by showing that comple-
mentary experimental approaches that inhibit NE activity
also reduce NKT cell accumulation in the livers of leptin-
sufficient mice. Evidence that NE regulates hepatic NKT
cells independently of leptin suggests that like leptin-de-
ficient ob/ob mice, obese hyperleptinemic mice and hu-
HEPATOLOGY, Vol. 40, No. 2, 2004 LI ET AL. 439

mans may also develop hepatic NKT cell depletion.
Therefore, hepatic depletion of NKT cells may be in-
volved in a common mechanism that promotes obesity-
related liver damage in both hyper- and hypoleptinemic
individuals.
The possibility that tissue-specific reductions in NKT
cell populations promote organ damage has already been
suggested by other groups.
41
Indeed, given the strong as-
sociation between obesity, diabetes, and NAFLD, it is
particularly intriguing that NKT cell depletion has been
implicated in the pathogenesis of both diabetes and LPS
liver injury in certain nonobese animal models for these
conditions.
6,41
In those models, organ damage is attrib-
uted to relative excesses of proinflammatory cytokines
that develop when tissue NKT cell populations are re-
duced. For example, overabundance of proinflammatory
cytokines such as IFN-
␥
sensitizes hepatocytes to liver
injury upon subsequent exposure to doses of LPS that are
well-tolerated when IFN-
␥
is not excessive.
25
NKT cell
populations normally control proinflammatory Th-1
cytokine activities by promoting the production of
anti-inflammatory Th-2 cytokines.
42
Thus, when
proinflammatory cytokine activity is not tempered by an-
ti-inflammatory cytokines, sustained Th-1 polarization,
chronic inflammation, and progressive tissue injury ensue
in response to stimuli that typically signal a self-limited
inflammatory response. Treatment with NE in ob/ob
mice restores hepatic NKT cell population, reverses
proinflammatory polarization, and significantly reduces
hepatic inflammation.
Thymic selection of NKT cells is blocked during de-
velopment in CD1d⫺/⫺mice. Consequently, in adult
CD1d⫺/⫺mice, NKT cell populations are reduced in
many peripheral tissues, including the liver. As discussed
earlier, studies in normal mice infected with P. acnes,as
well as in leptin-deficient ob/ob mice, have correlated
decreased hepatic NKT cell populations with sensitiza-
tion to hepatotoxicity from LPS, a potent inducer of
proinflammatory cytokines. However, because P. acnes
infection and leptin deficiency influence multiple compo-
nents of the immune system, a cause–effect relationship
between hepatic NKT cell depletion and sensitization to
LPS liver injury remains speculative. Studies in genetically
altered mice with selective depletions of certain NKT cell
populations have yielded discrepant results—one group
reported that NKT cells protect against LPS toxicity,
43
but another group reported the opposite result.
44
Present
evidence that otherwise healthy, CD1d⫺/⫺mice develop
worse liver injury after LPS exposure than their littermate
controls provides compelling additional support for the
importance of reduced hepatic NKT cells in enhancing
sensitivity to toxicity from LPS. Our results complement
and extend those reported by Emoto et al.,
43
who showed
that mice that are deficient in V
␣
14⫹NKT cells due to
targeted disruption of
␤
2microglobulin, a factor that is
necessary for CD1d activity, are profoundly sensitive to
LPS toxicity. Together with our other findings, this result
suggests that the decreases in NE that occur in leptin
deficiency may mediate progression of obesity-related
liver disease, because decreased NE promotes apoptosis of
hepatic NKT cells. This in turn reduces NKT cell popu-
lations in the liver. Dysregulation of other hepatic cyto-
kine-producing cells ensues, leading to the accumulation
of Th-1 polarized lymphocytes. When confronted by a
stimulus for proinflammatory cytokine production (e.g.,
LPS), production of proinflammatory cytokines is rela-
tively unconstrained in the liver, enhancing acute LPS
hepatotoxicity while also promoting a sustained (i.e.,
chronic) inflammatory response.
References
1. Day CP, James O. Steatohepatitis: a tale of two “hits”? Gastroenterology
1998;114:842–845.
2. Zhang Y, Proenca R, Maffei M, Barone M, Leopold L, Friedman JM.
Positional cloning of the mouse obese gene and its human homologue.
Nature 1994;372:425–432.
3. Yang SQ, Lin HZ, Diehl AM. Fatty liver vulnerability to lipopolysaccaride
despite NF-kB induction and caspase 3 inhibition. Am J Physiol 2001;281:
G382–G392.
4. Guebre-Xabier M, Yang SQ, Lin HZ, Schwenk R, Kryzch U, Diehl AM.
Altered hepatic lymphocyte subpopulations in obesity-related fatty livers.
HEPATOLOGY 1999;31:633–640.
5. Benlagha K, Kyin T, Beavis A, Teyton L, Bendelac A. A thymic precursor
to the NK T cell lineage. Science 2002;296:553–555.
