of November 13, 2013.

This information is current as

Cells and TNF
Exotoxin A in Mice Depends on T aeruginosa
Pseudomonas Acute Hepatotoxicity of
Lohoff and Gisa Tiegs
Jens Schmann, Sabine Angermller, Renate Bang, Michael
http://www.jimmunol.org/content/161/10/5745
1998; 161:5745-5754; ; J Immunol
References
http://www.jimmunol.org/content/161/10/5745.full#ref-list-1
, 21 of which you can access for free at: cites 54 articles This article
Subscriptions
http://jimmunol.org/subscriptions
is online at: The Journal of Immunology Information about subscribing to
Permissions
http://www.aai.org/ji/copyright.html
Submit copyright permission requests at:
Email Alerts
http://jimmunol.org/cgi/alerts/etoc
Receive free email-alerts when new articles cite this article. Sign up at:
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Immunologists All rights reserved.
Copyright 1998 by The American Association of
9650 Rockville Pike, Bethesda, MD 20814-3994.
The American Association of Immunologists, Inc.,
is published twice each month by The Journal of Immunology

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m


b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Acute Hepatotoxicity of Pseudomonas aeruginosa Exotoxin A in
Mice Depends on T Cells and TNF
1
Jens Schu mann,* Sabine Angermu ller,

Renate Bang,* Michael Lohoff,

and Gisa Tiegs

2
*
The most potent virulence factor of Pseudomonas aeruginosa, its exotoxin A (PEA), inhibits protein synthesis, especially in the
liver, and is a weak T cell mitogen. This study was performed to correlate hepatotoxic and possible immunostimulatory features
of PEA in vivo. Injection of PEA to mice caused hepatocyte apoptosis, an increase in plasma transaminase activities, and the
release of TNF, IFN-, IL-2, and IL-6 into the circulation. Most strikingly, liver damage depended on T cells. Athymic nude mice
or mice depleted of T cells by anti-Thy1.2 mAb pretreatment failed to develop acute hepatic failure, and survival was signicantly
prolonged following T cell depletion. Neutralization of TNF or lack of TNF receptors prevented liver injury. In the liver, TNF was
produced by Kupffer cells before hepatocellular death occurred. After T cell depletion, Kupffer cells failed to produce TNF.
Transaminase release was signicantly reduced in perforin knockout mice, and it was even elevated in lpr/lpr mice. These results
demonstrate that PEA induces liver damage not only by protein synthesis inhibition but also by TNF- and perforin-dependent,
Fas-independent, apoptotic signals. The Journal of Immunology, 1998, 161: 57455754.
P
seudomonas aeruginosa is a Gram-negative, opportunistic
pathogen that evokes respiratory, urinary tract, or skin in-
fections as well as gastrointestinal disorders associated with
enterocolitis and bowel perforation. Moreover, P. aeruginosa fre-
quently causes septicemia in immunocompromised patients, and a
high incidence of P. aeruginosa bacteremia was observed in pa-
tients with impaired barrier function of the liver (1, 2). P. aerugi-
nosa is the fourth most common cause of primary, hospital-ac-
quired, Gram-negative bacteremia and is usually associated with
high resistance to antibiotic treatment and high mortality rates. P.
aeruginosa bacteremia clinically resembles other forms of Gram-
negative sepsis, i.e., common symptoms are fever, hypotension,
refractory shock, adult respiratory distress syndrome, and renal
failure (1). However, jaundice appears to occur more often than in
other forms of Gram-negative sepsis (1).
The two bacterial products most likely implicated in the sys-
temic toxicity of P. aeruginosa are its LPS and exotoxin A (PEA).
3
Puried exotoxin A is highly lethal for animals, including subhu-
man primates, and produces shock in dogs and rhesus monkeys.
The biologic signicance of exotoxin A for the pathogenicity of P.
aeruginosa became evident by studies showing that patients with
high levels of serum Abs to exotoxin A at the onset of P. aerugi-
nosa septicemia have a better chance of survival than those with
low Ab titers (1, 3). Moreover, in a mouse model it was shown that
PEA-producing P. aeruginosa strains were more toxic than non-
producers, and that their toxicity could be weakened by Abs
to PEA (4).
Hepatic injury due to systemic inammatory processes has been
reported to occur in the pathophysiology of septic shock (5). LPS
from Gram-negative bacteria have been intensively studied with
respect to their capacity to induce shock by stimulation of mono-
cytic cells. These cells release TNF and other proinammatory
cytokines that mediate multiorgan failure and lethality in experi-
mental animal models of endotoxic shock. More recently, it be-
came evident that activation not only of macrophages but also of
T cells may result in a systemic inammatory response syndrome
and organ injury (6). We recently described two models of T lym-
phocyte-dependent apoptotic and secondary necrotic liver injury
induced by either the anti-mouse CD3 mAb 145 2C11 or the bac-
terial superantigen Staphylococcus aureus enterotoxin B (SEB) in
D-galactosamine (GalN)-sensitized mice (7, 8). In these models the
T cell stimuli evoke a cytokine release syndrome. TNF was found
to be the central mediator of hepatocellular apoptosis and the en-
suing severe liver failure (7). The amino sugar GalN depletes ura-
cil nucleotides selectively in the liver, thereby inhibiting hepatic
transcription and translation (9) and sensitizing the liver toward T
cell stimuli (7, 8), LPS (1012), or TNF (13, 14).
Like GalN, PEAinhibits protein synthesis in mammalian cells, e.g.,
in hepatocytes (15, 16), thus being a bacterial toxin that possibly sen-
sitizes the liver toward proinammatory cytokines such as TNF. In
human ovarian tumor cell lines, PEA induced apoptosis that was ac-
celerated by TNF (17). This property of PEA led to the use of PEA-
derived fusion proteins as immunotoxins for cancer treatment (18). In
clinical trials in humans, the major side effects of immunotoxins are a
vascular leak syndrome, fever, and hepatotoxicity (18, 19).
In the past, several reports have tried to draw a connection be-
tween PEA and T cells. In vitro, PEA led to mitogenic activation
of T lymphocytes in murine whole splenocyte cultures and gener-
ated cytolytic T lymphocytes active against EL4 target cells (20).
This apparent lack of an apoptotic signal on T cells may be ex-
plained by a minor sensitivity of lymphoid cells to inhibition of
protein synthesis by PEA compared with the sensitivity of other
mammalian cells (20). Moreover, PEA exhibited properties that
*Institute of Experimental and Clinical Pharmacology and Toxicology, and

Institute
of Clinical Microbiology and Immunology, University of Erlangen-Nurnberg, Erlan-
gen, Germany; and

Department of Anatomy and Cell Biology II, University of Hei-
delberg, Heidelberg, Germany
Received for publication December 18, 1997. Accepted for publication July 10, 1998.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1
This work was supported by Deutsche Forschungsgemeinschaft Grant Ti 169/3-4
and by Interdisziplinares Zentrum fur Klinische Forschung, Universitat Erlangen-
Nurnberg.
2
Address correspondence and reprint requests to Dr. Gisa Tiegs, Institute of Exper-
imental and Clinical Pharmacology and Toxicology, University of Erlangen-Nurn-
berg, Universitatsstrasse 22, D-91054 Erlangen, Germany. E-mail address:
gt@macpost.pharmakologie.uni-erlangen.de
3
Abbreviations used in this paper: PEA, Pseudomonas aeruginosa exotoxin A; SEB,
Staphylococcus aureus enterotoxin B; GalN, D-galactosamine; HSA, human serum
albumin; LAL, Limulus amebocyte lysate; ALT, alanine aminotransferase.
Copyright 1998 by The American Association of Immunologists 0022-1767/98/$02.00

