| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Institut National de la Santé et de la Recherche Médicale, Unité 437, Nantes, France;
Molecular Medicine and Gene Therapy Unit, University of Manchester School of Medicine, Manchester, United Kingdom; and
Department of Cell Biology and Immunology, School of Medicine, Free University, Amsterdam, The Netherlands
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, TNF-
, and Fas
ligand. Administration of AdmIL-4 to IFN-, TNF- receptor type I,
or TNF- receptor type II knockout mice also resulted in lethal
hepatitis, whereas a moderate protection was observed in Fas-deficient
lpr mice. IL-4-dependent hepatocyte apoptosis could be
abolished by treatment with caspase inhibitory peptides. Our results
thus demonstrate that IL-4 causes hepatocyte apoptosis, which is only
partially dependent on the activation of Apo-1-Fas signaling and is
largely independent of any immune cells in the
liver. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
IL-4 is a major cytokine released during T cell responses polarized toward a Th2 phenotype. IL-4 itself displays a variety of either anti- or proinflammatory actions, mainly on cells of various hemopoietic lineages (8). Its actions include antiapoptotic actions on B cells (12) and proapoptotic effects on macrophages (13), mast cells (14), or eosinophils (15). Although IL-4 has also been described as modulating functions of nonhemopoietic cells, such as endothelial cells, fibroblasts, and hepatocytes (8), much less is known about its actions on these cells compared with those on hemopoietic cells.
We recently showed that adenovirus-mediated IL-4 expression in the liver by intraportal administration of 2.5 x 109 PFU of adenovirus coding for rat IL-4 (AdrIL-4)4 results in nonlethal, reversible hepatitis in rats (16). At this dose of AdrIL-4 the hepatitis is partially due to an increase in cellular and humoral anti-adenovirus immune responses (17). In the present study, we show that higher doses of AdrIL-4 (1010 PFU) induce lethal hepatitis in rats, we reproduced this lethal hepatitis model in mice using an adenovirus coding for mouse IL-4 (AdmIL-4) and investigated the inflammatory and immune mechanisms as well as the role of apoptosis underlying acute and lethal IL-4-mediated hepatitis.
Our present results show that expression of IL-4 in the liver results
in extensive hepatocyte apoptosis and mortality, which are prevented by
treatment with the caspase inhibitory YVADcmk peptide and was
statistically significantly reduced in lpr mice. Lethal
hepatitis was associated with production of IFN-, TNF-, and Fas
ligand (FasL). In mice, IL-4-dependent lethal hepatitis was also not
reduced in recombinase-activating gene-2
(RAG2)-/-, IFN-, TNF- receptor type I
(TNF-RI), or TNF-RII knockout mice. Although the livers of rats
or mice expressing IL-4 were heavily infiltrated with
monocytes/macrophages, prevention of hepatic monocyte infiltration,
depletion of resident macrophages, and abolition of T cell immune
responses in rats failed to protect against hepatic lesions and death.
Hepatitis was independent of anti-adenovirus immune responses,
since systemic IL-4 production as a result of administration of AdrIL-4
into the hind limb of rats also resulted in hepatitis. Finally, rIL-4
or transduction with AdrIL-4 induced apoptosis of primary cultures of
rat hepatocytes, which was prevented by caspase inhibitors.
In summary, our results indicate that besides the already known
chemotactic and immune activator actions of IL-4, lethal hepatitis
induced by this cytokine is mediated through apoptosis of hepatocytes
by mechanisms independent of immune response activation,
proinflammatory effects on leukocytes, or TNF- or IFN-
production. Therefore, IL-4 secretion could play an important role in
various types of hepatitis by different mechanisms. Treatment with
IL-4-blocking agents or antiapoptotic caspase peptides could prevent
IL-4-mediated liver injury in these situations.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Male Wistar rats (8 wk old; CERJ, Le Genest St. Isle, France) and nude rats (Rowett Hooded rnu/rnu, 8 wk old; Charles River, St. Aubin les Elbeuf, France) were injected via the portal vein with the indicated numbers of PFUs of recombinant adenovirus as previously described (16). Gene transfer to the hind limb was performed by i.m. injection of 2 x 1010 PFU (in a total volume of 100 µl at four injection points).
C57BL/6 mice were purchased from Charles River, and
RAG2-/- and lpr mice
(Fas-/-) were obtained from Centre National de
la Recherche Scientifique-Centre de Developpement des Techniques
Arducées (Orléans, France). Mice deficient in IFN-,
TNF-RI, or TNF-RII were purchased from The Jackson Laboratory
(Bar Harbor, ME). All mutated mice were of the C57BL/6 background. Male
mice (6 wk old) received the indicated doses of recombinant adenovirus
by i.v. injection.
Recombinant adenovirus
AdrIL-4 was constructed, propagated, and titrated as previously
described (16). Briefly, AdrIL-4 was constructed by
homologous recombination in 293 cells using DNA from Addl324 (an
Ad5-derived, E1- and E3-deleted adenovirus). Rat IL-4 cDNA sequences
were placed under transcriptional control of the human elongation
factor-1 promoter. AdmIL-4 was provided by F. Graham (McMaster
University, Hamilton, Canada). AdlacZ has been previously described
(18). Adenoviruses were purified using two consecutive
CsCl gradient ultracentrifugations and then were aliquoted and stored
at -80°C until used. Adenovirus preparations were tittered on 293
cells using PFU assays. Analysis of adenoviral particle concentration,
as determined by A260 absorbancy, allowed us to
calculate the ratio of PFUs to total physical adenovirus particles
(PFU/particle ratio). The PFU/particle ratio varied from 1/20 to 1/40
for all of the recombinant adenovirus stocks used in this work.
