Content uploaded by Mark A Kay
Author content
All content in this area was uploaded by Mark A Kay
Content may be subject to copyright.
JOURNAL OF VIROLOGY,
0022-538X/97/$04.0010
Nov. 1997, p. 8798–8807 Vol. 71, No. 11
Copyright © 1997, American Society for Microbiology
The Role of Kupffer Cell Activation and Viral Gene Expression in
Early Liver Toxicity after Infusion of Recombinant
Adenovirus Vectors
ANDRE LIEBER,
1
CHENG-YI HE,
1
LEONARD MEUSE,
1
DAVID SCHOWALTER,
1
IRINA KIRILLOVA,
2
BRIAN WINTHER,
1
AND MARK A. KAY
1,2,3
*
Division of Medical Genetics, Department of Medicine,
1
and Departments of Pathology
2
and Pediatrics,
3
University of Washington, Seattle, Washington 98195
Received 12 May 1997/Accepted 31 July 1997
Systemic application of first-generation adenovirus induces pathogenic effects in the liver. To begin unrav-
eling the mechanisms underlying early liver toxicity after adenovirus infusion, particularly the role of mac-
rophage activation and expression of viral genes in transduced target cells, first-generation adenovirus or
adenovirus vectors that lacked most early and late gene expression were administered to C3H/HeJ mice after
transient depletion of Kupffer cells by gadolinium chloride treatment. Activation of NF-kB, and the serum
levels of the proinflammatory cytokines tumor necrosis factor (TNF) and interleukin-6 (IL-6) were studied in
correlation with liver damage, apoptosis, and hepatocellular DNA synthesis. While Kupffer cell depletion
nearly eliminated adenovirus-induced TNF release, it resulted in a more robust IL-6 release. These responses
were greatly reduced in animals receiving the deleted adenovirus. Although there were quantitative differences,
NF-kB activation was observed within minutes of first-generation or deleted adenovirus vector administration
regardless of the status of the Kupffer cells, suggesting that the induction is related to a direct effect of the virus
particle on the hepatocyte. Early liver toxicity as determined by serum glutamic-pyruvic transaminase elevation
and inflammatory cell infiltrates appeared to be dependent on adenovirus-mediated early gene expression and
intact Kupffer cell function. Kupffer cell depletion had little effect on adenovirus-mediated hepatocyte apop-
tosis but did increase hepatocellular DNA synthesis. Finally, Kupffer cell depletion decreased the persistence
of transgene (human a1-antitrypsin [hAAT]) expression that was associated with a more pronounced humoral
immune response against hAAT. The elucidation of these events occurring after intravenous adenovirus
injection will be important in developing new vectors and transfer techniques with reduced toxicity.
Viral proteins expressed in cells after transduction with first-
generation adenoviruses elicit a host immune response leading
to inflammation of the target organ and extinction of transgene
expression (3, 5). The immune response in the liver following
systemic adenovirus administration is divided into two phases
(15, 28, 29). The first phase occurs between days 1 and 4
postinfection (p.i.) and is characterized in part by a periportal
polymorphonuclear leukocyte infiltration and elevated liver
enzymes, in some cases with lethal outcome, in response to
high viral doses. The second phase begins 5 to 7 days p.i. and
is associated with an antigen-dependent immune response spe-
cific to viral and/or transgene products. Substantial efforts have
been undertaken to study and modulate the specific immune
response against viral antigens (24, 29, 51, 52). However, little
is known concerning the mechanisms behind the early, innate
inflammatory response in the mouse liver following systemic
application of first-generation adenovirus vectors. Possible el-
ements of an antiviral innate immune response include the
activation of macrophages, NK cells, complement, and cyto-
kine release.
Kupffer cells (KC) are large liver macrophages. Because of
their topological localization within the liver sinusoids, they
represent the first line of defense against viruses entering the
liver through the portal circulation. KC functions are activated
by a variety of particles and substances, including viruses (Sen-
dai virus and Newcastle disease virus), bacterial lipopolysac-
charide (LPS), muramyl dipeptide, gamma interferon, and tu-
mor necrosis factor alpha (TNF) (for a review, see reference
11). The phagocytosis of parasites by KC is accompanied by the
release of proinflammatory cytokines that act primarily as a
paracrine signal on neighboring hepatocytes and induce che-
motaxis and aggregation of neutrophils. Furthermore, KC ex-
press class II major histocompatibility complex molecules, as
well as processing and presenting antigens.
Recently, KC activation by first-generation adenovirus was
reported, although active transduction of KC by adenovirus
was not demonstrated (47, 48). In these reports, KC function
was thought to be responsible for the elimination of ;90% of
viral vector DNA within the first 24 h after intravenous appli-
cation. Intravenous injections of gadolinium chloride (GdCl
3
),
a member of the rare earth metals, were able to prevent acti-
vation of KC by LPS (8, 18, 34). Hardonk et al. (18) speculated
that GdCl
3
formed a colloidal precipitate in the bloodstream at
neutral pH which is phagocytosed by KC and dissolved again in
the acidic environment within the macrophage lysosomes, re-
sulting in cell destruction. A single injection of GdCl
3
blocks
phagocytosis in more than 90% of KC and selectively elimi-
nates the large periportal macrophages (18). Repopulation of
KC by immature macrophages and monocytes begins 4 days
after injection (18). Macrophages in the red pulp of the spleen
are less vulnerable to GdCl
3
(18, 34). There is no evidence that
GdCl
3
exerts any direct toxic effects on hepatocytes, biliary
epithelial cells, endothelial cells, Ito cells, circulating mono-
cytes, or lymphocytes (34).
Activation of cytokines, especially TNF and interleukin-6
* Corresponding author. Mailing address: Division of Medical Ge-
netics, Department of Medicine, Box 357720, University of Washing-
ton, Seattle, WA 98195. Phone: (206) 685-9182. Fax: (206) 685-8675.
E-mail: mkay@u.washington.edu.
8798
(IL-6), is a central element of the innate immune response
against viruses (7, 14, 17). TNF is produced primarily by acti-
vated KC as a biologically active membrane-bound precursor
that is cleaved to produce the mature cytokine, which is re-
leased into the serum with systemic effects (17, 33). Released
TNF binds to two different receptors on a variety of target cell
types, including KC and hepatocytes. The two TNF receptors
p55 and p75 have different intracellular domains involved in
activation of different signal transduction pathways leading to
cell cycle progression, apoptosis, or differentiation. Moreover,
TNF activates KC and induces synthesis of leukotrienes that in
turn attract cells involved in the inflammatory and/or immune
response. TNF has at least two direct activities against wild-
type adenovirus: inhibition of virus replication and killing of
virus-infected cells and sensitization by E1a of cells to TNF-
induced p53-mediated apoptosis (17, 19, 40, 46). Four of the
;25 early adenovirus proteins (E1b-19K and E3-14.7K and
-10.4K/14.5K) prevent early TNF lysis (17). TNF is the princi-
pal mediator of endotoxemic shock, and in humans there is a
good correlation between high TNF serum levels and the se-
verity of adenovirus-induced pulmonary disease (31).