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

are similar, but not identical with those of microbial superantigens;
PEA selectively stimulated the proliferation of murine thymocytes
expressing the V8.2 chain in their TCR (21), and thymocyte pro-
liferation depended on the presence of APCs (22, 23). However, in
contrast to conventional superantigens, lymphoproliferative activ-
ity required intracellular processing (24), and no superantigenic
activity of PEA was found by others (25). Hence, the objective of
this study was to further investigate the role of TNF and T cells in
the toxicity of PEA in vivo.
Materials and Methods
Mice
Six- to 8-wk-old BALB/c mice were obtained from the institutes internal
animal breeding house. TNFR-decient and the corresponding wild-type
(C57BL/6 129/Sv) mice were provided by Dr. H. Bluethmann, Hoff-
mann-La Roche (Basel, Switzerland). Perforin knockout mice were pro-
vided by Drs. J. Tschopp and M. Schroter, University of Lausanne, Insti-
tute of Biochemistry (Epalinges, Switzerland). Athymic BALB/c-nu/nu
mice, Fas-decient lpr/lpr mice, and mice of the corresponding wild-type
strain MRL/Mp, as well as C3H/HeN and C3H/HeJ mice, were purchased
from Harlan Winkelmann (Borchen, Germany). Animals received humane
care according to National Institutes of Health guidelines as well as the
legal requirements in Germany and were maintained under controlled con-
ditions (22C, 55% humidity, 12-h day/night rhythm) and were fed a stan-
dard laboratory chow (Altromin 1313, Altromin, Lage, Germany) ad
libitum.
Animal treatments
All reagents were injected in a total volume of 250 l/25 g mouse. PEA
(Sigma, St. Louis, MO) was injected i.v. in pyrogen-free saline containing
0.1% human serum albumin (HSA). The LPS content of PEA was deter-
mined with the help of a commercially available Limulus amebocyte lysate
(LAL) kit (Coatest Endotoxin, Chromogenix, Molndal, Sweden). In some
experiments mice were treated with one of the following combinations of
reagents: 1) 30 mg/kg i.v. of benzyloxycarbonyl-Val-Ala-Asp uorometh-
ylketone (Bachem Biochemica, Heidelberg, Germany) in saline/6% DMSO
followed by PEA 15 min later; 2) 20 l/mouse i.v. of the IgG fraction of
a sheep anti-mouse TNF polyclonal antiserum (provided by Dr. A. Wendel,
University of Konstanz, Konstanz, Germany) in saline/0.1% HSA followed
by PEA 15 min later; 3) 200 l/mouse i.v. of rabbit anti-mouse IFN-
polyclonal antiserum in saline/0.1% HSA followed by PEA 15 min later;
4) 100 l/mouse i.v. of anti-mouse Thy1.2 mAb (IgG2a; Bio-Yeda, Re-
hovot, Israel; this treatment leads to depletion of 82% of T cells 24 h later
(26)) in saline followed by PEA 24 h later; 5) 100 l/mouse i.v. of control
Abs, i.e., rat IgG2a (PharMingen, Hamburg, Germany), in saline followed
by PEA 24 h later; 6) 50 mg/kg i.v. of cyclosporin A in saline/10% placebo
(Sandoz, Nurnberg, Germany) 15 and 1 h before PEA; 7) 2.5 mg/kg i.p. of
SEB (Sigma) in saline/0.1% HSA followed by PEA 48 h later; 8) 2.5
mg/kg i.p. of SEB in saline/0.1% HSA followed by GalN/SEB (GalN, 700
mg/kg i.p.; SEB, 2.5 mg/kg i.p.) 48 h later; 9) 2.5 mg/kg i.p. of SEB in
saline/0.1% HSA followed by GalN/LPS (GalN, 700 mg/kg i.p.; LPS from
Salmonella abortus equi, S form (Metalon, Ragow, Germany), 10 g/kg
i.p.) 48 h later; 10) 10 g/kg recombinant murine TNF (provided by Dr.
G. R. Adolf, Bender & Co., Vienna, Austria) i.v. 15 min after injection of PEA
in saline/0.1%HSA; 11) 10 g/kg LPS fromS. abortus equi, S form, given i.p.
15 min after injection of 3 g/kg PEA in saline; and 12) 2.5 mg/kg SEB given
i.p. 15 min after injection of 3 g/kg PEA in saline/0.1% HSA.
Sampling of material
Mice were lethally anesthetized with 150 mg/kg pentobarbital i.v. contain-
ing a dose of 15 mg/kg heparin. Blood was withdrawn by cardiac puncture.
Livers were perfused via the portal vein for 10 s with cold perfusion buffer
(89 mM Tris-HCl, 89 mM boric acid, 2 mM EDTA, and 10 mM EGTA, pH
8.0) before excision. In some experiments, perfusion was performed for
20 s via the right ventricle of the heart before excision of liver, lung,
cecum, kidneys, and heart. Cecum was luminally perfused with saline to
remove contents. All organs were blotted dry and disintegrated in cold
perfusion buffer with an Elvehjem-type homogenizer. The 20% organ ho-
mogenates were centrifuged at 13,000 g for 15 min. One part of the
supernatant was to detect oligonucleosome-bound DNA fragments (see
below). The other part was used to precipitate DNA by addition of 1 ml of
ice-cold ethanol plus 50 l of 3 M sodium acetate.
Analysis of liver enzymes and released proteins
Hepatocyte damage was assessed by measuring plasma enzyme activities
of alanine aminotransferase (ALT), aspartate aminotransferase, and sorbi-
tol dehydrogenase according to Bergmeyer (27) using an automated pro-
cedure. The cytokines TNF, IFN-, IL-2, and IL-6 were determined by
ELISA (PharMingen). For determination of TNF, a polyclonal sheep anti-
mouse TNF capture Ab (28), puried on protein G columns (Pharmacia,
Freiburg, Germany) was used to replace the PharMingen capture mAb.
IL-10, IL-1, and IL-1 were measured using commercial ELISA kits
(IL-10: Genzyme, Cambridge, MA; IL-1: Endogen, Cambridge, MA; IL-
1: Paesel & Lorei Co., Hanau, Germany).
DNA fragmentation
DNA fragmentation was quantied (29) by measuring cytosolic oligonu-
cleosome-bound DNA using an ELISA kit (Boehringer Mannheim, Mann-
heim, Germany). Briey, the cytosolic fraction (13,000 g supernatant)
from approximately 200 g of liver was employed as Ag source in a
sandwich ELISA with a primary anti-histone Ab coated to the microtiter
plate and a secondary anti-DNA Ab coupled to peroxidase. DNA fragmen-
tation was also analyzed semiquantitatively after extraction of the
13,000 g supernatant (corresponding to 80 mg of liver) by the phenol/
chloroform method, precipitation by ethanol, and electrophoresis on 1.0%
agarose gels (30).
Electron microscopic studies
For electron microscopic studies, the livers were xed by perfusing the
portal vein with a xative containing 0.25% glutaraldehyde and 2% sucrose
in 100 mM PIPES buffer at pH 7.4 for 5 min. Sections were postxed with
2% aqueous osmium tetroxide, dehydrated in graded ethanol, and embed-
ded in Epon 812. Ultrathin sections were counterstained with lead citrate
and examined in a Philips EM 301 electron microscope (Eindhoven, The
Netherlands).
Immunouorescent staining and confocal laser imaging
Twelve-micrometer-thick cryostat sections of livers were thawed onto
glass slides, air-dried, and xed in acetone/methanol (1/1) for 10 min at
4C before they were incubated in PBS containing 3% BSA for 30 min at
room temperature. After the slides had been rinsed in PBS, incubation was
continued with polyclonal rabbit anti-mouse-TNF Ab (Genzyme Virotech,
Russelsheim, Germany; 1/750) together with rat mAb against murine pan-
macrophage marker (clone BM8, Dianova, Hamburg, Germany; 1/100), or
mouse CD4 (clone RM4-5, PharMingen; 1/50), or mouse CD8 (clone Ly-2,
PharMingen; 1/50) in PBS containing 3% BSA overnight at 4C. After
rinsing with PBS, binding sites were detected using swine anti-rabbit IgG
tagged with FITC (Dako, Hamburg, Germany; 1/30) and goat anti-rat IgG
tagged with Texas Red (Dianova; 1/200) in PBS containing 3% BSA for 1 h
at room temperature. After rinsing with PBS, sections were coverslipped with
TBS/glycerol (1/1), pH 8.6. Sections processed for immunouorescence were
examined by confocal laser scanning microscopy (MRC 1000, Bio-Rad,
Richmond, CA).
Statistical analysis
The results were analyzed using Students t test if two groups were com-
pared and the Dunnetts test if more groups were tested against a control
group. If variances were inhomogeneous, the results were analyzed using
the Welsh test. The signicance of prolonged survival was tested by com-
parison of survival curves with the log-rank test. All data in this study are
expressed as the mean SEM. p 0.05 was considered signicant.
Results
PEA-induced liver injury
The i.v. injection of PEA into BALB/c mice dose-dependently
induced liver cell damage within 12 h as assessed by an increase
in the activities of plasma transaminases as well as enhanced ac-
tivities of the liver-specic enzyme sorbitol dehydrogenase in
plasma (Fig. 1). Administration of the highest dose, i.e., 300 g/kg
PEA, resulted in lethality within 16 h. The release of liver enzymes
was preceded by the appearance of cytosolic oligonucleosome-
bound DNA within the liver that was signicantly enhanced 6.5 h
following PEA administration (Fig. 2). Internucleosomal DNA
fragmentation as a measure of programmed cell death was also
5746 T CELL DEPENDENCE OF Pseudomonas EXOTOXIN A HEPATOTOXICITY IN MICE