Endotoxin levels in purified recombinant adenovirus preparations were
<0.1 endotoxin units/1010 PFU as detected
by the Limulus assay (Sigma, St. Louis, MO). AdrIL-4 and
AdmIL-4 preparations contained <10 PFU replication-competent
adenovirus/109 PFU recombinant adenovirus, as
determined using adenovirus replication-permissive A549 cells and
Addl7001 (replication-competent adenovirus) as a positive control
(provided by P. Moullier, Nantes, France).
Depletion of leukocyte subpopulations and immunosuppression
On day 0, animals were injected with recombinant adenoviruses.
To evaluate the role of CD8+ cells in vivo, rats
received i.p. injections of anti-CD8 -chain mAb (OX8 clone,
IgG1; European Cell Culture Collection, Wiltshire, U.K.) or an
irrelevant control, anti-human CD16 mAb (3G8 clone, IgG1; American
Type Culture Collection, Manassas, VA) on days -1, 2, 5, and 8.
CD8+ lymphocyte depletion in peripheral blood
(>98%) was confirmed using the OX8 mAb and anti-
TCR R.7.3
mAb between days 0 and 8 in both naive control rats as well as in the
experimental animals (up to day 2, since death by IL-4-mediated
hepatitis occurred between days 3 and 5). Cyclosporine A (CsA) was
administered orally at daily doses of 10 mg/kg from day -1 for the
duration of the experiment (CsA serum levels were confirmed at >960
µg/ml). Cyclophosphamide (CyP) was administered to rats and mice i.p.
at 70 mg/kg on day -5 and then at 35 mg/kg every 3 days. Peripheral
leukocytes in rats from day 0 onward were between 0.5 and 2 x
109/liter (normal values, 12–17 x
109/liter). Prolonged treatment with CyP (>10
days) was associated with some toxicity, since Addl324-infused animals
showed 20% mortality. CyP was not hepatotoxic, since this group showed
normal hepatic biochemical values and had no hepatic lesions (data not
shown). Depletion of Kupffer cells was obtained by i.v. injection of 1
ml of liposome-encapsulated dichloromethylene diphosphonate
(liposomes-Cl2MDP; a gift from Roche, Mannheim, Germany)
(17) on day -1 and confirmed on day 0 by immunostaining
using the ED2 anti-macrophage mAb (liver ED2+
cells from 8.3 ± 1 in nontreated controls to 0.4 ± 0.2% in
liposomes-Cl2MDP-treated ones).
Liposomes-Cl2MDP have been previously shown to deplete the
liver from ED2+-resident macrophages, such as
Kupffer cells, rather than ED1+ peripheral blood
monocytes or newly arrived tissue macrophages (19).
Treatment with anti-apoptotic YVADcmk peptide
The synthetic peptide (Ac-Tyr-Val-Ala-Asp-chloromethylketone
(YVADcmk), Bachem, Bubendorf, Switzerland) is an irreversible inhibitor
of the IL-1-converting enzyme (ICE)-like family of caspases,
previously used in vivo (20, 21, 22, 23). Rats infused
intraportally with AdrIL-4 (day 0) received i.p. injections of 2 mg of
YVADcmk daily from days 0 to 6.
Analysis of liver histopathology and liver function
Liver samples were fixed in 4% formaldehyde, embedded in paraffin, sectioned, and stained with hematoxylin-eosin-saffron (HES). Alanine and aspartate transaminases (ALAT and ASAT), bilirubin levels, and alkaline phosphatase activity were measured in serum as indicators of hepatocellular damage.
Immunohistochemistry
Immunohistochemistry on frozen sections was performed as
previously described in detail (17). Briefly, tissue
sections were incubated for 60 min with rabbit anti-FasL (N-20;
Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-rat TNF-
(Serotec, Oxford, U.K.), or the following mouse mAb: anti-leukocyte
CD45 (OX1 and OX30), anti-monocyte/macrophage (ED1),
anti-macrophage (ED2; a large proportion of tissue macrophages
coexpress the ED1 and ED2 Ags, but at early and late time points of
macrophage differentiation there is a preferential expression of ED1
and ED2 Ags, respectively) (22), anti- TCR
(R.7.3), anti-CD4 (W3/25), anti-CD8 -chain (OX8),
anti-monomorphic MHC class II Ags (OX6) (all of these mAb were
obtained from European Cell Culture Collection), and
anti-NKRP1 (3.2.3; Serotec). After washing the sections
exhaustively, they were incubated for 45 min with biotin-conjugated
anti-rabbit or anti-mouse Ig Abs (Vector Laboratories,
Burlingame, CA) and then with HRP-conjugated streptavidin (Vector
Laboratories; 45 min) and revealed using the VIP substrate (Vector
Laboratories). Quantification of labeled cellular infiltrates was
performed using the point-counting technique and expressed as the
percent area of biopsies occupied by positive cells as previously
described (17). The numbers of cells immunoreactive for
either TNF- or FasL were determined using a microscope grid (0.0625
mm2; x400 magnification).
Histological determination of hepatocyte apoptosis
Apoptosis in liver sections was investigated using three different methods.
First, hepatocyte apoptotic bodies were detected in HES-stained liver sections, and their numbers were determined using a microscope grid (x400 magnification).
Secondly, apoptosis was also analyzed using the TUNEL technique. Paraffin sections were deparaffinized, washed twice in PBS, digested with 20 µg/ml proteinase K (Sigma; 15 min at room temperature), and incubated with hydrogen peroxide (10%) in methanol (1/5, v/v, 30 min). The TUNEL ApopTag kit (Oncor, Gaithersburg, MD; for detection by immunoperoxidase using diaminobenzidene as the substrate) or the Apodetect kit (Roche, Meylan, France; for detection by immunofluorescence using FITC-labeled Abs) were used according to the manufacturers’ instructions.