IL-6 is produced by a variety of cell types; a major source of
IL-6 following virus infection or LPS stimulation is splenic
macrophages (1, 35). A specific IL-6 receptor on hepatocytes is
implicated in acute-phase reactions in the liver and in the
regulation of liver regeneration after partial hepatectomy. IL-6
transcription is activated by TNF in KC, hepatocytes, and other
cell types (32, 38). Its synthesis is rapidly induced after pulmo-
nary adenovirus administration (14). Furthermore, IL-6 may
be involved in the activation of cytotoxic T lymphocytes (CTL)
that can affect immune responses against the vector, vector-
transduced cells, and/or transgene product (16).
NF-kB is a ubiquitous transcription factor which is activated
by a variety of stimuli, including viruses such as cytomegalovi-
rus (CMV), human immunodeficiency virus, hepatitis B virus,
Epstein-Barr virus and Sindbis virus (2, 30). It represents a
master switch for the cellular immune response against viruses.
This is achieved in part by its potential to coordinately trans-
activate transcription of inflammatory cytokine genes, includ-
ing the TNF and IL-6 genes (32, 38). NF-kB is also involved in
apoptotic pathways leading to elimination of virally infected
cells independent of an immune response (2).
Recently, we generated recombinant adenoviruses lacking
the genes for the immunogenic products encoded in the E1,
E2, E3, and L1-L4 regions (DBP, polymerase, pTP, hexon,
fiber, etc.) (29). The deleted adenovirus containing the human
a1-antitrypsin (hAAT) expression cassette (DAd.hAAT) can
be generated at high titers by a technique based on Cre/Lox
recombination. The small 9-kb-deleted viral genome is pack-
aged into capsids that are structurally similar to those of un-
deleted adenovirus vectors and can efficiently transduce mouse
hepatocytes in vivo. However, transgene expression in vivo
lasted only 2 days before the deleted genome was degraded.
Expression from deleted genomes can be stabilized in trans,
suggesting that viral proteins encoded in the excised region are
needed for genome persistence.
MATERIALS AND METHODS
Adenoviruses. Ad/RSVhAAT and DAd.hAAT were produced as previously
described (29). The viral titers given in transducing units were determined on
HeLa cells by hAAT immunofluorescence (29). The number of E1-deleted
helper viral particles was less than 5 plaques (on 293 cells)/10
6
transducing units
in the DAd.hAAT preparations. Viruses with a titer of 5 3 10
10
transducing
units/ml were stored at 280°C in 10 mM Tris-Cl (pH 8.0)–1 mM MgCl
2
–10%
glycerol. Tests for replication-competent virus were performed as described
previously (3). All adenovirus preparations were analyzed for endotoxin by using
the Mulus amebocyte lysate (Pyrotell) test (Associates of Cape Cod, Inc., Fal-
mouth, Mass.) according to the protocol of the manufacturer. The detection limit
of the test was 0.05 endotoxin unit/10
10
PFU/ml.
Animal studies. Animal studies were performed in accordance with the insti-
tutional guidelines set forth by the University of Washington. Five- to six-week-
old female C3H/HeJ and C3H-SCID mice (Jackson Laboratory, Bar Harbor,
Maine) were used. All animals were housed in specific-pathogen-free facilities.
Adenovirus injection was performed via tail vein infusion with 200 ml of adeno-
virus. Generally 10
10
transducing units, a dose that transduces ;100% of hepa-
tocytes (27), was injected per mouse. Blood samples for analysis were obtained
by retroorbital bleeding. Serum samples for cytokine and hAAT analysis were
stored at 280°C; samples for serum glutamic-pyruvic transaminase (SGPT) and
hAAT antibody analysis were stored at 4°C. For DNA replication studies, [meth-
yl-
3
H]thymidine (1 mCi/g of body weight) diluted in pyrogen-free physiologic
saline was injected intraperitoneally in 200 ml at 12 and 1 h before sacrifice.
GdCl
3
injections. For transient KC blockage, a protocol described previously
(18) was adapted. GdCl
3
was dissolved in H
2
O, and 10 mg/kg of body weight was
injected via the tail vein at 30 and 6 h prior to adenovirus administration in a total
volume of 200 ml. Control animals were injected with 200 ml of saline. To test the
phagocytic capacity of KC, colloidal carbon (0.8 ml/kg of body weight; Sigma, St.
Louis, Mo.) was injected 30 min before sacrifice in mice treated with GdCl
3
or
untreated at 0, 24, 48, and 72 h (18). The carbon uptake by liver macrophages,
scored by light microscopy of liver sections was reduced by ;80% in mice that
received the double GdCl
3
injection compared to untreated animals. Similar
results were obtained when the dose of GdCl
3
was 20 mg/kg. This is consistent
with results obtained in previous studies (8, 18, 34, 39).
Biochemical analysis of serum samples. Cytokine standards and antibodies
were from Pharmingen. For the IL-6 enzyme-linked immunosorbent assay
(ELISA), the rat anti-mouse IL-6 monoclonal antibody (MAb) MP5-20F3 was
used as the capture antibody and the biotinylated anti-IL-6 MAb MP5-32C11
was used as the detecting antibody. Binding was detected with avidin D-horse-
radish peroxidase (A-2004; Vector Laboratories). For the TNF ELISA, anti-
mouse TNF MAbs MP6-XT22 and Mp6-XT3 were used. The detection sensi-
tivities were 15 and 50 pg/ml for IL-6 and TNF, respectively. Serum hAAT
concentrations were determined by ELISA as previously described (22). A Sigma
diagnostic kit was used for colorimetric determination of the SGPT activity, using
10 ml of serum (Sigma procedure no. 505).
NF-kB electrophoretic mobility shift assay (EMSA). Nuclear extracts were
obtained from mouse livers as described previously (13) and stored at 280°C
until used. Protein concentrations were measured by the Bradford method. The
NF-kB binding sequence from the class I major histocompatibility complex
enhancer element (H2
k
) was used as a probe. Double-stranded oligonucleotides
were end labeled with [g-
32
P]ATP, using T4 polynucleotide kinase. Ten micro-
grams of nuclear protein was incubated with 0.2 ng of labeled oligonucleotide
probe for 30 min at room temperature and separated in 5% polyacrylamide-
Tris-glycine-EDTA gels. For antibody supershift assays, anti-p50- and anti-p65
specific polyclonal antibodies (Santa Cruz Biotechnology) were used. One mi-
crogram of the corresponding antibody was added to the samples after 30 min of
incubation with the labeled probe. Gels were dried and exposed to Kodak-AR
film for 12 h.
Histological analysis. For histological analysis, liver samples (left half of the
large upper lobe) were fixed in 10% neutral formalin, embedded in paraffin,
sectioned (6 mm), and stained with hematoxylin-eosin. Autoradiography after
[methyl-
3
H]thymidine labeling was performed on paraffin sections that were
dip-coated with Kodak NTB-2 emulsion diluted 1:1 (vol/vol) with water and
developed after a 2-week exposure as described previously (28). All slides were
counterstained with hematoxylin-eosin.
For terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling
(TUNEL) analysis, liver tissue was frozen in OCT compound (Miles, Inc.,
Elkhart, Ind.) and cryosectioned in 10-mm sections. A Boehringer Mannheim in
situ cell death detection kit was used to quantify apoptosis in hepatocytes as
specified by the manufacturer. Liver sections were counterstained with hema-
toxylin and analyzed by light microscopy.
Southern blotting. For genomic DNA preparation, mouse livers were flushed
with 5 ml of phosphate-buffered saline via the portal vein. After removal, livers
were homogenized, and a 100-mg portion was used for DNA extraction as
described previously (44). DNA concentrations were determined spectrophoto-
metrically. Ten micrograms of genomic DNA was digested with BamHI, run on
a 0.8% agarose gel, and electrotransferred to a Hybond membrane (Amersham).
The blots were hybridized in rapid hybridization buffer (Amersham) with an
[a-
32
P]dCTP-labeled hAAT probe, using a random priming kit (Gibco BRL). As
a control, DNA from uninfected animals was spiked with 10 pg of purified
Ad/RSVhAAT viral DNA and loaded on each gel. The relative amount of
adenovirus DNA was determined by phosphorimager analysis as a ratio between
sample signal and control signal. Small variations in DNA loading or transfer
between lanes were adjusted by rehybridization of the blots with probes for the
mouse metallothionein gene. The following plasmid DNA fragments were used
for labeled probes: 1.4-kb hAAT cDNA (EcoRI fragment of pAd.RSVhAAT
(23) and 2-kb mouse metallothionein I gene (HindIII/EcoRI fragment of
pmMMT-I (44).
hAAT antibodies. Anti-hAAT antibodies were determined by ELISA as de-
scribed previously (37). Briefly, ELISA plates coated with anti-hAAT capture
MAb were blocked and then incubated with hAAT protein (calibrator serum no.
VOL. 71, 1997 INNATE IMMUNITY AFTER ADENOVIRUS VECTOR ADMINISTRATION 8799
4; Atlantic Antibodies, Stillwater, Minn.) diluted 1:50 in blocking buffer for 2 h
at room temperature. For each two-sample well, two additional wells were mock
loaded with blocking buffer to determine whether individual high serum hAAT
levels would interfere with the assay. Mouse serum samples, diluted 1:1,000, were
loaded onto the plate along with a similarly diluted naive serum as a negative
control. Murine immunoglobulin G2b anti-hAAT MAb 178260 (Calbiochem, La
Jolla, Calif.) serially diluted in blocking buffer from 10
22
to 10
26
was also
included on each plate as a positive control. Following another2hofincubation
at room temperature, the plates were incubated with horseradish peroxidase-
labeled sheep anti-mouse immunoglobulin whole-molecule antibody (A-6782;
Sigma) for another 2 h.
RESULTS
Block of KC function. GdCl
3
was extensively used in liver
regeneration studies of rats and mice as an agent to block KC
function (8, 18, 34, 39). In agreement with these studies, we
confirmed that more than 75% of KC in the mouse liver can be
blocked by two intravenous injections of GdCl
3
at 30 and 6 h
before analysis (see Materials and Methods). The block lasted
for at least 3 days after the second injection. GdCl
3
had no
obvious side effects on the animals, as the levels of SGPT,
TNF, IL-6, and NF-kB activation were indistinguishable from
those for control mice at the time of analysis (data not shown).
Because bacterial LPS can copurify together with adenovirus
particles and influence the parameters measured, only virus
preparations that tested negative for LPS endotoxin were used
in these studies. Furthermore, all animal studies were per-
formed with C3H/HeJ mice, which are known to be resistant to
bacterial endotoxin due to a mutation in the LPS-responsive
gene in chromosome 4 (45).
Cytokine release. To determine the cytokine profiles in mice
in response to adenovirus, Ad/RSVhAAT, a first-generation
adenovirus, or DAd.hAAT, lacking E1, E2, E3, and late gene
expression (29), were injected at a dose of 10
10
transducing
particles per mouse after GdCl
3
or saline injection. In mice
that received Ad/RSVhAAT without GdCl
3
treatment, serum
TNF concentrations were elevated by ;13-fold over baseline
at 3 h after adenovirus administration, decreased at 12 h, and
increased again, with a major peak at 36 h (;20-fold above
normal) (Fig. 1A). KC depletion almost completely prevented
the early TNF release at 3 h and significantly reduced the
second peak (fourfold above baseline). DAd.hAAT injection
induced early TNF release, but to a lesser degree (;7.5-fold)
than Ad/RSVhAAT injection. Importantly, the second peak at
36 h was clearly less pronounced in animals injected with the
vector depleted for viral gene expression. GdCl
3
-pretreated
mice that received DAd.hAAT had baseline TNF levels.
Serum IL-6 concentrations were elevated ;20-fold, with a
peak about 12 h after infusion of Ad/RSVhAAT, and then
gradually decreased during the next 4 days, reaching levels
4-fold above baseline (Fig. 1B). Infusion of DAd.hAAT caused
a slight IL-6 elevation (;2 times normal) during the first 24 h.
Interestingly, GdCl
3
pretreatment significantly stimulated IL-6
release after injection of Ad/RSVhAAT and DAd.hAAT.
NF-kB is a transcription factor present in an inactive cyto-
plasmic form in almost all cell types (2). Upon stimulation, the
NF-kB dimer is released from its inhibitory subunit IkB and
translocated to the nucleus, where it binds to specific DNA
sites. Measuring the DNA binding activity in nuclear (liver)
cell extracts by EMSA is a means to analyze the activation
status of NF-kB (Fig. 2). Supershift analysis using antibodies to
the NF-kB subunits p50 and p65 (not shown) designated the
upper band as a p50/p65 heterodimer whose function is im-
portant in terms of transcriptional transactivation (2). The
NF-kB binding activity in nuclear liver cell extracts was ana-
lyzed at different time points: 15 min, 3 h, and 3 days after
adenovirus infusion. There was a remarkable NF-kB (p50/p65)
activation immediately after infusion of Ad/RSVhAAT or
DAd.hAAT in both GdCl
3
-treated and untreated mice. The
heterodimer p65/p50 binding activity was significantly lower by
3 h postinfusion. At day 3 after Ad/RSVhAAT injection, we
observed a slight increase in NF-kB binding activity that was
more pronounced in mice not receiving GdCl
3
treatment.
NF-kB binding activity was almost absent at day 3 in animals
receiving DAd.hAAT.