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

demonstrated on an agarose gel. The appearance of the DNA lad-
der again preceded the increase in plasma ALT (Fig. 3, lanes 13).
The unspecic tripeptidic caspase inhibitor benzyloxycarbonyl-
Val-Ala-Asp uoromethylketone (Z-VAD.fmk), given in a dose
that protected mice from GalN/TNF-induced liver injury (31) (data
not shown), also signicantly inhibited transaminase release fol-
lowing PEA challenge (buffer plus PEA: ALT, 8970 2355 U/l;
Z-VAD.fmk plus PEA: ALT, 990 215 U/l ( p 0.05); n 6).
The inhibitor had no effect on PEA-induced TNF production, i.e.,
Z-VAD.fmk did not reduce the amount of circulating TNF (data
not shown). Together these data clearly suggest a role for apoptotic
proteases in PEA-dependent liver failure.
To explore whether other organs besides the liver were injured
by PEA, we quantied DNA fragmentation in lung, cecum, kid-
ney, and heart (32) compared with that in liver. An increase in
cytosolic oligonucleosome-bound DNA was most prominent
within the liver and the lung (data not shown).
To determine which cells in the liver died by apoptosis, we set
up transmission electron micrographs of mouse livers excised 12 h
following PEA injection. Micrographs clearly show that the dying
cells were hepatocytes (Fig. 4, a and b). They died by apoptosis
(Fig. 4, a and b) and necrosis (Fig. 4b), whereas neither apoptosis
nor necrosis was seen in livers of saline-treated mice (33). Necrotic
liver areas contained inltrated polymorphonuclear neutrophils
(Fig. 4b). As a cause of damage, perfusion of the livers with the
xative was incomplete. Hence, blood cells were still located in
the hepatic sinusoids (Fig. 4c). Fig. 4c also shows a lymphocyte
sticking to sinusoidal endothelium. Some lymphocytes inltrating
into the hepatic tissue were observed. One example is shown in
Fig. 4d. Fig. 4, ad, also demonstrates that hepatocytes of PEA-
treated mice contained numerous lipid droplets. The development
of a fatty liver most likely results from an inhibition of apolipopro-
tein synthesis by PEA.
Inhibition of transcription by GalN sensitizes the liver toward
LPS (10, 12, 33). Because PEA, like GalN, has protein synthesis
inhibitory activities, it had to be excluded that PEA was merely
sensitizing the mice to contaminating LPS present either in the
PEA preparation or in the intestines of mice. However, the fol-
lowing experiments argue against this idea. 1) As tested by the
LAL assay, the contaminating LPS dose (1.4 ng/kg) injected to-
gether with PEA (300 g/kg) was 7000-fold lower than the LPS
dose (10 g/kg) used in our laboratory to induce liver injury in
GalN-sensitized BALB/c mice. 2) Translocation of LPS from the
intestine was excluded by showing that LPS was below 15 pg/ml
(detection limit) in plasma 3.5, 8, and 12 h following injection of
300 g/kg PEA to mice as determined by the LAL assay. 3) LPS-
resistant C3H/HeJ mice were as sensitive as LPS-responsive C3H/
HeN mice toward PEA (ALT, 1247 881 vs 833 525 U/L; n
3). Therefore, the involvement of LPS in the induction of PEA-
induced liver damage is very unlikely.
PEA-induced cytokine release
As PEA has been described to stimulate the proliferation of T cells
in vitro (20, 21), it seemed feasible that the toxin also stimulates
cytokine production in vivo. By determination of plasma cytokines
following PEA administration to mice we found a time-dependent
release of the proinammatory cytokines TNF and IL-6 as well as
of the T cell cytokines IL-2 and IFN- (Fig. 5). IL-1, IL-1, and
IL-10 were not detectable in plasma of PEA-treated animals (IL-
1, 6 pg/ml; IL-1, 1 pg/ml; IL-10, 15 pg/ml). Thus, cyto-
kine analysis suggests that PEA also activates T lymphocytes in
vivo.
FIGURE 1. Dose dependence of PEA-induced liver cell damage.
BALB/c mice received different doses of PEA i.v.; 12 h after challenge the
activities of plasma transaminases and sorbitol dehydrogenase were deter-
mined. Data are expressed as the mean SEM (n 36). , p 0.05 vs
0 g/kg treatment. AST, aspartate aminotransferase.
FIGURE 2. Time course of oligonucleosomal hepatic DNA fragmenta-
tion and ALT release in PEA-treated BALB/c mice. BALB/c mice received
300 g/kg PEA i.v.; at the indicated times blood was withdrawn for ALT
determination, and liver homogenates were prepared for quantication of
oligonucleosomal hepatic DNA fragmentation with the help of a commer-
cially available cell death detection ELISA. Data are expressed as the
mean SEM (n 36). , p 0.05 vs 0 h.
FIGURE 3. Internucleosomal DNA fragmentation in livers from PEA-
treated BALB/c mice and its prevention by anti-TNF or anti-Thy1.2 pre-
treatment. BALB/c mice received 300 g/kg PEA i.v. and were treated as
described below in the different lanes. DNA was prepared from 13,000
g supernatants of liver homogenates and was stained with ethidium bro-
mide after analysis on a 1% agarose gel. Lane M, Marker of a 123-nucle-
otide pair ladder; lane 1, control (12 h after HSA treatment); lane 2, PEA
(8 h after treatment); lane 3, PEA (12 h after treatment); lane 4, PEA/anti-
TNF (12 h after PEA treatment); lane 5, PEA/anti-Thy1.2 (12 h after PEA
treatment).
5747 The Journal of Immunology

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

TNF as a mediator of PEA hepatotoxicity
Because TNF has been identied as a common mediator of hep-
atocellular apoptosis and liver injury in experimental mouse mod-
els (7, 29, 33, 34), we wondered whether TNF is also a mediator
of PEA-induced liver damage.
Although TNF was not detectable in plasma until 12 h after
administration of PEA (Fig. 5), i.e., at a point when liver damage
had already developed, TNF was locally produced within the liver
as early as 3 h after administration of PEA to BALB/c mice. This
was shown by means of immunouorescent staining followed by
confocal laser imaging (Fig. 6). TNF was colocalized with resident
liver macrophages, which were stained with the BM8 murine pan-
macrophage marker (Fig. 6A, PEA 3 h), indicating that TNF was
produced by Kupffer cells at this early time point. CD4- and CD8-
positive T cells were present within the liver 3 h after PEA injec-
tion in amounts comparable to those observed at time zero. These
cells did not produce TNF (Fig. 6, B and C, PEA 3 h). TNF was
also detectable on the surface of hepatocytes, but not inside these
cells, suggesting that macrophage-produced TNF bound to liver
parenchymal cells (Fig. 6, PEA 3 h). At later time points, i.e., 6 and
9 h after administration of PEA, hepatic TNF was not detectable by
immunostaining (data not shown). Twelve hours after challenge
with PEA the number of CD8-positive T cells within the liver was
slightly increased (Fig. 6C, PEA 12 h), whereas the number of
CD4-positive T cells remained unchanged (Fig. 6B, 12 h). Mac-
rophages had disappeared (Fig. 6A, PEA 12 h), and TNF was still
undetectable in liver sections (Fig. 6, PEA 12 h). This indicates
that the systemic TNF concentrations observed 12 h following
FIGURE 4. Electron micrographs of mouse livers 12 h after treatment with 300 g/kg PEA. a, Liver tissue containing an apoptotic cell with typical
condensed chromatin at the nuclear membrane, and chromatin emerged outside the nuclear membrane (arrows). Apoptotic bodies (AB) have been
phagocytized by neighboring hepatocytes. Hepatocytes contain numerous lipid droplets (L). Magnication, 3500. b, Necrotic liver area with two
polymorphonuclear neutrophils (PMN) and an apoptotic body (arrows). Hepatocytes contain numerous lipid droplets (L). Magnication, 5200. c, A
polymorphonuclear neutrophil (PMN), a lymphocyte (LYM), and two erythrocytes are seen in a liver sinusoid. Hepatocytes contain numerous lipid droplets
(L). Magnication, 3100. d, A lymphocyte (LYM) has moved between two hepatocytes, touching one of them. Hepatocytes contain numerous lipid
droplets (L). Magnication, 9000.
5748 T CELL DEPENDENCE OF Pseudomonas EXOTOXIN A HEPATOTOXICITY IN MICE