6. Matsui K, Yoshimoto T, Tsutsui H, Hyodo Y, Hayashi N, Hiroishi K, et
al. Propionibacterium acnes treatment diminishes CD4⫹NK1.1⫹T cells
but induces type 1 T cells in the liver by induction of IL-12 and IL-18
production from Kupffer cells. J Immunol 1997;159:97–106.
7. Faggioni R, Feingold KR, Grunfeld C. Leptin regulation of the immune
response and the immunodeficiency of malnutrition. FASEB J 2001;15:
2565–2571.
8. Zarkesh-Esfahani H, Pockley G, Metcalfe RA, Bidlingmaier M, Wu Z,
Ajami A, et al. High-dose leptin activates human leukocytes via receptor
expression on monocytes. J Immunol 2001;167:4593–4599.
9. Lord GM, Matarese G, Howard JK, Baker RJ, Bloom SR, Lechler RI.
Leptin modulates the T-cell immune response and reverses starvation-
induced immunosuppression. Nature 1998;394:897–901.
10. Loffreda S, Yang SQ, Lin HZ, Bulkley G, Bregman M, Nobel P, et al.
Leptin regulates proinflammatory immune responses. FASEB J 1998;12:
57–65.
11. Li Z, Lin HZ, Yang SQ, Diehl AM. Murine leptin deficiency alters kupffer
cell production of cytokines that regulate the innate immune system. Gas-
troenterology 2002;123:1304–1310.
12. Howard JK, Lord GM, Matarese G, Vendetti S, Ghatei MA, Ritter MA, et
al. Leptin protects mice from starvation-induced lymphoid atrophy and
increases thymic cellularity in ob/ob mice. J Clin Invest 1999;104:1051–
1059.
13. Makimura H, Mizuno TM, Roberts J, Silverstein J, Beasley J, et al. Adre-
nalectomy reverses obese phenotype and restores hypothalamic melanocor-
tin tone in leptin-deficient ob/ob mice. Diabetes 2000;49:1917–1923.
14. Knehans AW, Romsos DR. Reduced norepinephrine turnover in brown
adipose tissue of ob/ob mice. Am J Physiol 1982;242:E253–E261.
440 LI ET AL. HEPATOLOGY, August 2004

15. Takeda S, Elefteriou F, Levasseur R, Liu X, Zhao L, Parker KL, et al. Leptin
regulates bone formation via the sympathetic nervous system. Cell 2002;
111:305–317.
16. Cohen P, Zhao C, Cai X, Montez JM, Rohani SC, Feinstein P, et al.
Selective deletion of leptin receptor in neurons leads to obesity. J Clin
Invest 2001;108:1113–1121.
17. Minagawa M, Oya H, Yamamoto S, Shimizu T, Bannai M, Kawamura H,
et al. Intensive expansion of natural killer T cells in the early phase of
hepatocyte regeneration after partial hepatectomy in mice and its associa-
tion with sympathetic nerve activation. HEPATOLOGY 2000;31:907–915.
18. Thomas SA, Matsumoto AM, Palmiter RD. Noradrenaline is essential for
mouse fetal development. Nature 1995;374:643–646.
19. Dubuisson L, Desmouliere A, Decourt B, Evade L, Bedin C, Boussarie L,
et al. Inhibition of rat liver fibrogenesis through noradrenergic antagonism.
HEPATOLOGY 2002;35:325–331.
20. Suzuki S, Toyabe S, Moroda T, Tada T, Tsukahara A, Iiai T, et al. Circa-
dian rhythm of leucocytes and lymphocytes subsets and its possible corre-
lation with the function of the autonomic nervous system. Clin Exp
Immunol 1997;110:500–508.
21. Kayagaki N, Yamaguchi N, Nakayama M, Takeda K, Akiba H, Tsutsui H,
et al. Expression and function of TNF-related apoptosis-inducing ligand
on murine activated NK cells. J Immunol 1999;163:1906–1913.
22. Lin HZ, Yang SQ, Kujhada F, Ronnet G, Diehl AM. Metformin reverses
nonalcoholic fatty liver disease in obese leptin-deficient mice. Nat Med
2000;6:998–1003.
23. Shigeno M, Nakao K, Ichikawa T, Suzuki K, Kawakami A, Abiru S, et al.
Interferon-alpha sensitizes human hepatoma cells to TRAIL-induced apo-
ptosis through DR5 upregulation and NF-kappaB inactivation. Oncogene
2003;22:1653–1662.
24. Li Z, Yang S, Lin H, Huang J, Watkins PA, Moser AB, et al. Probiotics and
antibodies to TNF inhibit inflammatory activity and improve nonalco-
holic fatty liver disease. HEPATOLOGY 2003;37:343–350.
25. Shimizu Y, Margenthaler JA, Landeros K, Otomo N, Doherty G, Flye
MW. The resistance of P. acnes-primed interferon gamma-deficient mice
to low-dose lipopolysaccharide-induced acute liver injury. HEPATOLOGY
2002;35:805–814.