Liver toxicity. We reported earlier (29) that intravenous
injection of 10
10
PFU of Ad/RSVhAAT resulted in biphasic
elevations in levels of SGPT, an early and sensitive marker for
hepatocyte injury. Early elevations in SGPT were detected
within the first 36 h after adenovirus administration, followed
by chronic SGPT elevation lasting as long as viral and trans-
gene expression were detectable. The second phase did not
occur in immunodeficient animals, suggesting that it was the
result of antigen-specific immunity. Here we focused on the
early phase of adenovirus-induced liver toxicity (Fig. 3). In
C3H mice, SGPT levels were elevated by ;15-fold during the
first 2 days after infusion of Ad/RSVhAAT. In GdCl
3
-pre-
treated mice, SGPT levels were increased only ;3-fold at 36 h
postinfusion. Injections of DAd.hAAT resulted in an ;2-fold
increase in the serum concentrations of the enzyme (with a
small peak around 3 h). SGPT levels were normal in GdCl
3
-
treated mice that received DAd.hAAT. Taken together, these
FIG. 1. Serum cytokine concentrations. Serum TNF (A) and IL-6 (B) serum concentrations after injection of Ad/RSVhAAT or DAd.hAAT in C3H mice with
(1Gd) GdCl
3
or without (2Gd) GdCl
3
administration at 30 and 6 h before adenovirus (n 5 4 animals per group). Gd, GdCl
3
without adenovirus.
8800 LIEBER ET AL. J. VIROL.
results suggest that adenovirus gene expression and interac-
tions with KC are linked to hepatocyte injury.
To investigate whether the events triggered by adenovirus
infection in vivo can lead to hepatocyte death by apoptosis,
TUNEL studies were carried out on liver sections obtained at
different times after adenovirus infusion. Representative mi-
crophotographs shown in Fig. 4b revealed that at 15 min and
3 h (data not shown), TUNEL signals were detected only in
small nonparenchymal cells in mice treated with GdCl
3
. Posi-
tive signals in hepatocytes were found at day 3 p.i., with ;1
to 3% apoptotic hepatocytes in livers of mice receiving Ad/
RSVhAAT in both GdCl
3
-treated and untreated animals
(Fig. 4c to f). Fewer apoptotic cells (,,1%) were found in
mice infused with DAd.hAAT (Fig. 4g and h).
A neutrophil/lymphocyte infiltrate that was most pro-
nounced in the periportal regions was found in liver sections
from mice after infusion of Ad/RSVhAAT (Fig. 4c). Notably
fewer inflammatory cells were present at day 3 in livers from
GdCl
3
-pretreated mice that received Ad/RSVhAAT and in
mice after infusion of DAd.hAAT.
The liver responds to damage and cytokine release with
regeneration, including hepatocellular DNA synthesis. The
percentage of regenerating hepatocytes was quantified by use
of [
3
H]thymidine, the incorporation of which into DNA during
replication can be detected in liver sections (Fig. 5). The peak
number of replicating hepatocytes (69 and 37% in two different
animals [Fig. 5b and d]) was found at day 4 after Ad/RSV
hAAT infusion, with a trend toward greater DNA replication
after GdCl
3
pretreatment (87 and 49% [Fig. 5c and e]).
DAd.hAAT infusion induced DNA synthesis in less than 1% of
hepatocyte nuclei in livers without GdCl
3
pretreatment and in
;2% in GdCl
3
-injected mice (Fig. 5f and g). [
3
H]thymidine
incorporation before and after day 4 was significantly reduced
in all animal groups, and naive controls had less than 0.01%
labeling in hepatocytes (not shown).
The extent of hepatocyte replication after infusion of first-
generation adenovirus is comparable to that observed during
liver regeneration after partial hepatectomy. In both situations,
NF-kB activation and/or high IL-6 levels could be among the
initial triggers that induce hepatocellular DNA synthesis (50).
Interestingly, KC depletion enhanced the adenovirus-induced
DNA synthesis. This finding is in agreement with published
data; treatment that inhibits TNF release from KC increases
DNA replication after partial hepatectomy (9, 10, 34).
Effects on level and persistence of transgene expression.
Innate immune mechanisms clearly influence the antigen-de-
pendent responses by affecting antigen presentation and re-
cruitment of immunologic effector cells. This may alter the
final fate of adenovirus-transduced cells and transgene expres-
sion. We wanted to determine whether GdCl
3
treatment that
depleted KC would influence antigen-specific immune mech-
anisms that limit transgene expression after adenovirus-medi-
ated liver gene transfer. To explore this possibility, we followed
serum hAAT concentrations after Ad/RSVhAAT injection in
C3H mice with and without prior KC depletion. From earlier
studies, it was known that this mouse strain represents a so-
called short expressor because the transgene product, serum
hAAT, declined to undetectable levels within the first 2 to 3
FIG. 2. NF-kB EMSA. Ad/RSVhAAT or DAd.hAAT was infused with or without GdCl
3
treatment. Controls (Gd) received GdCl
3
without adenovirus. Mice were
sacrificed at 15 min, 3 h, and 3 days (3d) postinfusion, and nuclear extracts from livers were analyzed by EMSA, using 10 mg of nuclear protein per lane. The positions
of the NF-kB p50/p65 heterodimer and p50 homodimer are indicated. Reticulocyte lysate (RL) was used as a marker to determine the position of the p50/p65 complex
(13). Each lane represents an individual animal.
FIG. 3. Liver injury. SGPT concentrations were determined in mice infused
with Ad/RSVhAAT or DAd.hAAT. 1Gd, after GdCl
3
pretreatment; 2Gd, with-
out GdCl
3
. Controls (Gd) received GdCl
3
without adenovirus infusion. n 5 3
animals per point.
VOL. 71, 1997 INNATE IMMUNITY AFTER ADENOVIRUS VECTOR ADMINISTRATION 8801
FIG. 4. TUNEL apoptosis assays. Apoptotic cell death was analyzed by the TUNEL technique in liver sections. (a) Liver section without adenovirus or GdCl
3
treatment; (b) GdCl
3
treatment only; (c and e) livers from two animals at day 3 after Ad/RSVhAAT without GdCl
3
pretreatment; (d and f) livers from two animals
at day 3 after Ad/RSVhAAT with GdCl
3
pretreatment; (g) liver section after DAd.hAAT without GdCl
3
; (h) liver section at day 3 after DAd.hAAT with GdCl
3
. Slides
are counterstained with hematoxylin. Note the neutrophil infiltration in panels c and e.
8802
FIG. 5. Hepatocellular DNA synthesis. Analysis of hepatocellular DNA synthesis in liver sections at day 4 after adenovirus infusion. A total of 10
10
transducing
particles of Ad/RSVhAAT (b to e) or DAd.hAAT (f to g) were infused in mice after GdCl
3
pretreatment (c, e, and g) or without GdCl
3
pretreatment (b, d, and f).
(a) Liver section after GdCl
3
administration only. At 12 and 1 h before sacrifice, the animals were infused with [methyl-
3
H]thymidine. Liver sections were exposed to
film emulsion for 2 weeks and counterstained with hematoxylin-eosin.
8803
months after administration of Ad.RSVhAAT (3). In our stud-
ies, serum hAAT concentrations dropped after 80 to 100 days
in C3H mice and were persistent in C3H-SCID mice (Fig. 6).
Interestingly, in GdCl
3
-pretreated mice, hAAT levels were ini-
tially two to three times higher but declined to zero ;50 days
earlier than in nontreated C3H mice.