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

PEA injection were produced extrahepatically. T cells were essen-
tial for TNF production of liver macrophages. In mice depleted
from T cells by an anti-mouse Thy1.2 mAb (26) (see also below),
there was no detectable TNF production within the liver 3 h after
administration of PEA (Fig. 6, anti-Thy1.2/PEA 3 h), whereas
macrophage staining remained unaltered (Fig. 6A, anti-
Thy1.2/PEA 3 h). CD4- and CD8-positive T cells were completely
depleted from hepatic tissue in anti-Thy1.2-pretreated animals
(Fig. 6, B and C, anti-Thy1.2/PEA 3 h). Because there was no
staining by the anti-TNF Ab within the livers of untreated mice
and anti-Thy1.2-pretreated mice 3 h after PEA administration, un-
specic binding of this Ab to macrophages can be excluded. Taken
together, these results show that T cells are required for Kupffer
cell-dependent early TNF production following PEA intoxication
in mice.
To prove a functional involvement of TNF in the development
of PEA-induced liver injury, we pretreated mice with a specic
polyclonal anti-mouse TNF Ab that completely neutralized circu-
lating TNF following PEA challenge (TNF plasma concentration
12 h after PEA alone, 97 18 pg/ml vs undetectable amounts
(10 pg/ml) in the Ab-pretreated group; n 5). This Ab signif-
icantly inhibited DNA fragmentation as well as ALT release (Fig.
7A). Inhibition of internucleosomal DNA fragmentation in the
liver by anti-TNF Ab pretreatment was also demonstrated on an
agarose gel (Fig. 3, lanes 3 and 4). Moreover, TNFR1 knockout
(tnfr1) as well as TNFR2 knockout (tnfr2) mice were signi-
cantly protected from PEA-induced liver failure (Fig. 7B). In con-
trast to PEA-challenged tnfr1 mice, DNA fragmentation in livers
of PEA-treated tnfr2 mice was not signicantly reduced com-
pared with that in wild-type animals, indicating that hepatic apo-
ptosis is primarily mediated via TNFR1. However, because tnfr2
mice were signicantly protected from severe hepatic failure as
measured by ALT release, it seems likely that either costimulatory
signals provided by TNFR2 (35) or ligand passing (36) are nec-
essary for TNFR1-induced hepatocellular death.
Inhibition of transcription has been described to sensitize mice
toward i.v. administered recombinant murine TNF, which, in turn,
develop severe liver injury and eventually die (13, 14, 29). To
study whether PEA sensitizes the liver against TNF, we treated
mice with a subtoxic dose of PEA (3 g/kg) and recombinant
murine TNF (10 g/kg). Neither PEA nor TNF given alone in-
duced the release of liver enzymes, whereas the combination of
both resulted in severe liver injury (PEA alone: ALT, 36 8 U/L;
TNF alone: ALT, 46 12 U/L; PEA plus TNF: ALT, 12,795
2,496 U/L ( p 0.05); n 3). Similar results were obtained when
mice were treated with subtoxic doses of PEA (3 g/kg) plus LPS
(10 g/kg) or PEA (3 g/kg) plus SEB (2.5 mg/kg; PEA alone:
ALT, 36 8 U/L; LPS alone: ALT, 52 6 U/L; PEA plus LPS:
ALT, 4744 2227 U/L ( p 0.05); SEB alone: ALT, 10 3 U/L;
PEA plus SEB: ALT, 8,663 2,242 ( p 0.05) U/L; n 3). We
conclude that PEA sensitizes the liver toward TNF, e.g., induced
by LPS during systemic Pseudomonas infections or coinfections
with other bacteria.
T cell dependence of PEA-induced toxicity
In vitro, PEA was shown to be a polyclonal activator of cytolytic
T lymphocytes, effective against Con A-treated target cells (20).
Furthermore, as described above, T cells were required for Kupffer
cell-dependent early TNF production following PEA intoxication
in mice. To test whether T lymphocytes accounted for part of the
in vivo toxicity of PEA, we pretreated mice with the T cell-de-
pleting anti-Thy1.2 mAb (26). This treatment signicantly pro-
tected mice from PEA-induced liver damage (Fig. 8A) and pre-
vented hepatic internucleosomal DNA fragmentation (Fig. 3, lanes
3 and 5). Pretreatment with control Abs, i.e., rat IgG2a, failed to
protect mice from PEA-induced liver failure (data not shown).
Similarly, athymic BALB/c-nu/nu mice lacking mature T cells
were resistant toward PEA-induced liver injury (Fig. 8A). Circu-
lating cytokine levels, including TNF, were signicantly reduced
or were not detectable in anti-Thy1.2 mAb-pretreated or nu/nu
mice (Fig. 8B). In conclusion, the production of TNF and other
cytokines as well as liver injury clearly depends on T lymphocytes.
As cyclosporin A is known to suppress T cell activation, the role
of T cells in PEA-induced liver injury was supported by our nd-
ing that ALT release was inhibited in cyclosporin A-pretreated
compared with placebo-pretreated mice following PEA injection
(placebo-pretreated: ALT, 1679 247 U/L; cyclosporin A-pre-
treated: ALT, 544 47 U/L ( p 0.05); n 3).
Recently, Muraille et al. described a V-unrestricted, T cell-
unresponsive state together with an APC defect and a selective
decrease in splenic dendritic cell numbers in mice 48 h after SEB
injection (37). A similar unresponsive state has been described by
Miethke et al., who observed that SEB pretreatment protected mice
not only from lethality induced by the very same superantigen, i.e.,
SEB, but also by a third party superantigen with nonoverlapping
V specicity, i.e., toxic shock syndrome toxin-1 (38). Hence, we
wondered whether SEB pretreatment would also protect mice from
PEA-induced liver injury. Indeed, PEA-induced ALT release was
FIGURE 5. PEA-induced cytokine release. BALB/c mice received 300
g/kg PEA i.v.; at the time points indicated blood was withdrawn for
cytokine determination. All cytokines (TNF, IL-6, IFN-, and IL-2) were
determined by ELISA. Data are expressed as the mean SEM (n 36).
, p 0.05 vs 0 h.
5749 The Journal of Immunology

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

signicantly reduced in SEB-pretreated mice (Fig. 8C). The same
SEB pretreatment also protected the animals from GalN/SEB-de-
pendent transaminase release but not from GalN/LPS-induced
liver failure (Fig. 8C), again backing up a prominent role of T cells
for PEA hepatotoxicity. The release of TNF into the circulation,
determined at the time of maximal systemic TNF production (6,
39), was suppressed in SEB-pretreated mice challenged with PEA
or GalN/SEB, but not in SEB-pretreated mice challenged with
GalN/LPS (Fig. 8C).
To study a positive effect of T cell depletion on survival, mice
were pretreated with the anti-Thy1.2 mAb and challenged with 30
g/kg PEA i.v. Survival was signicantly prolonged by anti-
Thy1.2 mAb pretreatment (Fig. 8d). Mice treated with 30 g/kg
PEA developed liver injury within 24 h (ALT, 1227 271 U/L;
n 5) that was again prevented by anti-Thy1.2 mAb-dependent T
cell depletion (ALT, 139 18 U/L; n 5; p 0.05). The en-
hanced survival of the anti-Thy1.2 mAb-pretreated animals points
to a prominent role of T cells for PEA-induced toxicity. This nd-
ing was corroborated by a signicantly prolonged survival of SEB-
pretreated compared with nonpretreated mice following intoxica-
tion with 30 g/kg PEA (Fig. 8D).
Due to the fact that PEA was described to be a polyclonal ac-
tivator of cytolytic T cells in vitro (20) and our observation of a
slight increase in the amount of CD8-positive T cells in the liver
12 h after challenge with PEA (cf, Fig. 6C, PEA 12 h), we won-
dered whether effector molecules of cytolytic T cells might be
involved in PEA-induced hepatotoxicity. Perforin knockout ( per-
forin) mice (40, 41) were signicantly, but not completely, pro-
tected from PEA-induced liver injury (Fig. 9). In contrast, mice
lacking functional Fas/CD95/APO-1 (lpr/lpr) were more suscep-
tible to PEA than corresponding wild-type mice (Fig. 9). Two of
ve lpr/lpr mice did not survive the 12 h, while all wild-type
MRL/Mp mice were alive. These results suggest that CTL are at
least partially involved in PEA-induced hepatotoxicity using the
FIGURE 6. Confocal laser images of liver
sections from PEA-treated mice following immu-
nouorescent staining of TNF together with mac-
rophages, CD4