Finally, nuclear condensation and fragmentation typical of apoptotic cells were assessed morphologically by qualified liver histopathologists using Hoechst 33258 staining. Liver cryosections or hepatocyte cytospins were incubated (10 min, room temperature) in an acetic acid/methanol (1/3, v/v) solution, washed twice in HBSS without red phenol (Life Technologies, Gaithersburg, MD), and incubated (for 30 min, at room temperature, protected from light) with 0.1 µg/ml of Hoechst 33258 (diluted in HBSS). DNA staining was visualized using an epifluorescence microscope with a 365-nm filter, and the percentage of apoptotic cells was determined by counting at least 500 cells from each experimental condition.
Quantitative competitive RT-PCR
RNA isolation, RT, PCR conditions, oligonucleotides, and standard DNA constructs used in this study have been described previously (24, 25).
The quantitative RT-PCR technique used allows precise quantification of transcripts using PCR amplification run to saturation. The basis of this technique is to coamplify the cDNA to be quantified with known amounts of an internal standard DNA whose sequence is identical with the sequence to be assayed except for a four-base deletion. Quantification is performed after an additional nested PCR using a fluorescent dye-labeled oligonucleotide. This PCR method allows the two amplified species (i.e., the experimental cDNA sample and the standard DNA) to be visualized and quantified on an electrophoresis gel using a DNA sequencer (Applied Biosystems, Courtaboeuf, France) equipped with Immunoscope software (Pasteur Institute, Paris, France).
The number of copies of the housekeeping gene hypoxanthine phosphoribosyltransferase (HPRT) was quantified using the same technique, and results were expressed as the ratio between the number of copies of the mRNA analyzed and the number of copies of HPRT in the same sample.
Hepatocyte culture and adenovirus-mediated in vitro transduction
Rat hepatocytes were isolated by collagenase digestion as
described previously (26) and were provided by C.
Guguen-Guillouzo (Institut National de la Santé et de la
Recherche Médicale, Unité 49, Rennes, France). Hepatocytes
(>98% pure) were transduced with recombinant adenovirus at the
indicated number of PFU per cell and were cultured in Williams medium
supplemented with 1 g/L albumin, 5 µg/ml insulin, 7 x
10-5 M dexamethasone, 1 mM glutamine, 100 U/ml
penicillin, and 1 mg/ml streptomycin (all of these reagents were from
Sigma) for 72 h at 37°C in 5% CO2.
Hepatocytes were also cultured under the same conditions with rat rIL-4
(Biosource, Rungis, France) or human rTNF- (gift from Müller
Neumann, BASF/Knoll, Ludwigshafen, Germany). Incubations with
human rTNF- were performed in the presence of cycloheximide (5
µg/ml). The caspase inhibitors YVADcmk, z-VADfmk (both from
Bachem), and z-DEVDfmk (Calbiochem, San Diego, CA) and the cathepsin
inhibitor I z-FAfmk (Calbiochem), were dissolved at 10 mM in DMSO and
diluted freshly in culture medium (50 µM final dilution). Hepatocytes
were cultured with these peptides for 20 min at 37°C before addition
of AdrIL-4. Hepatocyte culture supernatants were harvested and
centrifuged, and hepatocyte-derived ALAT was measured in cell-free
medium. Nonadherent hepatocytes obtained following this centrifugation
were pooled with adherent cells and used to obtain cytospins for the
quantification of apoptotic cells using Hoechst 33258 histochemical
stain.
Statistics
Statistical significance was evaluated using a one-way ANOVA test. A p < 0.05 was considered to be statistically significant. The survival of different groups of animals was compared using the Kaplan-Meier test.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Intraportal infusion of AdrIL-4 in rats resulted in lethal,
dose-dependent hepatitis (Fig. 1
).
Animals injected with 109, 2.5 x
109, 5 x 109, or
1010 PFU of AdrIL-4 showed 0, 6.6, 70, or 85%
mortality, respectively (Fig. 1). The time of death after
administration of 1010 PFU of AdrIL-4 was 4
± 1 days after gene transfer and was progressively delayed (up to day
14) with decreasing doses of AdrIL-4. Rats infused with
1010 PFU of Addl324 and with no further treatment
showed no mortality. Animals injected with 109,
2.5 x 109, 5 x
109, and 1010 PFU of
AdrIL-4 showed a dose-dependent increase (mean ± SEM;
n = 6–11) in serum bilirubin (5 ± 1.5, 27
± 10, 32 ± 3.8, and 51 ± 3.5 µmol/L, respectively) and
ALAT (42 ± 4, 68 ± 9, 81 ± 14, and 237 ± 22 IU,
respectively). Rats infused with 1010 PFU of
Addl324 showed values of bilirubin (4 ± 0.2 µmol/L) and ALAT
(34 ± 4.5 IU) in serum comparable to those in nontreated rats
(bilirubin, 5 ± 1 µmol/l; ALAT, 33 ± 3.5 IU).
|
, A and B). By
contrast, livers treated with AdrIL-4 revealed confluent hepatocyte
lesions predominantly in periportal areas associated with a leukocyte
infiltrate composed of mononuclear cells (Fig. 2C). This
leukocyte infiltrate also extended to centrilobular areas.