To investigate the source for these differences in transgene
expression, viral DNA was quantified in preparations of geno-
mic liver DNA at 5, 60, and 100 days after infusion of Ad/
RSVhAAT (Fig. 7). In GdCl
3
-treated mice, the concentration
of Ad/RSVhAAT DNA was ;2.5-fold higher at day 5 than in
untreated mice. This is in agreement with data from Worgall et
al. (48), who reported that .90% of vector DNA is lost as a
result of KC function soon after adenovirus administration. At
60 and 100 days postinfusion, viral DNA was still detectable at
low levels in livers from C3H mice; these levels were compa-
rable with vector DNA concentrations in livers from C3H-
SCID mice. Thus, at time points when serum hAAT was no
longer detectable, transduced vector DNA was still present in
the livers of C3H mice. This finding suggests that CTL re-
sponses directed against the transduced cells are not respon-
sible for loss of gene expression. One possible explanation is
that hAAT gene expression is blocked at the level of transcrip-
tion, translation, or posttranslational processing. Tsui et al.
(42) reported that CTL-derived cytokines can induce posttran-
scriptional clearance of hepatitis B virus RNA in infected
hepatocytes. We did not investigate this possibility but concen-
trated on an observation recently made by Schowalter et al.
(reference 37 and unpublished results), who demonstrated that
antibodies to the expressed transgene product (hAAT) reduce
the level of detectable serum hAAT. To evaluate this possibil-
ity, we determined anti-hAAT antibody titers in the serum of
C3H mice with and without prior GdCl
3
treatment (Table 1).
Early production of antibodies against hAAT was suppressed
in GdCl
3
-treated mice, possibly due to inhibition of antigen
presentation by KC. At later time points (weeks 4 and 10), the
concentration of hAAT-specific antibodies was higher in
GdCl
3
-treated mice. We hypothesize that other antigen-pre-
senting cells (e.g., splenic macrophages) were stimulated in
KC-depleted mice or that viral antigens and/or transgene prod-
ucts synthesized in transduced hepatocytes were taken up by
the replaced KC. As a result, perhaps an enhanced humoral
immune response to the transgene products was responsible
for the early falloff of serum hAAT concentrations in GdCl
3
-
treated C3H mice.
DISCUSSION
The pathologic changes in the liver that occur soon after
adenovirus infusion result from a combination of a direct cy-
totoxic effect of expressed viral proteins and innate immune
defenses to virus infection. We analyzed the changes in some
of the important elements of the innate immune response,
NF-kB, TNF, and IL-6, in correlation to liver damage and
transgene expression after injection of a first-generation ade-
novirus (Ad/RSVhAAT) and a vector deleted for E1, E2, E3,
and late gene expression (DAd.hAAT) in mice that received
GdCl
3
, an agent that selectively eliminates large periportal KC
for 3 to 4 days after intravenous administration.
We observed a biphasic elevation of serum TNF, with the
first peak occurring shortly after adenovirus administration
and a second major peak at 36 h. We assume that the peak at
3 h represents TNF release as a result of KC activation occur-
FIG. 6. Transgene expression in mice. Serum levels of hAAT after in vivo
gene transfer. C3H mice were injected with 10
10
transducing particles of Ad/
RSVhAAT after GdCl
3
pretreatment (filled squares) or without GdCl
3
pretreat-
ment (open squares). Serum samples were collected periodically and analyzed
for hAAT by ELISA. For comparison, the same dose of Ad/RSVhAAT was
injected into immunodeficient C3H-SCID mice without GdCl
3
(crosses). Each
line represents an individual animal.
FIG. 7. DNA analysis. (A) Southern blot analysis of transduced adenovirus
vector DNA in genomic liver DNA. Animals (C3H or C3H-SCID mice) were
sacrificed at different time points (days 5, 60, and 100) after infusion of 10
10
transducing particles of Ad/RSVhAAT after GdCl
3
pretreatment (filled squares)
or without GdCl
3
pretreatment (open squares). Crosses represent SCID mice
without GdCl
3
. Ten micrograms of BamHI-digested genomic DNA was loaded
in each lane. Blots were hybridized with a
32
P-labeled 1.4-kb hAAT probe. The
18.3-kb band is a specific BamHI fragment containing the hAAT expression
cassette. Liver DNA for the 100-day time point was obtained from animals that
were monitored for serum hAAT in Fig. 6. The filters for the day 5 and day 60
and 100 time points were exposed for 2 and 72 h, respectively. (B) DNA quan-
tification. The relative amounts of adenovirus DNA were determined by phos-
phorimager analysis and expressed as a ratio between the sample signal and a
control signal (10 pg of Ad/RSVhAAT DNA).
8804 LIEBER ET AL. J. VIROL.
ring during binding and internalization of adenovirus particles
(Ad/RSVhAAT and DAd.hAAT) because it can be blocked by
GdCl
3
. The nature of the second TNF peak is correlated with
the onset of early viral gene expression in Ad/RSVhAAT-
transduced cells (29); it was significantly reduced when viral E2
gene expression was absent (DAd.hAAT). The major source of
the second peak of TNF probably represents de novo produc-
tion induced by early viral gene expression in cells transduced
with Ad/RSVhAAT.
Liver damage measured by SGPT elevation is temporally
correlated with serum TNF elevations and can be substantially
prevented by GdCl
3
treatment. Thus, we conclude that TNF
release/synthesis after KC activation is one of the etiologic
factors for adenovirus-induced liver pathology. In this context,
de novo synthesis of early viral antigens contributed, probably
via TNF stimulation, to the observed hepatocellular injury,
because intravenous infusion of DAd.hAAT particles did not
lead to significant hepatocellular damage. Notably, hepatocel-
lular apoptosis is not the major mechanism in TNF-induced
liver injury after administration of first-generation adenovi-
ruses. In contrast, infection with E1a-competent (wild-type)
adenovirus sensitizes cells to apoptosis by TNF (17, 46). The
infiltration of inflammatory cells observed in liver sections at
day 3 after Ad/RSVhAAT infusion was efficiently blocked by
GdCl
3
, pretreatment. Reduced TNF release as a result of KC
depletion could be among the factors that inhibit chemotaxis of
neutrophils.
Intravenous administration of the first-generation adenovi-
rus, Ad/RSVhAAT, resulted in substantial increases in serum
IL-6 concentrations over the analyzed time period of 4 days,
with a peak at 12 h postinfusion. GdCl
3
pretreatment did not
block but rather enhanced IL-6 release, suggesting that the
IL-6 production originated from cells other than KC. Possible
candidates are splenic macrophages (not affected by GdCl
3
)or
endothelial cells, which may receive a greater viral load in
KC-depleted mice. This could lead to their enhanced activa-
tion and cytokine production. High IL-6 levels affect hepato-
cyte metabolism (1) and stimulate CTL activation and infiltra-
tion (16). Further experiments are required to test this
hypothesis.