T cells, or CD8

T cells.
BALB/c mice were treated as indicated (PEA,
300 g/kg i.v.; anti-Thy1.2 mAb, 24 h before
PEA). Twelve-micrometer cryostat sections of
livers from these mice were double stained by use
of a polyclonal rabbit anti-mouse TNF Ab (sec-
ondary swine anti-rabbit IgG tagged with FITC,
green uorescence) together with one of the fol-
lowing Abs: A, rat anti-mouse macrophage, clone
BM8 (secondary goat anti-rat IgG tagged with
Texas Red, red uorescence); B, rat anti-mouse
CD4, clone RM4-5 (secondary goat anti-rat-IgG
tagged with Texas Red, red uorescence); and C,
rat anti-mouse CD8, clone Ly-2 (secondary goat
anti-rat IgG tagged with Texas Red, red uores-
cence). All sections were examined by confocal
laser scanning microscopy. Costaining is repre-
sented by yellow uorescence.
5750 T CELL DEPENDENCE OF Pseudomonas EXOTOXIN A HEPATOTOXICITY IN MICE

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

perforin-dependent pathway. Fas, in turn, expressed by the T cells
could possibly slow down the liver damage by mediating activa-
tion-induced T cell death (42, 43).
Discussion
Previous experimental work provided evidence for sensitization of
mice by inhibitors of hepatic transcription and translation, e.g., the
amino sugar GalN, toward either bacterial toxins such as LPS,
SEB, and toxic shock syndrome toxin-1 or their common mediator
TNF. All these models, which are often referred to as endotoxic or
T cell-mediated lethal shock in mice (6, 10, 13, 44, 45), have in
common the development of a fulminant hepatitis (7, 11, 12, 14,
29, 33, 45). The mechanism of sensitization has been discussed as
inhibition of synthesis of TNF-induced cytoprotective proteins
(29). The sensitizing agents used in these studies were either GalN,
which specically inhibits hepatic transcription (9, 33), or actinomy-
cin D (29) or -amanitin (11), which usually do not affect patients
suffering from septicemia.
Our results presented here demonstrate for the rst time that a
bacterial toxin affecting the host during bacteremia, i.e., the protein
synthesis inhibitory PEA, sensitizes the liver and probably also
other organs toward either LPS or SEB or their common mediator
TNF. This was shown by injection of subtoxic doses of PEA to-
gether with recombinant murine TNF, LPS, or SEB to mice and by
demonstrating the development of liver damage. Moreover, toxic
doses of PEA induced liver damage by themselves and stimulated
the release of TNF and other cytokines by the immune system.
Experiments using neutralizing Abs clearly identied TNF as a
central mediator of this hepatic failure. Liver injury was charac-
terized by early apoptosis that preceded the release of transami-
nases. A similar effect of TNF has been described in the case of
coadministration of actinomycin D and TNF (29), GalN and TNF
(33), or GalN and SEB (7). PEA-induced hepatocyte death was
inhibited by the unspecic caspase inhibitor Z-VAD.fmk, as it was
recently shown for GalN/TNF-induced apoptotic and secondary
necrotic liver cell death (31).
In contrast to LPS or SEB, which induce systemic peak con-
centrations of TNF within 1 or 3 h, respectively (39, 6), and ful-
minant hepatitis in GalN-sensitized mice within 8 h (7, 12), PEA-
induced plasma TNF was only measurable 12 h after injection of
the toxin, i.e., at a time when hepatic necrosis was already detect-
able. However, circulating cytokine levels only insufciently re-
ect local tissue concentrations, a concept that has been described
as the tip of the iceberg (46). Indeed, after administration of very
high doses of PEA to mice, plasma TNF was detectable by 90 min
(data not shown). Furthermore, following injection of 300 g/kg
PEA, local hepatic TNF production was observed by 3 h after
intoxication, as shown by means of immunouorescent staining.
This early appearing TNF was produced by Kupffer cells, but not
by T cells. However, T cells were required for Kupffer cell-de-
pendent early TNF production, providing evidence of cross-talk
between T lymphocytes and liver macrophages. Such cross-talk
has been previously suggested to occur upon Con A stimulation of
lymphocyte/macrophage cocultures in vitro (47). Interestingly, we
observed two phases of TNF production in PEA-treated mice. The
rst peak occurred after 3 h, and this TNF was found to be locally
produced by liver macrophages. The second increase, measurable
as systemic TNF, appeared after 12 h. A similar time course of
TNF release was recently demonstrated by Toyabe et al. in a T
cell-dependent hepatitis model in mice (48). It seems likely that
Kupffer cells become activated by resident liver T cells upon PEA
stimulation, because there was no additional inltration of CD4- or
CD8-positive T cells into liver tissue following PEA injection.
Twelve hours after PEA challenge, Kupffer cells were completely
depleted. This is probably a result of PEA-induced ongoing ADP
ribosylation of translation elongation factor 2 in these phagocytic
cells, which is the mechanism by which PEA inhibits protein syn-
thesis. In contrast, T cells seem to be resistant to direct killing by
PEA, because their cell number did not decrease in liver sections
even 12 h after PEA injection. This is in line with earlier in vitro
ndings showing only marginal and short-lived inhibitory effects
of PEA on protein synthesis in lymphoid cells (20). The number of
CD8-positive T cells in the liver even increased 12 h after PEA
challenge, providing a possible explanation for the dependence of
liver injury on perforin. These ndings support an earlier report
describing PEA as a polyclonal activator of cytolytic T lympho-
cytes in vitro (20). TNF was not detectable in liver sections 12 h
after administration of PEA, demonstrating that the systemic TNF
measurable at this time point was produced outside the liver. This
late TNF might be extensively synthesized by T cells in lymphoid
organs such as spleen or lymph nodes.
Besides the protective effect of anti-TNF Ab against PEA, fur-
ther evidence of the critical role of TNF was shown by the resis-
tance of tnfr1 and tnfr2 mice to PEA-induced hepatotoxicity. As
described earlier, the p55 TNFR1 mediates liver injury and lethal-
ity in GalN/LPS-treated mice (45, 49) as well as TNF-induced
hepatocyte apoptosis (50). In contrast, the blockade or the absence
of p75 TNFR2 failed to protect mice from GalN/LPS-induced le-
thality (49, 51), and tnfr2 mice exhibited enhanced sensitivity to
GalN/TNF-induced hepatic damage compared with the corre-
sponding wild-type mice (G. Tiegs and G. Kunstle, unpublished
observations). However, we recently found in a different TNF- and
T cell-dependent hepatitis model, which is induced by Con A and
is independent of transcriptional inhibition (34), that tnfr1 as well
as tnfr2 mice were protected from liver injury. This was explained
by providing evidence of a critical role of transmembrane TNF
(52) that has previously been described to cooperatively signal via
both TNFRs (35). Hence, PEA-induced hepatocellular toxicity is
possibly mediated by direct cytolytic action of Kupffer cell-derived
transmembrane TNF and its activation of both TNF receptors (35).
An alternative explanation may be given in view of the low TNF
concentrations observed following PEA injection and the concept
of ligand passing (36) from the high afnity TNFR2 to TNFR1,
which, in turn, transduces cytotoxic signaling. We have in vivo
evidence for the latter mechanism from experiments proving a sig-
nicant hepatoprotection of tnfr1, but enhanced sensitivity of
FIGURE 7. TNF as a mediator of PEA hepatotoxicity. A, BALB/c mice
received 300 g/kg PEA i.v. 15 min after pretreatment with neutralizing
anti-mouse polyclonal TNF Abs. Control animals received 0.1% HSA con-
taining saline instead of Abs. B, tnfr1 and tnfr2 mice as well as wild-type
mice of the corresponding strain C57BL/6 129 received 300 g/kg PEA
i.v. A and B, Twelve hours after challenge blood was withdrawn for ALT
determination, and liver homogenates were prepared for quantication of
oligonucleosomal hepatic DNA fragmentation with the help of a commer-
cially available cell death detection ELISA. Data are expressed as the
mean SEM (A, n 6; B, n 3). , p 0.05 vs control or wild-type
mice.
5751 The Journal of Immunology

