ARTICLE
Auteur(s) : Noemi Hernadfalvi1, Wolfgang
Langhans1, Claudia von Meyenburg1, Brigitte
Onteniente2, Denis Richard3, Denis
Arsenijevic1
1Institute of Animal Sciences, ETH Zurich,
Schwerzenbach, Switzerland
2INSERM UMR421, Creteil, France
3Department of Anatomy and Physiology, Laval University,
Quebec, Canada
accepté le 4 Juin 2007
Inflammation, infection and administration of bacterial
lipopolysaccharide (LPS) cause oxidative stress, and this entails
an increase in tissue concentrations of reactive oxygen species,
which are believed to mediate tissue dysfunction and/or
destruction. Consequently, antioxidants such as glutathione (GSH)
[1] are major determinants of the degree of pathology in models of
tissue inflammation [2]. GSH is mainly produced in the liver and
controls the redox state in many tissues [3]. GSH has also been
supposed to limit LPS-induced pathology [4], but so far it is
unknown whether GSH also affects LPS anorexia. GSH can modify
cytokine production by several mechanisms [5] and may therefore
regulate anorexic cytokine production in response to LPS. Tissue
GSH may, in fact, be one determinant of cytokine pathology. Some
evidence indicates that, in addition to an increase in total GSH,
an increase in mitochondrial GSH is important for its protective
effect [6-9]. In this study, we investigated the role of GSH in the
food intake response to LPS in LPS-naïve and LPS-pretreated mice.
LPS pretreatment normally reduces the pathophysiological responses
and the anorexia in response to subsequent LPS administrations, a
state often referred to as LPS tolerance [10]. If GSH modulates LPS
anorexia, its production and tissue levels in response to LPS may
differ between LPS-naïve and LPS-pretreated animals. To examine
further a possible causal relationship between GSH and LPS
anorexia, we also determined whether pharmacological GSH depletion
in LPS-pretreated animals changes the feeding response to LPS, and
whether total or mitochondrial GSH levels in brain and liver are
indicative of tolerance to LPS-induced anorexia.
Methods
Mice, housing and diets
We used male, 8-16 week old, 25-30g (body weight) c57bl6 mice for
all experiments. Mice were housed individually in stainless steel
cages with a grid floor and kept on a 12:12 hour light:dark cycle.
Drug solutions were freshly prepared and always injected 1-2 hours
before the onset of the dark cycle. Mice had continuous ad libitum
access to water and powdered laboratory chow (Nafag, Gossau,
Switzerland). All procedures were approved by the Canton of Zurich
Veterinary Office.
Experiment 1: Effects of LPS
Design (table 1) and feeding
LPS from Eschericha coli O111:B4 (L2630, Sigma, Buchs, Switzerland)
was dissolved in pyrogen-free saline (Braun, Emmenbrucke,
Switzerland) and injected intraperitoneally (ip) at a dose of
4μg/mouse twice, on days 1 and 4 of the experiment, when the
animals’ food intake had returned to pre-injection values. Control
animals received an equivalent volume of vehicle. The LPS dose
chosen was the lowest dose which caused a reliable food intake
reduction for 24 hours in a preliminary experiment in WT mice with
0, 0.4, 4 and 40 μg per 25 g body weight (n = 8/dose).
Food intake was measured (± 0.1g) at 6, 12 and 24 hours after
injection by subtracting the weight of the food container and
taking into account any spillage. At the end of the experiment,
i.e., on day 5, we collected blood, brain and liver to determine
total reduced GSH (see below). Each treatment group shown in table 1 consisted of 7-8 mice.
Table 1 Time lines for treatments and food intake
measurements in mice
Glutathione (GSH) in serum, brain and liver
For determination of trGSH, mice were sacrificed with
CO2 24 hours after the second injection of LPS or
saline. Blood was collected by heart puncture and centrifuged.