We observed a strong NF-kB activation in the liver after
intravenous adenovirus infusion. The mechanisms by which
adenovirus activates NF-kB are not known. Since the activa-
tion occurs within 20 min after vector administration, it is
unlikely that viral gene expression is involved. Moreover,
DAd.hAAT exerts the same effect, indicating that the virus
particle itself triggers the process. The same NF-kB activation
pattern is observed in mice after KC depletion, suggesting that
activation takes place in hepatocytes. We speculate that fiber
binding to the adenovirus receptor or subsequent events such
as the interaction of pentons with integrins or endosome lysis
activate kinases that phosphorylate IkB. A possible candidate
for such a kinase is the Raf/mitogen-activated protein kinase
(6). There was no correlation between the observed NF-kB
activation occurring in both control and KC-depleted mice and
the increased TNF and SGPT levels. This is different than in
other studies, where at later time points, the activation and
release of inflammatory cytokines could be prevented by block-
ing NF-kB based on ectopic overexpression of a nucleus-local-
ized IkB after adenovirus-mediated gene transfer (49).
To make space for cloning larger inserts, E3 deletions were
introduced into first-generation adenovirus vectors, based on
the premise that E3 proteins are not necessary for the adeno-
virus life cycle (4). However, studies by Ginsberg et al. (14) and
Sparer et al. (40) in a mouse model for adenovirus pneumonia
demonstrated that adenovirus lacking the E3-TNF resistance
genes (14.7K, 10.4/14.5K, and gp19K) induce a more severe
pulmonary disease, characterized by alveolar infiltration, than
wild-type virus. Importantly, the E3 promoter is the only ade-
novirus promoter that contains binding sites for NF-kB, allow-
ing for E1a-independent expression (12). Based on our results,
it is appealing to consider that during adenovirus infection,
NF-kB activates the E3 promoter and thereby induces the E3
proteins that block the antiviral and cytotoxic effect of TNF in
hepatocytes. Moreover, a recent study (19) shows that antigen-
dependent immunity may be reduced with constitutive E3 gene
expression; however, the exact mechanism and its effects on
early liver toxicity are not known. Thus, a restoration of E3
gene expression under the endogenous E3 promoter may re-
duce the hepatotoxicity of first-generation adenoviruses.
Another practical consequence of NF-kB activation in hepa-
tocytes after adenovirus infusion is that certain (viral) promot-
ers with NF-kB binding sites, for example, the CMV promoter
(36), that were used to drive transgene expression were acti-
vated by NF-kB binding. The observation that after adenovi-
rus-mediated liver gene transfer the CMV promoter is only
transiently active in hepatocytes may be related to NF-kB
activation (21, 22). This possibility should be taken into con-
sideration when viral promoters are used for transgene expres-
sion.
The etiology of hepatocellular apoptosis observed at day 3
after adenovirus infusion is not clear. The number of apoptotic
cells is relatively low in livers after DAd.hAAT or AdRSV
hAAT. However, both vectors still contain the E4 region, and
expression of E4 proteins in transduced hepatocytes could be
responsible for the low level of apoptosis seen in liver sections
at day 3 after vector infusion (41). On the other hand, apopto-
sis could be linked to NF-kB activation occurring immediately
after adenovirus infusion. Although NF-kB activation has been
shown for many viruses, only its activations by Sindbis virus
and dengue virus (30) are definitive examples of virus-induced
apoptosis.
Although we did not study the effects of GdCl
3
treatment on
the development of an antigen-specific cellular immune re-
sponse systematically, we observed an enhanced humoral re-
sponse to the transgene product hAAT in KC-depleted mice at
later time points. Thus, blocking elements of the innate im-
mune response by depletion of KC while reducing early ade-
novirus-mediated liver toxicity shortens transgene expression
due to an enhanced antigen-specific humoral immune re-
TABLE 1. Anti-hAAT MAb levels
Treatment
Serum anti-hAAT IgG
a
Wk 2 p.i. Wk 4 p.i. Wk 10 p.i.
Ad/RSVhAAT 2 Gd 11, 11, 11 11, 111, 11 11, 111, 11
Ad/RSVhAAT 1 Gd 1, 2, 111,111, 111 1111, 111, 1111
a
Semiquantitative measurement for each animal in comparison to a standard curve of anti-hAAT MAbs and expressed as not detected (2) or detection similar to
that of a standard MAb diluted 1:10,000 (1), 1:1,000 (11), 1:100 (111), or 1:10 (1111).
VOL. 71, 1997 INNATE IMMUNITY AFTER ADENOVIRUS VECTOR ADMINISTRATION 8805
sponse. This is probably a result of virus spillover into other
organs like the spleen or may be related to repopulation of KC
and antigen presentation from viral proteins leaking from the
hepatocytes into these cells.
Recently Wolff et al. (47) and Kuzmin et al. (26) analyzed in
BALB/c mice the effect of macrophage depletion by intrave-
nous injection of liposome-encapsulated dichloromethylene bi-
phosponate (Cl
2
MBP) on transgene expression after adenovi-
rus-mediated gene transfer; however, its effect on the innate
immune response and the correlation to liver pathology were
not investigated. The authors found an inhibition of early an-
tibody production and increased transgene expression level in
KC-depleted mice compared with untreated mice. However, in
contrast to our results, KC depletion by Cl
2
MBP-liposomes in
BALB/c mice extended the persistence of hAAT expression
after adenovirus-mediated gene transfer (48). The humoral
and cellular immune responses to viral and transgene product
vary between different mouse strains (3). The decline in serum
hAAT levels observed in C3H mice appears to be related to an
antibody response against the transgene product, whereas in
BALB/c mice, the loss of DNA in transduced hepatocytes is
thought to be responsible for extinction of gene expression
(36a). Second, the block of reticuloendothelial cell function by
Cl
2
MBP-liposomes and GdCl
3
is qualitatively different; where-
as only large liver macrophages are vulnerable to GdCl
3
,
Cl
2
MBP-liposomes affect other macrophage populations as
well, including macrophages of the spleen, an organ that does
receive adenovirus after intravenous adenovirus administra-
tion (29) and is known to secrete a different spectrum of cy-
tokines (35). Third, macrophage studies of BALB/c mice may
be problematic, because macrophages of this inbred strain
have a reduced capacity to inactivate parasites such as Myco-
bacterium bovis BCG, Mycobacterium smegmatis, Salmonella
typhimurium,orLeishmania donovani due to a mutation in the
Bcg and/or Nramp gene (43). The C3H/HeJ mouse strain used
in our studies represents a Bcg-resistant (wild-type) strain (43).
In general, the clinical application of KC-depleting agents
such as GdCl
3
and Cl
2
MBP is limited due to their pharmaco-
logical side effects (8, 20, 35) and their potential damage of
organs other than liver resulting from viral spillover beyond the
KC (47). We noticed that the treatment with GdCl
3
, while
preventing liver damage and cytokine release, made the ani-
mals more susceptible to high viral doses. When more than 4 3
10
10
PFU of adenovirus per mouse (n 5 3 or more per group)
was injected, the mortality 24 h after infection was twofold
greater in GdCl
3
-treated mice than in untreated animals (not
shown).