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

tnfr2, mice following administration of very high doses of PEA
(data not shown).
The PEA-induced hepatotoxicity was clearly dependent on T
cells. This was proven by the protective effect of T cell depletion
by anti-Thy1.2 mAb, by the inability of nu/nu mice to develop
liver injury, by the protective effect of cyclosporin A, and by the
prevention of PEA-dependent liver failure following SEB pretreat-
ment. Notably, a substantial inhibition of protein synthesis by PEA
in mice was only observed in the liver, whereas the spleen was
much less affected (15). Furthermore, T cells were not depleted
from the liver by PEA in our experiments. These ndings suggest
that despite inhibition of hepatic protein synthesis, T cell stimula-
tion by PEA was still possible in vivo. The lack of T cells in nu/nu
mice or following anti-Thy1.2 mAb-dependent depletion corre-
lated with signicantly reduced levels of plasma cytokines upon
PEA challenge, and T cell depletion abolished early TNF produc-
tion by Kupffer cells, indicating that T lymphocytes are necessary
for the production of TNF and other cytokines. Suppression of T
cell activation by cyclosporin A provided protection from PEA-
induced liver injury, as it did in the case of Con A-induced, T
cell-dependent liver failure (26) or lethal toxicity of the superan-
tigen SEB in GalN-sensitized mice (6). Moreover, the induction of
an immunosuppressed state characterized by V-unrestricted T
cell unresponsiveness and defective APC functions 48 h after SEB
pretreatment (37) also provided protection from PEA-dependent
liver injury accompanied by suppression of TNF release. Because
SEB pretreatment only protects from the deleterious effects of T
cell mitogenic agents, not from LPS, this result again emphasizes
the key role of T cells in PEA-induced liver failure. Furthermore,
the prominent role of T cells in PEA-induced toxicity was shown
by the prolonged survival of anti-Thy1.2- as well as SEB-pre-
treated mice. However, T cell depletion was not life saving. Fur-
thermore, immediately before death, high activities of plasma
transaminases were detectable in T cell-depleted mice (data not
FIGURE 8. T cell dependence of PEA-induced hepatotoxicity, cytokine
release, and lethality. A, BALB/c mice pretreated with anti-Thy1.2 mAb
24 h before challenge as well as athymic nude/nude mice (nu/nu) of the
strain BALB/c received 300 g/kg PEA i.v. Twelve hours after challenge
blood was withdrawn for ALT determination, and liver homogenates were
prepared for quantication of oligonucleosomal hepatic DNA fragmenta-
tion with the help of a commercially available cell death detection ELISA.
Data are expressed as the mean SEM (n 5). , p 0.05 vs wild-type
control. B, In the same experiment as that described in A, the blood samples
were also employed for cytokine determinations. All cytokines (TNF, IL-6,
IFN-, and IL-2) were determined by ELISA. Data are expressed as the
mean SEM (n 5). , p 0.05 vs wild-type control. n.d., Not detect-
able. C, BALB/c mice pretreated with SEB 48 h before challenge received
PEA (300 g/kg i.v.) or GalN/SEB (GalN, 700 mg/kg i.p.; SEB, 2.5 mg/kg
i.p.) or GalN/LPS (GalN, 700 mg/kg i.p.; LPS, 10 g/kg i.p.). Control
animals were pretreated with 0.1% HSA instead of SEB. Twelve hours
(PEA) or 8 h (GalN/SEB and GalN/LPS) after challenge blood was with-
drawn for ALT determination. TNF was determined 12 h (PEA), 3 h
(GalN/SEB), or 1.5 h (GalN/LPS) following mouse treatment, respectively.
FIGURE 9. PEA hepatotoxicity in perforin-decient mice and Fas-mu-
tated lpr/lpr mice. Perforin mice, lpr/lpr mice, as well as mice of the
corresponding wild-type mice (C57BL/6 129 and MRL/Mp, respec-
tively) received 300 g/kg PEA i.v.; 12 h after challenge blood was with-
drawn for ALT determination. Data are expressed as the mean SEM
(n 3 for perforin; n 5 for lpr/lpr). Two of ve lpr/lpr mice did not
survive 12 h. , p 0.05 vs corresponding wild-type mice.
Data are expressed as the mean SEM (n 3). , p 0.05 vs 0.1%
HSA-pretreated mice. D, BALB/c mice received 30 g/kg PEA i.v. 24 h
after pretreatment with anti-Thy1.2 mAb or 48 h after pretreatment with
SEB, and survival was monitored. n, Number of mice per group. , p
0.05 vs saline-pretreated mice.
5752 T CELL DEPENDENCE OF Pseudomonas EXOTOXIN A HEPATOTOXICITY IN MICE

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

shown). Hence, liver damage might be the cause of death in these
mice also. This points to a multifactorial mechanism leading to
liver injury and death in PEA-treated mice that includes T cell
activation and production of TNF. The release of transaminases in
T cell-depleted mice at the time of death could be explained in
terms of ongoing ADP ribosylation of elongation factor-2 in hepa-
tocytes (15). Nevertheless, it has to be emphasized that T cell
activation signicantly contributes to PEA toxicity. In athymic
BALB/c nude mice, which were shown to be protected from acute
liver injury, survival was not signicantly improved compared
with that in wild-type BALB/c mice (data not shown). Again, at
the time of death, high amounts of plasma transaminases were
detectable in nude mice (data not shown). Hence, the results ob-
tained from nude mice corroborate the ndings from T cell-de-
pleted mice concerning the role of T cells in acute hepatotoxicity
and the signicance of liver failure for death. However, the lack of
prolonged survival of nude mice points to the existence of addi-
tional mechanisms that might have developed in these mice as a
consequence of a whole life without T cells. These differences
include higher amounts of NK cells and the presence of extrathy-
mically matured T cells, which possibly caused the increased sus-
ceptibility of nude mice compared with that of mice only decient
in responsive peripheral T cells.
Previous in vitro ndings suggested that PEA is a superantigen,
as it selectively stimulated the proliferation of murine thymocytes
expressing V8.2 in their TCR (21). PEA induced the proliferation
of murine splenocytes (20) (data not shown). However, we failed
to observe specic accumulation of either CD4