AdrIL-4-infused livers contained a large number of hepatocytes showing
the typical features of apoptotic cells: cell shrinkage, intensely
acidophilic cytoplasm with condensed and fragmented chromatin, or even
an absence of nuclei.
|
D). By contrast, after administration of
AdrIL-4 large numbers of apoptotic TUNEL-positive cells were detected
(Fig. 2E). Most of these cells were hepatocytes, while a few
were leukocytes. TUNEL signals were detected in apoptotic bodies and
also in cells with condensed chromatin in the periphery of their
nuclei. Apoptosis was confirmed by two-color immunofluorescence in
liver sections showing colocalization of condensed chromatin (Fig. 2F) and DNA fragmentation using the TUNEL assay (Fig. 2G). Animals injected with AdrIL-4 showed no overt tissue histopathologic lesions in heart, lung, kidney, or brain, whereas thymus and spleen showed areas of atrophy (thymic cortex and spleen white pulp) or necrosis (thymic medulla and spleen red pulp; data not shown).
TNF-, FasL, and IFN- expression is increased in livers
infused with AdrIL-4
The expression of TNF-, FasL, and IFN- has been implicated
in the induction of apoptosis of hepatocytes in different liver
pathological processes (25, 27, 28, 29). Immunohistology using
Abs directed against TNF- or FasL showed no reactivity in livers
infused with Addl324 (Fig. 3, A and C). In contrast, livers treated with
AdrIL-4 contained numerous positive cells recognized by
anti-TNF- and FasL Abs (Fig. 3, B and D,
respectively). Positive cells were mainly situated in areas infiltrated
by leukocytes, but TNF- reactivity showed a more diffuse pattern,
with weak reactions in certain hepatocyte areas. Morphometric
quantification of anti-TNF-- and FasL-labeled cells showed a
significant increase in livers infused with AdrIL-4 compared with
untreated or Addl324-treated livers (Fig. 3E).
|
and FasL mRNA levels using a quantitative RT-PCR
method showed significantly increased accumulation of specific mRNAs in
livers infused with AdrIL-4 compared with untreated or Addl324-treated
livers (Fig. 3F). Production of IFN- following
administration of low doses of AdrIL-4 was previously described
(16), and high levels of IFN--positive cells were also
observed by immunohistology and mRNA after gene transfer with
1010 PFU in the present study (data not shown).
These results show that levels of liver TNF-, FasL, and IFN- and
their corresponding mRNAs are increased after adenovirus-mediated IL-4
expression and suggest that they could mediate IL-4-dependent
hepatocyte apoptosis. Lethal hepatitis induced by adenovirus-derived IL-4 is largely independent of T cell-mediated anti-adenovirus immune responses
To study the role of CTLs in mortality induced by Ad-derived IL-4
we pretreated rats with a depleting anti-CD8 mAb before injecting
them with AdrIL-4. AdrIL-4-induced lethal hepatitis was not
significantly reduced (57%, Fig. 1) in animals depleted of
CD8+ T cells and subsequently infused with
1010 PFU. Liver histology in animals infused with
1010 PFU of AdrIL-4 in the absence of
CD8+ cells showed hepatitis with the same
features (apoptotic hepatocytes, mixed leukocyte infiltration, and
reticulofibrosis) as that in animals with intact
CD8+ cells, although increases in serum markers
were of lower magnitude (data not shown). These results indicate that
depletion of CD8+ cells had a modest protective
effect against AdrIL-4-induced lethal hepatitis. To further investigate
the role of T cells in AdrIL-4-induced hepatitis, we injected AdrIL-4
in T cell-deficient nude rats. This resulted in a high mortality (67%;
Fig. 1), an intense leukocyte infiltrate composed of
ED1+/ ED2+
monocytes/macrophages (Table I), and
hepatic histopathology displaying apoptotic hepatocytes, leukocyte
infiltration, and reticulofibrosis (Fig. 4). Hepatic biochemical values in
CD8-depleted and nude rats were significantly raised (in both
nonsurvivors and survivors; data not shown) compared with those in
Addl324-infused or untreated rats.
|
|
), massive leukocyte infiltration, or increased
serum hepatic biochemistry values (data not shown). Injection of
Addl324 in CsA-treated or nude animals was perfectly tolerated (0%
mortality; Fig. 1). Altogether, these results indicate that lethal
hepatitis induced by high doses of AdrIL-4 is largely independent of T
cells. IL-4-dependent lethal hepatitis is independent of Kupffer cells or infiltrating leukocytes
Since Kupffer cells comprise the predominant type of leukocyte present in normal liver and IL-4 induced a massive increase in infiltrating monocytes/macrophages, we asked whether depletion of Kupffer cells or prevention of liver infiltration by monocytes would protect against the development of hepatitis.
Animals depleted of Kupffer cells by injection of liposomes containing
Cl2MDP prior to infusion of AdrIL-4 showed no protection
against mortality (Fig. 1) compared with animals treated with AdrIL-4
alone despite greatly reduced macrophage counts at the moment of death
(9 ± 1.1 vs 36.3 ± 2%, respectively; Table I).
Nevertheless, numbers of MHC class II and ED1+
cells, probably newly arrived monocytes, were significantly increased
in animals depleted of Kupffer cells and injected with AdrIL-4
(ED1+, 24.6 ± 2 vs 3.7 ± 0.7% in
controls; Table I). Hepatic histology showed the typical morphology of
IL4-induced acute hepatitis, e.g., ballooned hepatocytes, apoptotic
cells, and leukocyte infiltration (Fig. 4).
To further completely prevent infiltration of the liver by circulating
blood monocytes, leukocytes were depleted by a partially myeloablative
treatment with CyP before the infusion of AdrIL-4. Animals treated with
CyP and infused with AdrIL-4 were not protected against acute lethal
hepatitis (Fig. 1). CyP effectively prevented infiltration of the liver
by monocytes (ED1+ cells, 8.5 ± 2.3 vs
32 ± 1% in AdrIL-4-treated animals), macrophages
(ED2+ cells), and T cells (Table I). Liver
lesions included apoptotic hepatocytes, reticulofibrosis, and abnormal
liver architecture (Fig. 4). Liver enzymes showed raised hepatic
biochemistry values (data not shown).