The interactions of recombinant adenovirus with the host
liver are complex, and here we have started to uncover some of
these early events. Clearly, host responses to adenovirus can
vary greatly between animals with different genetic back-
grounds, and they will be more difficult to unravel in a genet-
ically diverse, human population. To reduce early toxicity in-
duced by adenoviruses, further improvements of adenovirus-
based gene delivery should focus on the use of vectors depleted
for viral gene expression; on the restoration of E3 genes that
protect cells from TNF-induced cytolysis; on the pursuit of
other routes for vector delivery, such as infusion into the bile
duct to reduce activation of KC; or on the transient repression
of serum TNF or NF-kB, for example, by TNF antibodies,
chimeric TNF receptors (25), or steroids.
ACKNOWLEDGMENTS
This work was supported by NIH-R01 grants DK49022 and
DK51807.
REFERENCES
1. Andus, T., J. Bauer, and W. Gerok. 1991. Effects of cytokines on the liver.
Hepatology 13:364–375.
2. Baeuerle, P., and D. Baltimore. 1996. NF-kB: ten years after. Cell 87:13–20.
3. Barr, D., J. Tubb, D. Ferguson, A. Scaria, A. Lieber, C. Wilson, J. Perkins,
and M. A. Kay. 1995. Strain related variations in adenovirally mediated
transgene xpression from mouse hepatocytes in vivo: comparisons between
immunocompetent and immunodeficient inbred strains. Gene Ther. 2:151–
155.
4. Bett, A. J., W. Haddara, L. Prevec, and F. L. Graham. 1994. An efficient and
flexible system for construction of adenovirus vectors with insertions or
deletions in early regions 1 and 3. Proc. Natl. Acad. Sci. USA 91:8802–8806.
5. Boucher, R. C. 1996. Current status of CF gene therapy. Trends Genet.
12:81–84.
6. Bruder, J., and I. Kovesti. 1997. Adenovirus infection stimulates the Raf/
MAPK signaling pathway and induces interleukin-6 expression. J. Virol.
71:398–404.
7. Busam, K. J., A. Homfeld, R. Zawatzky, S. Kastner, J. Bauer, W. Gerok, and
K. Decker. 1990. Virus-vs endotoxin-induced activation of liver macro-
phages. Eur. J. Biochem. 191:577–582.
8. Callery, M. P., T. Kamei, and W. Flye. 1990. Kupffer cell blockade increases
mortality during intra-abdominal sepsis despite improving systemic immu-
nity. Arch. Surg. 125:36–41.
9. Cornell, R. P. 1985. Restriction of gut-derived endotoxin impairs DNA
synthesis for liver regeneration. Am. J. Physiol. 249:R563–R569.
10. Cornell, R. P., B. L. Liljequist, and K. F. Bartizal. 1990. Depressed liver
regeneration after partial hepatectomy of germ-free, athymic and lipopo-
lysaccharide-resistant mice. Hepatology 11:916–922.
11. Decker, K. 1990. Biologically active products of stimulated liver macro-
phages (Kupffer cells). Eur. J. Biochem. 192:245–261.
12. Deryckere, F., and H.-G. Burger. 1996. Tumor necrosis factor a induces the
adenovirus early 3 promoter by activation of NF-kB. J. Biol. Chem. 47:
30249–30255.
13. Fitzgerald, M., E. Webber, J. Donovan, and N. Fausto. 1995. Rapid DNA
binding by nuclear factor kB in hepatocytes at the start of liver regeneration.
Cell Growth Differ. 6:417–427.
14. Ginsberg, H. S., L. L. Moldawer, P. B. Sehgal, M. Redington, P. L. Kilian,
R. M. Chanock, and G. A. Prince. 1991. A mouse model for investigating the
molecular pathogenesis of adenovirus pneumonia. Proc. Natl. Acad. Sci.
USA 88:1651–1655.
15. Ginsberg, H. S., U. Lundholm-Beauchamp, R. L. Horswood, B. Pernis, W. S.
Wold, R. M. Chanock, and G. A. Pronce. 1989. Role of the early region 3
(E3) in pathogenesis of adenovirus disease. Proc. Natl. Acad. Sci. USA
86:3823–3827.
16. Ginsberg, H. S. 1996. The ups and downs of adenovirus vectors. Bull. N. Y.
Acad. Med. 73:53–58.
17. Gooding, L. R. 1994. Regulation of TNF-mediated cell death and inflamma-
tion by human adenovirus. Infect. Agents Dis. 3:106–115.
18. Hardonk, M. J., F. W. J. Dihkhuis, C. E. Hulstaert, and J. Koudstaal. 1992.
Heterogeneity of rat liver and spleen macrophages in gadolinium chloride-
induced elimination and repopulation. J. Leukocyte Biol. 52:296–302.
19. Ilan, Y., G. Droguett, N. R. Chowdhury, Y. Li, K. Sengupta, N. R. Thummala,
A. Davidson, J. R. Chowdhuty, and M. S. Horwitz. 1997. Insertion of the
adenoviral E3 region into a recombinant viral vector prevents antiviral hu-
moral and cellular immune responses and permits long-term gene expres-
sion. Proc. Natl. Acad. Sci. USA 94:2587–2592.
20. Ishiyama, H., M. Sato, K. Matsumura, M. Sento, K. Ogino, and T. Hobara.
1995. Proliferation of hepatocytes and attenuation from carbon tetrachloride
hepatotoxicity by gadolinium chloride in rats. Pharmacol. Toxicol. 77:293–
298.
21. Kay, M. A., P. Baley, S. Rosenberg, F. Leland, L. Fleming, K. Parker-Ponder,
T. Liu, G. Darlington, W. Pokorny, and S. L. C. Woo. 1992. Expression of
human a1-antitrypsin in dogs after autologous transplantation of retroviral
transduced hepatocytes. Proc. Natl. Acad. Sci. USA 89:89–93.
22. Kay, M. A., Q. Li, F. Leland, C. Toman, M. Finegold, and S. L. C. Woo. 1992.
Hepatic gene therapy: persistent expression of human a1-antitrypsin in mice
after direct gene delivery in vivo. Hum. Gene Ther. 3:641–647.
23. Kay, M. A., F. Graham, F. Leland, and S. L. C. Woo. 1995. Therapeutic
serum concentrations of human a1-antitrypsin after adenoviral-mediated
gene transfer into mouse hepatocytes. Hepatology 21:815–819.
24. Kay, M. A., A.-X. Holterman, L. Meuse, A. Gown, H. Ochs, P. S. Linsley, and
C. B. Wilson. 1995. Long term hepatic adenovirus mediated gene expression
in mice following CTLA4Ig administration. Nat. Genet. 11:191–197.
25. Kolls, J., K. Peppel, M. Silva, and B. Beutler. 1994. Prolonged and effective
blockade of tumor necrosis factor activity through adenovirus-mediated gene
transfer. Proc. Natl. Acad. Sci. USA 91:215–219.
26. Kuzmin, A. I., M. J. Finegold, and R. C. Eisensmith. 1997. Macrophage
depletion increases the safety, efficacy and persistence of adenovirus-medi-
ated gene transfer in vivo. Gene Ther. 4:309–316.