Serum was kept at -80C° until analysis. Brains and livers were
collected, snap frozen in liquid nitrogen, and stored at - 80C°
until analysis. Throughout the extraction procedure for GSH,
samples were kept on ice. Tissues were homogenized using a polytron
PT1200 homogenizer (Kinematica, Lucern, Switzerland) with
phosphate-buffered saline (600 μl/100 mg tissue). The solution
underwent a precipitation reaction with phosphate solution (glacial
metaphosphoric acid 1.67g, disodium EDTA 0.2 g, NaCl
30 g, in 100 ml distilled water). This solution was
centrifuged, the supernatant was collected, and phosphate buffer
(0.3M Na2HPO4) added. From this an aliquot
was taken and the trGSH tissue levels were measured using a method
based on the formation of a chromophoric product resulting from the
reaction with 0.04% 5,5-dithiobis-2-nitrobenzoic acid (D8130, DTNB,
Sigma Chemicals, St. Louis, MO, USA) and GSH (G4251, Sigma, Buchs,
Switzerland) [11].
Mitochondrial GSH in brain and liver
In a separate group of mice, we also examined whether LPS resulted
in altered mitochondrial GSH levels in brain and liver. On day, 5
mice (n=4, each) were perfused with ice-cold PBS, brains and livers
were removed and mitochondria were isolated as follows [11]:
tissues were homogenized in a glass tissue grinder in a solution
corresponding to 600 μl (0.25 M sucrose – 1mM EDTA) per 100 mg of
tissue. This suspension was overlayed onto 0.34 M sucrose – 1mM
MgCl2, and 20 mM phosphate buffer (pH 7.4) and
centrifuged at 700 g (10 min, 4°C). The resulting supernatant
was centrifuged at 5000 g for 15 min at 4°C. The pellet was
re-suspended in 40 ml of 0.25 M sucrose – 1mM EDTA and centrifuged
for 10 min at 24,000 g. The resulting pellet was then
re-suspended in 3 ml of 0.25 M sucrose and 2 mM EDTA. Mitochondria
numbers were estimated using the succinate dehydrogenase (SDH)
activity assay from the mitochondrial preparation. Two μl were used
to determine activity by the reduction of 200 μl of
succinate/nitroblue tetrazolium solution. From the minimal activity
of 0.14 optical density (OD)/min, all isolates were normalized. GSH
was then determined as described above.
Experiment 2: Effects of GSH depletion on the responses to
LPS
Design and feeding response to LPS
To determine whether GSH depletion alters the animals’ feeding
response to 4μg LPS, we inhibited GSH production and, hence,
reduced GSH levels by injecting saline or LPS-pretreated mice with
diethylmaleate (DEM, D97703, Sigma, Buchs, Switzerland) and
L-buthionine sulfoximine (BSO, B2640, Sigma, Buchs, Switzerland)
(DEM/BSO = db) at 800μl/kg body weight sc and 200 mg/kg [12] body
weight ip, 2 hours before the second LPS injection on day 4.
Preliminary trials showed that at these doses, the combination of
both inhibitors reduced trGSH in the liver 2 hours after
administration. This resulted in the following six treatment groups
(n=4, each; first injection/second injection with or without db
treatment): 1) saline/saline no db (ss[-db]), 2) ss[+db], 3)
LPS/LPS no db (ll[-db]), 4) ll[+db], 5) LPS/saline db treatment
(ls[+db], and 6) saline/LPS db treatment (sl[+db]). Food intake was
measured at 6, 12 and 24 hours after the second LPS injection.
Serum acute phase protein response to LPS
To determine whether GSH depletion altered the acute phase protein
response, we determined serum proteins by cellulose gel
electrophoresis (Paragon SLE gel – Beckman, Nyon, Switzerland).
Mice groups (n = 4, each) were defined as ss[-db],
ssdb[+db], ls[-db], sl[+db], ll[-db], lldb[+db] and lsdb[-db]. Mice
were sacrificed with CO2, 24 hours after the second
injection of LPS or saline. Serum was collected by heart puncture
and stored at -80°C prior to use. Two μl of serum were loaded on a
tract and gels were run for 30 minutes in barbital buffer. Gels
were then stained with paragon blue and scanned with the Scion
Image program. We also determined whether GSH depletion by DEM/BSO
altered serum proteins. Total serum proteins levels were determined
by the Bradford protein assay (cat. No. 500-0006, Bio-Rad, Reinach,
Switzerland).
Serum tumour necrosis factor-α (TNFα) and interferon-γ (IFNγ)
response to LPS
To determine whether GSH depletion in LPS-pretreated animals
altered TNFα and IFNγ serum levels 24h after the second LPS
injection, we used serum from the electrophoresis experiment (see
above) and analyzed it for TNFα and IFNγ using IFNγ and TNFα murine
ELISA kits from Amersham (Otelfingen, Switzerland) [13].