Despite the prevention of liver infiltration by leukocytes, CyP
treatment did not deplete all Kupffer cells from the liver
(ED2+ cells, 6.6 ± 0.8 vs 8.3 ± 1 in
controls; Table I). To simultaneously eliminate both Kupffer cells and
infiltrating monocytes, animals were treated with
liposomes-Cl2MDP and CyP before being infused with AdrIL-4.
Immunohistology of leukocytes now demonstrated an efficient depletion
of resident Kupffer cells by >98% (ED2+ cells,
1.5 ± 0.1%), and the simultaneous inhibition by >99% of
monocytic infiltration (ED1+ cells, 0.9 ±
0.5%; Table I). However, such drastic elimination of monocytes and
macrophages was unable to protect animals from AdrIL4-dependent lethal
acute hepatitis (Fig. 1), with blood biochemistry values being grossly
abnormal (data not shown) and apoptotic bodies detected throughout the
liver (Fig. 4). These results indicate that lethal hepatitis induced by
IL-4 was totally independent of Kupffer cells or infiltrating
monocytes.
Induction of lethal hepatitis in mice following administration of AdmIL-4 and analysis of the molecular mechanisms using mutated or knockout mice
Injection of increasing doses of AdmIL-4 resulted in a
dose-dependent mortality, with 100% of DBA/2 and C57BL/6 mice dying
after administration of 4 x 109 or 8
x 109 PFU and 0% or 17% after administration
of 109 or 2 x 109 PFU, respectively (Table II). In contrast, administration of
4 x 109 PFU to BALB/c, 129/Sv, and SJL/J
resulted in 0, 0, and 20% mortality, respectively (Table II).
Administration to DBA/2 and C57BL/6 mice of up to 8 x
109 PFU of the noncoding Addl324 adenovirus was
nonlethal. Livers from DBA/2 and C57BL/6 mice injected with 4 x
109 PFU of AdmIL-4 showed a large proportion of
apoptotic hepatocytes as well as an intense leukocyte infiltrate mainly
composed of F4/80-positive macrophages (data not shown). Values of
bilirubin, ALAT, and ASAT were significantly elevated 6–9 days after
injection of 4 x 109 PFU of AdmIL-4 (12.5
µmol/L, 96 IU, and 324 IU, respectively) compared with those in
animals injected with Addl324 (3.7 µmol/L, 12 IU, and 72 IU) or not
treated (3.1 µmol/L, 12 IU, and 78 IU; p < 0.05 for
bilirubin, ALAT, and ASAT).
|
). Mice lacking IFN-, TNF-RI, or TNF-RII
showed kinetics and percentages of dead animals identical with those of
control mice (Table II). Fas-deficient mice (lpr)
showed a moderate, but statistically significant, prolongation of
survival compared with controls (10 ± 3 vs 7.5 ± 1.5,
respectively; p = 0.004; Table II).
These results indicate that the induction of lethal hepatitis by IL-4
is not restricted to rats, but can also be induced in mice. In both
species, the histopathologic characteristics and liver apoptosis are
very similar. As in rats, IL-4-induced hepatitis was not prevented by
elimination of T cell-dependent immune responses or liver infiltration
by peripheral leukocytes. Furthermore, IL-4-induced hepatitis was
independent of IFN-, TNF-RI, or TNF-RII expression and was
only partially dependent on the presence of an active Fas pathway.
Systemic high level expression of IL-4 from AdrIL-4 induces acute lethal hepatitis
The development of hepatic lesions independently of immune
responses or leukocyte infiltration strongly suggested that IL-4 could
be directly responsible for hepatocyte lesions in the absence of direct
hepatocyte transduction by adenovirus. To explore this possibility, we
injected AdrIL-4 in the hind limb of rats to obtain high levels of
circulating IL-4 (n = 2). Rats injected with AdlacZ as
controls (n = 2) showed undetectable levels of IL-4 in
serum (detection limit, 10 pg/ml), large numbers of X-galactosidase
(X-Gal)-positive cells in the injected muscle, but virtual absence of
X-Gal-positive cells in the liver (four X-Gal positive cells in 10
liver sections from two different hepatic lobes at x100 magnification)
and normal liver histology (Fig. 5A). Rats injected with
AdrIL-4 in the hind limb showed circulating IL-4 (49 and 36 pg/ml) for
at least 3 days following gene transfer. These values were in the lower
range of values observed in animals injected with AdrIL-4 into the
portal vein (43–190 pg/ml). Hepatic histology on day 7 after injection
showed apoptotic hepatocytes (Fig. 5B). Bilirubin and ALAT
levels were increased on day 3 and increased further on day 7 after
gene transfer in the hind limb in the two animals receiving AdrIL-4 (54
and 145 µmol/L of bilirubin; 96 and 109.2 IU of ALAT) compared with
those in animals receiving AdlacZ (3 and 4 µmol/L of bilirubin; 59.4
and 58.2 IU of ALAT). These results indicate that sustained levels of
circulating IL-4 do induce direct hepatotoxicity, which is completely
independent of hepatocyte transduction by adenoviruses.
|
To further identify the mechanisms underlying hepatocyte apoptosis
observed in vivo, we analyzed whether apoptosis could be induced in
purified hepatocytes by exposure to IL-4. Rat hepatocytes cultured for
48–72 h in the presence of rIL-4, AdrIL-4, or TNF- showed higher
numbers of apoptotic hepatocytes compared with either untreated cells
or hepatocytes transduced with AdlacZ at the same multiplicity of
infection (moi; Fig. 6A).