26. Brodbeck WG, Shive MS, Colton E, Ziats NP, Anderson JM. Interleu-
kin-4 inhibits tumor necrosis factor-alpha-induced and spontaneous apo-
ptosis of biomaterial-adherent macrophages. J Lab Clin Med 2002;139:
90–100.
27. Bendelac A, Rivera MN, Park SH, Roark JH. Mouse CD1-specific NK1 T
cells: development, specificity, and function. Annu Rev Immunol 1997;
15:535–562.
28. Mendiratta SK, Martin WD, Hong S, Boesteanu A, Joyce S, Van-Kaer L.
CD1d1 mutant mice are deficient in natural T cells that promptly produce
IL-4. Immunity 1997;6:469–477.
29. Cope K, Risby T, Diehl AM. Increased gastrointestinal ethanol production
in obese mice: implications for fatty liver disease pathogenesis. Gastroen-
terology 2000;119:1340–1347.
30. Billiar TR, Maddaus MA, West MA, Curran RD, Wells CA, Simmons
RL. Intestinal gram-negative bacterial overgrowth in vivo augments the
in vitro response of Kupffer cells to endotoxin. Ann Surg 1988;208:
532–540.
31. Yuan M, Konstantopoulos N, Lee J, Hansen L, Li Z-W, Karin M, et al.
Reversal of obesity- and diet-induced insulin resistance with salicylates or
targeted disruption of IKKbeta. Science 2001;293:1673–1677.
32. Kirchgessner TG, Uysal KT, Wiesbrock SM, Marino MW, Hotamisligil
GS. Tumor necrosis factor-alpha contributes to obesity-related hyperlep-
tinemia by regulating leptin release from adipocytes. J Clin Invest 1997;
100:2777–2782.
33. Zimmet P, Boyko EJ, Collier GR, de Courten M. Etiology of the meta-
bolic syndrome: potential role of insulin resistance, leptin resistance, and
other players. Ann N Y Acad Sci 1999;892:25–44.
34. Haddad JJ, Saade NE, Safieh-Garabedian B. Cytokines and neuro-im-
mune-endocrine interactions: a role for the hypothalamic-pituitary-adre-
nal revolving axis. J Neuroimmunol 2002;133:1–19.
35. Young JB, Landsberg L. Diminished sympathetic nervous system ac-
tivity in genetically obese (ob/ob) mouse. Am J Physiol 1983;245:
E148–E154.
36. Esler M, Lambert G, Brunner-La Rocca HP, Vaddadi G, Kaye D. Sympa-
thetic nerve activity and neurotransmitter release in humans: translation
from pathophysiology into clinical practice. Acta Physiol Scand 2003;177:
275–284.
37. Salvadori A, Fanari P, Giacomotti E, Palmulli P, Bolla G, Tovaglieri I, et al.
Kinetics of catecholamines and potassium, and heart rate during exercise
testing in obese subjects. Heart rate regulation in obesity during exercise.
Eur J Nutr 2003;42:181–187.
38. Malara A, Corsonello A, Buemi M, De Domenico D, Ientile R, Corica F.
Effects of alpha- and beta-adrenergic stimulation on free magnesium con-
centrations in platelets from healthy and obese individuals. Magnes Res
2001;14:263–272.
39. Faulds G, Ryden M, Ek I, Wahrenberg H, Arner P. Mechanisms be-
hind lipolytic catecholamine resistance of subcutaneous fat cells in the
polycystic ovarian syndrome. J Clin Endocrinol Metab 2003;88:2269 –
2273.
40. Luque CA, Rey JA. The discovery and status of sibutramine as an anti-
obesity drug. Eur J Pharmacol 2002;440:119–128.
41. Baxter AG, Kinder SJ, Hammond KJ, Scollay R, Godfrey DI. Association
between alphabeta TCR⫹CD4-CD8- T-cell deficiency and IDDM in
NOD/lt mice. Diabetes 1997;46:572–582.
42. Tanaka Y, Takahashi A, Watanabe K, Takayama K, Yahata T, Habu S, et
al. A pivotal role of IL-12 in Th1-dependent mouse liver injury. Int Im-
munol 1996;8:569–576.
43. Emoto M, Miyamoto M, Yoshizawa I, Emoto Y, Schaible UE, Kita E, et al.
Critical role of NK cells rather than Valpha14(⫹)NKT cells in lipopolysac-
charide-induced lethal shock in mice. J Immunol 2002;169:1426–1432.
44. Dieli F, Sireci G, Russo D, Taniguchi M, Ivanyi J, Fernandez C, et al.
Resistance of natural killer T cell-deficient mice to systemic Shwartzman
reaction. J Exp Med 2000;192:1645–1652.
HEPATOLOGY, Vol. 40, No. 2, 2004 LI ET AL. 441