27. Li, Q., M. A. Kay, M. Finegold, L. D. Stratford-Perricaudet, and S. L. C.
Woo. 1993. Assesment of recombinant adenoviral vectors for hepatic gene
therapy. Hum. Gene Ther. 4:403–409.
8806 LIEBER ET AL. J. VIROL.
28. Lieber, A., M.-J. Vrancken-Peeters, L. Meuse, N. Fausto, J. Perkins, and
M. A. Kay. 1995. Adenovirus mediated urokinase gene transfer induces liver
regeneration and allows for efficient retrovirus tranduction of hepatocytes in
vivo. Proc. Natl. Acad. Sci. USA 92:6210–6214.
29. Lieber, A., C.-Y. He, I. Kirillova, and M. A. Kay. 1996. Recombinant adeno-
viruses with large deletions generated by Cre-mediated excision exhibit dif-
ferent biological properties compared with first-generation vectors in vitro
and in vivo. J. Virol. 70:8944–8960.
30. Marianneau, P., A. Cardona, L. Edelman, V. Deubel, and P. Despres. 1997.
Dengue virus replication in human hepatoma cells activates NF-kB which in
turn induces apoptotic cell death. J. Virol. 71:3244–3249.
31. Noah, T. L., I. A. Wortman, P. Hu, M. W. Leigh, and R. C. Boucher. 1996.
Cytokine production by cultured human bronchial epithelial cells infected
with a replication-deficient adenoviral gene transfer vector or wild type
adenovirus type 5. Am. J. Cell. Mol. Biol. 14:417–424.
32. Pessara, U., and N. Koch. 1990. Tumor necrosis factor a regulates expression
of the major histocompatibility complex class II-associated invariant chain by
binding of an NF-kB-like factor to a promoter element. Mol. Cell. Biol.
10:4146–4154.
33. Pfizenmaier, K., H. Wajant, and M. Grell. 1996. Tumor necrosis factor in
1996. Cytokine Growth Factor Rev. 3:271–277.
34. Rai, R. M., S. Q. Yang, C. McClain, C. L. Karp, A. S. Klein, and A. M. Diehl.
1996. Kupffer cell depletion by gadolinium chloride enhances liver regener-
ation after partial hepatectomy in rats. Am. J. Physiol. 270:G909–G918.
35. Salkowski, C. A., R. Neta, T. A. Wynn, G. Strassmann, N. van Rooijen, and
S. N. Vogel. 1995. Effect of liposome-mediated macrophage depletion on
LPS-induced cytokine gene expression and radioprotection. J. Immunol.
155:3168–3179.
36. Sambucetti, L. C., J. M. Cherrington, G. W. G. Wilkinson, and E. Mocarski.
1989. NF-kB activation of the cytomegalovirus enhancer is mediated by a
viral transactivator and by T cell stimulation. EMBO J. 8:4251–4258.
36a.Schowalter, D., and M. Kay. Unpublished data.
37. Schowalter, D. B., L. Meuse, C. B. Wilson, P. S. Linsley, and M. A. Kay. 1997.
Constitutive expression of murine CTLA4Ig from recombinant adenovirus
vector results in prolonged transgene expression. Gene Ther. 4:853–860.
38. Shimizu, H., K. Mitomo, T. Watanabe, S. Okamoto, and K. Yamamoto. 1990.
Involvement of a NF-kB-like transcription factor in the activation of the
interleukin-6 gene by inflammatory cytokines. Mol. Cell. Biol. 10:561–568.
39. Shiratori, Y., S. Hongo, Y. Hikiba, K. Ohmura, T. Nagura, K. Okano, Y.
Komatsu, and M. Omata. 1996. Role of macrophages in regeneration of
liver. Dig. Dis. Sci. 41:1939–1946.
40. Sparer, T. E., R. A. Tripp, D. L. Dillehay, T. W. Hermiston, W. S. M. Wold,
and L. R. Gooding. 1996. The role of human adenovirus early region 3
proteins (gp19K, 10.4K, 14.5K, and 14.7K) in a murine pneumonia model.
J. Virol. 70:2431–2439.
41. Teodoro, J. G., and P. E. Branton. 1997. Regulation of apoptosis by viral
gene products. J. Virol. 71:1739–1746.
42. Tsui, L. V., L. G. Guidotti, T. Ishikawa, and F. V. Chrisari. 1995. Posttran-
scriptional clearance of hepatitis B virus RNA by cytotoxic T lymphocyte-
activated hepatocytes. Proc. Natl. Acad. Sci. USA 92:12398–12402.
43. Vidal, S. M., D. Malo, K. Vogan, E. Skamene, and P. Gros. 1993. Natural
resistance to infection with intracellular parasites: isolation of a candidate
for Bcg. Cell 73:469–495.
44. Vrancken Peeters, M.-J., A. Lieber, J. Perkins, and M. A. Kay. 1996. A
method for multiple portal vein infusions in mice: quantification of adeno-
virus-mediated hepatic gene transfer. BioTechniques 20:278–285.
45. Watson, J., K. Kelly, M. Largen, and B. A. Taylor. 1978. The genetic map-
ping of a defective LPS responsive gene on C3H/HeJ mice. J. Immunol.
120:422–424.
46. Wold, W. S. M. 1993. Adenovirus genes that modulate the sensitivity of virus
infected cells to lysis by TNF. J. Cell. Biochem. 53:329–335.
47. Wolff, G., S. Worgall, N. van Rooijen, W.-R. Song, B.-G. Harvey, and R. G.
Crystal. 1997. Enhancement of in vivo adenovirus-mediated gene transfer
and expression by prior depletion of tissue macrophages in the target organ.
J. Virol. 71:624–629.
48. Worgall, S., G. Wolff, E. Falk-Pedersen, and R. G. Crystal. 1997. Innate
immune mechanisms dominate elimination of adenoviral vectors following in
vivo administration. Hum. Gene Ther. 8:37–44.
49. Wrighton, C. J., R. Hofer-Warbinek, T. Moll, R. Eytner, F. H. Bach, and R.
de Martin. 1996. Inhibition of endothelial cell activation by adenovirus-
mediated expression of I-kBa, an inhibitor of the transcription factor NF-
kB. J. Exp. Med. 183:1013–1022.
50. Yamada, Y., I. Kirillova, J. J. Peschon, and N. Fausto. 1997. Initiation of
liver growth by tumor necrosis factor in mice lacking type I tumor necrosis
factor receptor. Proc. Natl. Acad. Sci. USA 94:1441–1446.
51. Yang, Y., F. A. Nunes, K. Berensci, E. E. Furth, E. Gonzol, and J. M. Wilson.
1994. Cellular immunity to viral antigens limits E1-deleted adenoviruses for
gene therapy. Proc. Natl. Acad. Sci. USA 91:4407–4411.
52. Yang, Y., G. Trinchieri, and J. M. Wislon. 1994. Inactivation of E2a in
recombinant adenoviruses improves the prospect for gene therapy in cystic
fibrosis. Nat. Genet. 7:362–369.
VOL. 71, 1997 INNATE IMMUNITY AFTER ADENOVIRUS VECTOR ADMINISTRATION 8807