Transduction of hepatocytes with 2.5, 25, or 100 PFU of AdrIL-4 or
AdlacZ resulted in 5, 50, and 90% of transduced cells as evaluated by
X-Gal staining of AdlacZ-transduced cells. Hepatocytes incubated with
heat-inactivated AdrIL-4 remained viable (data not shown). Cultures in
the presence of rat rIL-4 also resulted in increased numbers of
apoptotic cells, although lower numbers than those obtained with
saturating concentrations of TNF- (Fig. 6A). Control
hepatocytes incubated with Hoechst 33258 showed uniform chromatin
staining (Fig. 6B), whereas hepatocytes transduced with
AdrIL-4 or cultured with recombinant rat IL-4 showed chromatin
condensation and fragmentation typical of apoptosis (Fig. 6C).
|
Prevention of AdrIL-4-induced hepatitis and in vitro hepatocyte apoptosis by treatment with caspase inhibitors
Caspases play a central role in the executory phase of apoptosis
(20). Inhibition of ICE-like caspases by the decoy peptide
YVADcmk results in protection against Fas and TNF--mediated
apoptosis of hepatocytes in vivo (21, 22, 23). To confirm the
role of apoptosis in AdrIL-4 induced hepatitis and to explore potential
therapeutic strategies, rats infused with AdrIL-4 were treated with
YVADcmk. All animals treated with YVADcmk and infused with 5 x
109 PFU of AdrIL-4 had survived by day 14 after
gene transfer, whereas animals receiving AdrIL-4 alone showed 70%
mortality (Fig. 7A).
Bilirubin, but not ALAT, levels in rats receiving YVADcmk and AdrIL-4
were significantly lower (11.5 ± 4.9 µmol/L and 72.5 ± 12
IU; n = 6; p < 0.006 and
p > 0.05, respectively) than those in animals
receiving AdrIL-4 alone (32 ± 3.8 µmol/L and 81 ± 14 IU;
n = 6). Compared with those in Addl324-treated animals
or untreated controls, hepatic biochemistry values in rats receiving
YVADcmk and AdrIL-4 were slightly higher without reaching statistical
significance. Hepatic histology in animals treated with YVADcmk
revealed a large reduction in the number of apoptotic bodies and
leukocyte infiltration comparable to that in livers receiving AdrIL-4
alone (Fig. 7, B and C). Morphometric
quantification of apoptotic bodies revealed significantly lower numbers
in animals treated with AdrIL-4 and YVADcmk than in animals receiving
AdrIL-4 alone (Fig. 7D). This protection against
AdrIL-4-mediated lethal hepatitis conveyed by YVADcmk demonstrates the
key biological relevance of IL-4 induction of hepatocyte apoptosis in
this new model of hepatitis.
|
show that hepatocyte
death induced by expression of IL-4 was prevented by preincubation with
YVADcmk (60% inhibition), z-DEVDfmk, or z-VADfmk (both 
90%
inhibition), whereas z-FAfmk did not protect hepatocytes.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-, TNF-RI-, TNF-RII-, and
Fas-deficient (lpr) mice. Death was only delayed by a
few days in lpr mice, indicating that Fas-FasL interactions
play only a partial role in IL-4-induced apoptosis.
Production of type 1 cytokines (mainly IFN- and IL-2) is associated
with hepatic inflammatory damage in humans (30).
Production of IL-4, a typical type 2 cytokine, is also elevated in
various human liver diseases, such as autoimmune hepatitis
(5), chronic hepatitis (5, 30, 31), and
primary biliary cirrhosis (6). Furthermore, administration
of IL-4 in primate preclinical studies (4) or to humans
for cancer treatment (1, 2, 3) was associated with a high
incidence of hepatic damage in all reports. Finally, IL-4 production by
liver NK1+ cells is essential in the induction of
Con A-induced hepatitis (11). These previous results
together with the present demonstration that IL-4 induces hepatocyte
apoptosis strongly suggest that IL-4 plays a role in the
pathophysiology underlying human and experimental liver disease.
We demonstrate that although liver expression of IL-4 results in
recruitment and activation of macrophages, eosinophils, and mastocytes,
leading to tissue inflammation, the lethal hepatitis induced by high
levels of IL-4 is mainly due to a proapoptotic effect of IL-4 on
hepatocytes. We have previously shown that in animals that received
2.5 x 109 PFU of AdrIL-4,
anti-adenovirus immune responses were increased, and depletion
of CD8+ cells reduced the infiltration of the
liver by leukocytes (17). The high rates of mortality
after infusion of higher doses of AdrIL-4 (1010
PFU) in CD8-depleted, nude, and CsA- or CyP-treated rats clearly
indicate that anti-adenovirus immune responses do not play an
essential role in the lethal effects induced by AdrIL-4. Nevertheless,
CD8-depleted rats showed a partial protection against lethality, and
this may depend on the depletion of leukocytes other than T cells
expressing CD8, such as NK and NK1+ T cells.
Therefore, although part of the hepatic lesions induced at lower and
higher doses of AdrIL-4 are dependent on the recruitment of leukocytes
and the increase in anti-adenovirus immune responses, at higher
doses of AdrIL-4, leukocyte-independent pathological mechanisms
predominate. These results also indicate that lethal hepatitis only
develops when a relatively high production of IL-4 is obtained, which
should be taken into consideration when the role of IL-4 is evaluated
in other hepatitis models. Finally, the doses of AdrIL-4 and AdmIL-4
needed to obtain lethal hepatitis are comparable to those used in most
gene transfer protocols in the liver and correspond to an moi of 10
PFU/hepatocyte in the rat model (32, 33).