or CD8

T cells
expressing the V8.1 or V8.2 chain in their TCR (data not
shown). Hence, our preliminary results together with previous
ndings by others (25) argue against PEA being a superantigen.
Besides TNF, perforin was found to comediate PEA hepatotox-
icity, as suggested by a signicantly reduced susceptibility of per-
forin knockout mice to PEA-induced liver injury. TNF and other
proinammatory cytokines may indirectly activate cytotoxic T
cells (5355), thereby being responsible for additional hepatocy-
totoxic signals by the perforin/granzyme system. In contrast, the
second important mediator of cytotoxic lymphocytes, i.e., Fas li-
gand (40, 41), seems unlikely to induce hepatocellular damage in
our model; lpr/lpr mice lacking functional Fas suffered even more
from PEA toxicity than corresponding wild-type mice. This result
suggests a functionally different role of Fas in PEA-inducible T
cell-dependent liver injury compared with the results obtained by
Kondo et al., who demonstrated a protective effect of soluble
Fas-Fc in two different murine hepatitis models (56). One possible
explanation might be a suppression of Fas-dependent activation-
induced T cell death (42, 43) following PEA stimulation, thereby
aggravating liver injury via the TNF and perforin pathways. Pe-
ripheral clonal deletion of T cells has been shown to be impaired
in lpr/lpr mice following SEB (57) or Ag (58) injection in vivo.
Our results are well matched with previous ndings showing that
Fas-decient lpr/lpr mice were more susceptible than wild-type
mice to GalN/SEB-induced lethality (59), which is mediated by T
cells as well.
In conclusion, we propose a sequential stimulation of the im-
mune system by PEA resulting in hepatocytotoxicity (Fig. 10): T
cell activation by PEA results in the production of the central me-
diator TNF by Kupffer cells, which, in turn, induces apoptotic
signals in hepatocytes via the TNFR1, whereas TNFR2 possibly
passes the ligand or costimulatory signals to TNFR1. The inhibi-
tory effect of PEA on protein synthesis sensitizes the liver to the
cytotoxic action of TNF. TNF and other cytokines may also acti-
vate cytotoxic T cells, providing additional hepatocytotoxic signals
by the perforin/granzyme system. The continual immune stimula-
tory and hepatocytotoxic processes nally result in severe liver
failure, which causes lethality. Hence, protein synthesis inhibition
during Pseudomonas septicemia is likely to be a relevant sensiti-
zation mechanism to the cytotoxic action of TNF, which is pro-
duced not only upon direct stimulation of monocytes/macrophages
by LPS but also via T cell activation with PEA.
Acknowledgments
We thank Dr. H. Bluethmann (Hoffmann-La Roche, Basel, Switzerland)
for kindly providing TNF receptor knockout mice, as well as Drs.
J. Tschopp and M. Schroter (University of Lausanne, Institute of Biochem-
istry, Epalinges, Switzerland) for kindly providing perforin knockout mice.
We are indebted to Dr. G. R. Adolf (Bender & Co., Vienna, Austria) for
providing recombinant murine TNF and IFN-, and to Dr. A. Wendel
(University of Konstanz, Konstanz, Germany) for providing an antiserum
that neutralizes murine TNF. We thank Dr. W. Neuhuber (Institute of Anat-
omy, University of Erlangen-Nurnberg, Erlangen, Germany) for experi-
mental support regarding confocal laser scanning microscopy.
References
1. Pollack, M. 1990. Pseudomonas aeruginosa. In Principles and Practice of In-
fectious Diseases. G. L. Mandell, R. G. Douglas, and J. E. Bennett, eds. Churchill
Livingstone, New York, p. 1673.
2. Korvik, J. A., J. W. Marsh, T. E. Starzl, and V. L. Yu. 1991. Pseudomonas
aeruginosa bacteremia in patients undergoing liver transplantation: an emerging
problem. Surgery 109:62.
3. Cross, A. S., J. C. Sadoff, B. H. Iglewski, and P. A. Sokol. 1980. Evidence for the
role of toxin A in the pathogenesis of infection with Pseudomonas aeruginosa in
humans. J. Infect. Dis. 142:538.
4. Kohzuki, T., Y. Eguchi, M. Kato, K. Irie, H. Ohtsuka, A. Higuchi, and
H. Noguchi. 1993. Protective activity of anti-exotoxin A monoclonal antibody
against mice infected with toxin-producing Pseudomonas aeruginosa. J. Infect.
Dis. 167:119.
5. Cowley, R. A., J. R. Hankins, R. T. Jones, and B. F. Trump. 1982. Pathology and
pathophysiology of the liver. In Pathophysiology of Shock, Anoxia, and Ischemia.
R. A. Cowley and B. F. Trum, eds. Williams & Wilkins, Baltimore, p. 285.
6. Miethke, T., C. Wahl, K. Heeg, B. Echtenacher, P. H. Krammer, and H. Wagner.
1992. T cell-mediated lethal shock triggered in mice by the superantigen staph-
ylococcal enterotoxin B: critical role of tumor necrosis factor. J. Exp. Med. 175:
91.
7. Gantner, F., M. Leist, S. Jilg, G. Germann, M. A. Freudenberg, and G. Tiegs.
1995. Tumor necrosis factor-induced hepatic DNA fragmentation as an early
marker of T cell-dependent liver injury in mice. Gastroenterology 109:166.
8. Gantner, F., S. Kusters, A. Wendel, A. Hatzelmann, C. Schudt, and G. Tiegs.
1997. Protection from T cell-mediated murine liver failure by phosphodiesterase
inhibitors. J. Pharmacol. Exp. Ther. 280:53.
9. Decker, K., D. Keppler, J. Rudigier, and W. Domschke. 1971. Cell damage by
trapping of biosynthetic intermediates: the role of uracil nucleotides in experi-
mental hepatitis. Hoppe-Seylers Z. Physiol. Chem. 352:412.
10. Galanos, C., M. A. Freudenberg, and W. Reutter. 1979. Galactosamine-induced
sensitization to the lethal effects of endotoxin. Proc. Natl. Acad. Sci. USA 76:
5939.
FIGURE 10. Proposed sequential stimulation of the immune system by
PEA resulting in hepatotoxicity, liver damage, and death following admin-
istration to mice. For further explanation, see text.
5753 The Journal of Immunology

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

11. Keppler, D., W. Hagmann, S. Rapp, C. Denzlinger, and H. K. Koch. 1985. The
relation of leukotrienes to liver injury. Hepatology 5:883.
12. Wendel, A., and G. Tiegs. 1986. A novel biologically active seleno-organic com-
pound. VI. Protection by ebselen (PZ 51) against galactosamine/endotoxin-in-
duced hepatitis in mice. Biochem. Pharmacol. 35:2115.
13. Lehmann, V., M. A. Freudenberg, and C. Galanos. 1987. Lethal toxicity of li-
popolysaccharide and tumor necrosis factor in normal and D-galactosamine-
treated mice. J. Exp. Med. 165:657.
14. Tiegs, G., M. Wolter, and A. Wendel. 1989. Tumor necrosis factor is a terminal
mediator in galactosamine/endotoxin-induced hepatitis in mice. Biochem. Phar-
macol. 38:627.
15. Pavlovskis, O. R., and A. H. Shackelford. 1974. Pseudomonas aeruginosa exo-
toxin in mice: localization and effects on protein synthesis. Infect. Immun. 9:540.
16. Iglewski, B. H., P. V. Liu, and D. Kabat. 1977. Mechanism of action of Pseudo-
monas aeruginosa exotoxin A: adenosine diphosphate-ribosylation of mamma-
lian elongation factor 2 in vitro and in vivo. Infect. Immun. 15:138.
17. Morimoto, H., and B. Bonavida. 1992. Diphtheria toxin- and Pseudomonas A
toxin-mediated apoptosis: ADP ribosylation of elongation factor-2 is required for
DNA fragmentation and cell lysis and synergy with tumor necrosis factor-.
J. Immunol. 149:2089.
18. Vitetta, E. S., P. E. Thorpe, and J. W. Uhr. 1993. Immunotoxins: magic bullets
or misguided missiles? Trends Pharmacol. Sci. 14:148.
19. Vitetta, E. S., and P. E. Thorpe. 1991. Immunotoxins. In Biologic Therapy of
Cancer. V. T. DeVita, S. Hellman, and S. A. Rosenberg, eds. J. B. Lippincott Co.,
Philadelphia, p. 482.
20. Zehavi-Willner, T. 1988. Induction of murine cytolytic T lymphocytes by
Pseudomonas aeruginosa exotoxin A. Infect. Immun. 56:213.
21. Dixon, D. M., R. LeGrand, and M. L. Misfeldt. 1992. Selective activation of
murine V

.8.2 bearing T cells by Pseudomonas exotoxin A. Cell. Immunol. 145:

91.
22. Misfeldt, M. L., P. K. Legaard, S. E. Howell, M. H. Fornella, and R. D. LeGrand.
1990. Induction of interleukin-1 from murine peritoneal macrophages by Pseudo-
monas aeruginosa exotoxin A. Infect. Immun. 58:978.
23. Legaard, P. K., R. D. LeGrand, and M. L. Misfeldt. 1991. The superantigen
Pseudomonas exotoxin A requires additional functions from accessory cells for
T lymphocyte proliferation. Cell. Immunol. 135:372.
24. Legaard, P. K., R. D. LeGrand, and M. L. Misfeldt. 1992. Lymphoproliferative
activity of Pseudomonas exotoxin A is dependent on intracellular processing and
is associated with the carboxyl-terminal portion. Infect. Immun. 60:1273.
25. Fleischer, B., D. Gerlach, A. Fuhrmann, and K.-H. Schmidt. 1995. Superantigens
and pseudosuperantigens of gram-positive cocci. Med. Microbiol. Immunol.
184:1.
26. Tiegs, G., J. Hentschel, and A. Wendel. 1992. A T cell-dependent experimental
liver injury in mice inducible by concanavalin A. J. Clin. Invest. 90:196.
27. Bergmeyer, H. U. 1984. Methods of Enzymatic Analysis, 3rd Ed., Vol. 82. Verlag
Chemie, Weinheim, Germany, pp. 112, 416, and 444.
28. Kusters, S., F. Gantner, G. Kunstle, and G. Tiegs. 1996. Interferon- plays a
critical role in T cell-dependent liver injury in mice initiated by concanavalin A.
Gastroenterology 111:462.
29. Leist, M., F. Gantner, I. Bohlinger, P. G. Germann, G. Tiegs, and A. Wendel.
1994. Murine hepatocyte apoptosis induced in vitro and in vivo. by TNF- re-
quires transcriptional arrest. J. Immunol. 153:1778.
30. Wyllie, A. H. 1980. Glucocorticoid-induced thymocyte apoptosis is associated
with endogenous endonuclease activation. Nature 284:555.
31. Kunstle, G., M. Leist, S. Uhlig, L. Revesz, R. Feifel, A. MacKenzie, and
A. Wendel. 1997. ICE-protease inhibitors block murine liver injury and apoptosis
caused by CD95 or by TNF-. Immunol. Lett. 55:5.
32. Bohlinger, I., M. Leist, F. Gantner, S. Angermuller, G. Tiegs, and A. Wendel.
1996. DNA-fragmentation in mouse organs during endotoxic shock.
Am. J. Pathol. 149:1381.
33. Leist, M., F. Gantner, I. Bohlinger, G. Tiegs, P. G. Germann, and A. Wendel.
1995. Tumor necrosis factor-induced hepatocyte apoptosis precedes liver failure
in experimental murine shock models. Am. J. Pathol. 146:1220.
34. Gantner, F., M. Leist, A. W. Lohse, P. Germann, and G. Tiegs. 1995. Concanava-
lin A-induced T cell-mediated hepatic injury in mice: the role of tumor necrosis
factor. Hepatology 21:190.
35. Grell, M., E. Douni, H. Wajant, M. Lohden, M. Clauss, B. Maxeiner,
S. Georgopoulos, W. Lesslauer, G. Kollias, K. Pzenmaier, et al. 1995. The
transmembrane form of tumor necrosis factor is the prime activating ligand of the
80 kDa tumor necrosis factor receptor. Cell 83:793.
36. Tartaglia, L. A., and D. V. Goeddel. 1992. Two TNF receptors. Immunol. Today
13:151.
37. Muraille, E., T. De Smedt, F. Andris, B. Pajak, M. Armant, J. Urbain, M. Moser,
and O. Leo. 1997. Staphylococcal enterotoxin B induced an early and transient
state of immunosuppression characterized by V-unrestricted T cell unrespon-
siveness and defective antigen-presenting cell functions. J. Immunol. 158:2638.
38. Miethke, T., C. Wahl, K. Heeg, and H. Wagner. 1993. Acquired resistance to
superantigen-induced T cell shock. J. Immunol. 150:3776.
39. Barsig, J., S. Kusters, K. Vogt, H.-D. Volk, G. Tiegs, and A. Wendel. 1995.
Lipopolysaccharide-induced interleukin-10 in mice: role of endogenous tumor
necrosis factor-. Eur. J. Immunol. 25:2888.
40. Lowin, B., M. Hahne, C. Mattmann, and J. Tschopp. 1994. Cytolytic T-cell
cytotoxicity is mediated through perforin and Fas lytic pathway. Nature 370:650.
41. Kagi, D., F. Vignaux, B. Ledermann, K. Burki, V. Depraetere, S. Nagata,
H. Hengartner, and P. Golstein. 1994. Fas and perforin pathways as major mech-
anisms of T cell-mediated cytotoxicity. Science 265:528.
42. Dhein, J., H. Walczak, C. Baumler, K.-M. Debatin, and P. H. Krammer. 1995.
Autocrine T-cell suicide mediated by APO-1/(Fas/CD95). Nature 373:438.
43. Brunner, T., R. J. Mogil, D. LaFace, N. J. Yoo, A. Mahboubi, F. Echeverri,
S. J. Martin, W. R. Force, D. H. Lynch, C. F. Ware, et al. 1995. Cell-autonomous
Fas (CD95)/Fas-ligand interaction mediates activation-induced apoptosis in T-
cell hybridomas. Nature 373:441.
44. Miethke, T., K. Duschek, C. Wahl, K. Heeg, and H. Wagner. 1993. Pathogenesis
of the toxic shock syndrome: T cell mediated lethal shock caused by the supe-
rantigen TSST-1. Eur. J. Immunol. 23:1494.
45. Pfeffer, K., T. Matsuyama, T. M. Kundig, A. Wakeham, K. Kishihara,
A. Shahinian, K. Wiegmann, P. S. Ohashi, M. Kronke, and T. W. Mak. 1993.
Mice decient for the 55 kd tumor necrosis factor receptor are resistant to en-
dotoxic shock, yet succumb to L. monocytogenes infection. Cell 73:457.
46. Cavaillon, J. M., C. Munoz, C. Fitting, B. Misset, and J. Carlet. 1992. Circulating
cytokinesthe tip of the iceberg. Circ. Shock 38:145.
47. Gantner, F., M. Leist, S. Kusters, K. Vogt, H.-D. Volk, and G. Tiegs. 1996. T cell
stimulus-induced crosstalk between lymphocytes and liver macrophages results
in augmented cytokine release. Exp. Cell Res. 229:137.
48. Toyabe, S., S. Seki, T. Iiai, K. Takeda, K. Shirai, H. Watanabe, H. Hiraide,
M. Uchiyama, and T. Abo. 1997. Requirement of IL-4 and liver NK1

T cells for
concanavalin A-induced hepatic injury in mice. J. Immunol. 159:1537.
49. Sheehan, K. C. F., J. K. Pinckard, C. D. Arthur, L. P. Dehner, D. V. Goeddel, and
R. D. Schreiber. 1995. Monoclonal antibodies specic for murine p55 and p75
tumor necrosis factor receptors: identication of a novel in vivo role for p75.
J. Exp. Med. 181:607.
50. Leist, M., F. Gantner, S. Jilg, and A. W. Endel. 1995. Activation of the 55 kDa
TNF receptor is necessary and sufcient for TNF-induced liver failure, hepato-
cyte apoptosis, and nitrite release. J. Immunol. 154:1307.
51. Erickson, S. L., F. J. de Sauvage, K. Kikly, K. Carver-Moore, S. Pitts-Meek,
N. Gillett, K. C. F. Sheehan, R. D. Schreiber, D. V. Goeddel, and M. W. Moore.
1994. Decreased sensitivity to tumour-necrosis factor but normal T-cell devel-
opment in TNF receptor-2-decient mice. Nature 372:560.
52. Kusters, S., G. Tiegs, L. Alexopoulou, M. Pasparakis, E. Douni, G. Kunstle,
H. Bluethmann, A. Wendel, K. Pzenmaier, G. Kollias, et al. 1997. In vivo
evidence for a functional role of both TNF receptors and transmembrane TNF in
experimental hepatitis. Eur. J. Immunol. 27:2870.
53. Ridge, J. P., F. Di Rosa, and P. Matzinger. 1998. A conditioned dendritic cell can
be a temporal bridge between a CD4

T-helper and a T-killer cell. Nature 393:

474.
54. Bennett, S. R. M., F. R. Carbone, F. Karamalis, R. A. Flavell, J. F. A. P. Miller,
and W. R. Heath. 1998. Help for cytotoxic-T-cell responses is mediated by CD40
signalling. Nature 393:478.
55. Schoenberger, S. P., R. E. M. Toes, E. I. H. van der Voort, R. Offringa, and
C. J. M. Melief. 1998. T-cell help for cytotoxic T lymphocytes is mediated by
CD40-CD40L interactions. Nature 393:480.
56. Kondo, T., S. Takashi, H. Fukuyama, M. Adachi, and S. Nagata. 1997. Essential
roles of the Fas ligand in the development of hepatitis. Nat. Med. 3:409.
57. Scott, D. E., W. J. Kisch, and A. D. Steinberg. 1993. Studies of T cell deletion
and T cell anergy following in vivo administration of SEB to normal and lupus-
prone mice. J. Immunol. 150:664.
58. Singer, G. G., and A. K. Abbas. 1994. The Fas antigen is involved in peripheral
but not thymic deletion of T lymphocytes in T cell receptor transgenic mice.
Immunity 1:365.
59. Mountz, J. D., T. J. Baker, D. R. Borcherding, H. Bluethmann, T. Zhou, and
C. K. Edwards III. 1995. Increased susceptibility of fas mutant MRL-lpr/lpr mice
to staphylococcal enterotoxin B-induced septic shock. J. Immunol. 155:4829.
5754 T CELL DEPENDENCE OF Pseudomonas EXOTOXIN A HEPATOTOXICITY IN MICE

b
y

g
u
e
s
t

o
n

N
o
v
e
m
b
e
r

1
3
,

2
0
1
3
h
t
t
p
:
/
/
w
w
w
.
j
i
m
m
u
n
o
l
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m