The mouse IL-4R -chain (IL-4R) shows a polymorphism among
different mice strains that results in two groups of strains that
express receptors with either higher or lower affinity for IL-4
(34). Administration of AdmIL-4 resulted in lethal
hepatitis in strains with higher affinity (C57BL/6 and DBA/2), but not
in those with lower affinity (BALB/C, SJL/J; no data are available for
the IL-4R affinity in 129/Sv mice). The variability in strain
susceptibility to the IL-4-induced hepatitis may explain why in
previous studies systemic administration of high doses of rIL-4 or high
IL-4-expressing transgenic animals showed adverse effects when the
genetic background corresponded to the group with higher affinity IL-4R
(35), but not when the mice were from the lower affinity
group (36, 37). Furthermore, since polymorphisms in the
IL-4R-chain in humans also exist (38), IL-4 may play a
different role in hepatocellular damage in individuals of different
genetic backgrounds.
The fact that RAG2-/- immunodeficient and mice depleted of leukocytes by CyP treatment were not protected against lethal hepatitis induced by AdmIL-4 (but, in fact, mortality was slightly accelerated) further supports the idea that this hepatitis is independent of immune cells and that the pathophysiology of IL-4-induced hepatitis in rats and mice is identical.
Finally, the development of hepatic lesions following injection of AdrIL-4 in the hind limb and the induction of high levels of circulating IL-4 clearly indicate that hepatitis is independent of the presence of adenovirus particles in the liver undergoing hepatitis. Furthermore, the fact that IL-4 mediates apoptosis of purified hepatocytes in vitro further strengthens the evidence that apoptosis of hepatocytes in vivo is independent of inflammatory and immune responses, but is mediated by direct actions of IL-4 on hepatocytes.
IL-4R is associated with either IL-2R in hemopoietic cells or
IL-13R in endothelial cells (39, 40). Although
hepatocytes express receptors for IL-4 (7), the type of
second chain expressed in this cell type has not been reported. The
IL-4R intracellular segment contains distinct domains responsible
for cell proliferation, regulation of apoptosis, and gene expression
(39). Since the IL-4R does not contain death domains
present in receptors capable of directly transducing apoptotic signals
(such as Fas, TNF--RI, death receptor 3, and TNF-related
apoptosis-inducing ligand-R3), IL-4 may induce hepatocytes to
produce proapoptotic molecules. These molecules could be TNF-, Fas,
FasL, and reactive oxygen species acting in an autocrine manner, as has
been shown in several hepatic pathological conditions (25, 27, 28, 41). Alternatively, the IL-4R may directly induce hepatocyte
apoptosis, as has been described for members of the TNF receptor
superfamily (TNF-RII, CD40, CD30, and CD27) without death domains.
Nevertheless, apoptosis mediated by TNF-RII depends on intracellular
recruitment of adapter proteins (Fas-associated death domain,
TNFR-associated factor-2, and receptor-interacting protein)
(42), and these protein interactions are unlikely to be
possible with receptors other than those of the TNF receptor
superfamily. Despite the association of increased liver synthesis of
TNF- and IFN- during IL-4-mediated apoptosis of hepatocytes, the
fact that TNF-RI-, TNF-RII-, or IFN--deficient mice developed
hepatitis indicates that these cytokines do not mediate IL-4-induced
hepatocyte cell death. Hepatocytes are more sensitive to certain types
of apoptosis, probably due to the absence of Bcl-2 expression
(43). Hepatocyte apoptosis may depend on the production of
other proapoptotic molecules, such as TNF-related apoptosis-inducing
ligand, death receptor 3, or reactive oxygen species, and the
roles of these molecules need to be evaluated. In this regard, livers
from animals expressing IL-4 show reactive oxygen-mediated lesions, and
rats and mice treated with anti-oxidants such as melatonin or
cobalt protoporphyrin are protected from IL-4-induced lethal hepatitis
(I. Anegon, manuscript in preparation). The detailed molecular
mechanisms underlying the hepatocellular damage induced by IL-4 remain
to be determined.
Treatment with YVADcmk, a preferential, but not exclusive, inhibitor of
caspase-1 (ICE) (44), improved hepatic lesions and
survival after liver expression of IL-4. The protective effect of
YVADcmk in animals expressing IL-4 could be due to several nonexclusive
mechanisms. First, since caspase-1 cleaves and activates IL-1
(20) as well as IL-18 (45), a potent inducer
of IFN- production, the protection conferred by YVADcmk from
IL-4-induced hepatitis points to possible mechanisms involving IL-1
and/or IL-18. However, IFN- is not involved in IL-4-induced lethal
hepatitis, since hepatitis is unaffected in IFN--knockout mice. It
should be noted that IL-18 participates in other models of hepatitis,
such as that induced by endotoxin, which depend on mutual interactions
and a positive loop consisting of IL-18, the Fas-FasL system, and
IFN- and TNF- production (46). In this regard,
lpr mice show some protection against IL-4-induced
hepatitis. Second, it has been demonstrated that caspase-1 processes
and activates precaspase 3 (47) and that inhibition of
caspase-1 could inhibit certain steps of the caspase cascade leading to
apoptosis (20). Therefore, caspase-1 inhibition may
indirectly inhibit caspase-3 activation. The key role of caspase 1 in
apoptosis of hepatocytes is shown by the resistance of
caspase-1-deficient mice to Con A-induced hepatitis (48).
Finally, caspase 1 may not be the only caspase involved in IL-4-induced
hepatocyte apoptosis. The in vitro results showing higher protection of
hepatocytes from apoptosis induced by IL-4 with z-DEVDfmk (preferential
inhibitor for the effector caspase-3) and z-VADfmk (inhibitor for all
caspases) compared with YVADcmk, indicate that other proapoptotic
mechanisms activating other caspases, such as oxidative stress
(49), play a role. It should be stressed that YVADcmk has
been previously shown to protect hepatocytes from apoptosis in other
models of hepatitis (21, 22, 23) as well as neurons from
oxidative stress (50), in which IL-1 and IL-18 are not
involved.
IL-4 has been shown to regulate apoptosis in different cell types, being proapoptotic in endothelial cells (51), in mast cells from cord blood (14), and in eosinophils (15) and enhancing apoptosis in stimulated monocytes (13), while being antiapoptotic in B cells (12) and myeloid cells (8). The effects of IL-4 on the regulation of hepatocyte function (8, 9, 10) have been previously described in the literature, but the capacity of IL-4 to induce hepatocyte apoptosis has not been previously described.
Our results suggest that IL-4-mediated hepatitis will prove a valuable tool to evaluate new treatments for hepatic disorders in which IL-4 production has been reported or is suspected. In addition, this new model of hepatitis makes it possible to investigate the mechanisms by which cytokines, endothelial cells, monocyte/macrophages, T cells, and hepatocytes interact to generate liver inflammation and apoptosis.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 C.G. and H.C. contributed equally to this work.
3 Address correspondence and reprint requests to Dr. Ignacio Anegon, Institut National de la Santé et de la Recherche Médicale, Unité 437, 30 Boulevard Jean Monnet, 44093 Nantes, France.
4 Abbreviations used in this paper: AdrIL-4, adenovirus coding for rat IL-4; AdmIL-4, adenovirus coding for mouse IL-4; FasL, Fas ligand; RAG, recombinase-activating gene; TNF-RI, TNF- receptor type I; ALAT, alanine transaminase; ASAT, aspartate transaminase; CL2-MDP, liposome-encapsulated dichloromethylene diphosphonate; CsA, cyclosporine A; CyP, cyclophosphamide; HES, hematoxylin-eosin-saffron; HPRT, hypoxanthine phosphoribosyltransferase; ICE, IL-1-converting enzyme; moi, multiplicity of infection; X-Gal, X-galactosidase.
Received for publication August 22, 2000. Accepted for publication February 9, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Cancer Res. 51:2803. expression and apoptosis during allograft rejection and tolerance. Transplantation 66:887.[Medline]
mediate the lipopolysaccharide-induced acute-phase response and nitric oxide release by cultured rat hepatocytes. Eur. J. Biochem. 229:349.[Medline]
-induced cytotoxicity in mouse hepatocytes. Hepatology 21:1585.[Medline]
(IFN-) mRNA using liver biopsies at different phases of acute exacerbation of chronic hepatitis B. Clin. Exp. Immunol. 100:446.[Medline]
upregulates IgE synthesis and associates with atopic asthma. Nat. Genet. 19:119.[Medline]
chain. Blood 87:4286.-inducing factor and regulates LPS-induced IFN- production. Nature 386:619.[Medline]
, a mammalian homolog of CED-3, is a CrmA-inhibitable protease that cleaves the death substrate poly(ADP-ribose) polymerase. Cell 81:801.[Medline]
and FasL in concanavalin-A-induced hepatitis. J. Immunol. 160:4082.This article has been cited by other articles:
|
|
 |
|
  B. Brounais, C. Chipoy, K. Mori, C. Charrier, S. Battaglia, P. Pilet, C. D. Richards, D. Heymann, F. Redini, and F. Blanchard Oncostatin M Induces Bone Loss and Sensitizes Rat Osteosarcoma to the Antitumor Effect of Midostaurin In vivo Clin. Cancer Res., September 1, 2008; 14(17): 5400 - 5409. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Aoudjehane, P. Podevin, O. Scatton, P. Jaffray, I. Dusanter-Fourt, G. Feldmann, P.-P. Massault, L. Grira, A. Bringuier, B. Dousset, et al. Interleukin-4 induces human hepatocyte apoptosis through a Fas-independent pathway FASEB J, May 1, 2007; 21(7): 1433 - 1444. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. N. Ajuebor, J. A. Carey, and M. G. Swain CCR5 in T Cell-Mediated Liver Diseases: What's Going On? J. Immunol., August 15, 2006; 177(4): 2039 - 2045. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ali, G. D. King, J. F. Curtin, M. Candolfi, W. Xiong, C. Liu, M. Puntel, Q. Cheng, J. Prieto, A. Ribas, et al. Combined Immunostimulation and Conditional Cytotoxic Gene Therapy Provide Long-term Survival in a Large Glioma Model Cancer Res., August 15, 2005; 65(16): 7194 - 7204. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. N. Ajuebor, A. I. Aspinall, F. Zhou, T. Le, Y. Yang, S. J. Urbanski, S. Sidobre, M. Kronenberg, C. M. Hogaboam, and M. G. Swain Lack of Chemokine Receptor CCR5 Promotes Murine Fulminant Liver Failure by Preventing the Apoptosis of Activated CD1d-Restricted NKT Cells J. Immunol., June 15, 2005; 174(12): 8027 - 8037. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Guillonneau, C. Louvet, K. Renaudin, J.-M. Heslan, M. Heslan, L. Tesson, C. Vignes, C. Guillot, Y. Choi, L. A. Turka, et al. The Role of TNF-Related Activation-Induced Cytokine-Receptor Activating NF-{kappa}B Interaction in Acute Allograft Rejection and CD40L-Independent Chronic Allograft Rejection J. Immunol., February 1, 2004; 172(3): 1619 - 1629. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Jaruga, F. Hong, R. Sun, S. Radaeva, and B. Gao Crucial Role of IL-4/STAT6 in T Cell-Mediated Hepatitis: Up-Regulating Eotaxins and IL-5 and Recruiting Leukocytes J. Immunol., September 15, 2003; 171(6): 3233 - 3244. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