Statistics
Differences between groups were determined by using Kruskal-Wallis
test with a Dunn post hoc test. A value of p<0.05 was considered
significant. Non-parametrical Spearman correlation coefficients
were calculated using the Instat program (GraphPad Incorporation,
San Diego, CA, USA).
Results
Experiment 1: Effects of LPS
Feeding
Administration of LPS (4 μg/mouse) on day 4 reduced food intake
compared to twice saline injected (ss) mice at 6, 12, and 24 hours
in LPS naïve mice (sl) (figure 1).
LPS-pretreated mice which received a second LPS injection (ll)
showed a significant reduction in food intake at 6 and 12 hours
compared to saline (ss) controls (figure 1), but this
difference did not persist over 24 hours. In addition, ll mice ate
more than sl mice at 6, 12 and 24 hours after the second injection
(figure 1),
n = 8, p < 0.01).
Serum, brain and liver GSH response to LPS
LPS reduced serum and brain trGSH at 24 hours after the second
injection in LPS-naïve mice (sl) compared to twice saline-injected
mice (ss) (p < 0.01). LPS-pretreated mice, however,
responded to a second LPS injection (ll) with a significant
increase in serum and brain trGSH compared to their saline-injected
controls (p < 0.001) (figure 2). In
addition, ls mice had higher trGSH levels in liver than saline
controls (ss) (p < 0.001) (figure 2).
Mitochondrial GSH in brain and liver
Four days after injection, mitochondrial GSH levels were
significantly increased in brains and livers of LPS-treated mice
compared to their corresponding controls (figure 3).
Experiment 2: Effects of GSH depletion on the responses to
LPS
Feeding response to LPS
In Experiment 2, we confirmed our previous results that LPS
injection reduced food intake compared to saline injection in
saline-pretreated control mice (sl[-db] versus ss[-db]) at 6, 12
and 24 hours after injection (figure 4). Again, LPS
pretreatment attenuated the effect of LPS at all time points
measured (sl[-db] versus ll[-db], p < 0.01) (figure 4, also
see figure 1).
GSH depletion with DEM/BSO (db) per se (ss[+db] and ls[+db] mice)
did not affect food intake compared to saline pretreated mice
(ss[-db]) or LPS-pretreated hyposensitive mice (ll[-db]). Yet,
DEM/BSO treatment 2 hours prior to the second LPS injection
antagonized the hyposensitivity to LPS-induced anorexia, i.e., it
re-established the anorexia, in ll[+db] mice (figure 4). Ll[+db]
mice showed significant anorexia compared to their controls
(ll[-db], ss[+db]) at 6 hours and 12 hours (both time points
p < 0.01), but the effect did not persist over 24
hours.
Serum acute phase protein response to LPS
GSH depletion per se resulted in no major change in serum protein
levels (ss[+db] versus ss[-db]). Similar responses were observed in
all groups, with the exception of the ll[+db] group, in which serum
protein levels were significantly increased (figure 5A). Serum
electrophorograms indicate that this increase in ll[+db] mice was
due to a general increase rather than an increase in a specific
protein (figure 5B).
Serum TNFα and IFNγ response to LPS
Circulating serum IFNγ levels were increased in mice that received
LPS for the first time (sl[-db]) and in LPS-pretreated animals
treated with DEM/BSO and receiving a second LPS injection
(ll[+db]), whereas LPS-pretreated animals without db treatment
(ll[-db]) showed no increase in either cytokine one day after the
second LPS treatment (figure 6).
Discussion
Consistent with previous findings [14, 15], we report here that ip
LPS reduced food intake in LPS-naïve mice, whereas LPS-pretreated
mice showed an attenuated anorexia in response to the second LPS
injection. Interestingly, this phenomenon was associated with
increased glutathione (GSH) levels in brain and liver, raising the
possibility that the observed hyposensitivity to the second LPS
injection was due to high tissue GSH levels. Altering antioxidant
levels in various animal models has been shown to change the
release of cytokines in response to immune challenges. Thus,
LPS-induced cytokine production was modified by decreasing or
increasing the GSH content with DEM/BSO [18, 19] or
N-acetylcysteine [20, 21], respectively. Serum IFNγ levels appeared
to be sensitive to GSH changes in LPS-pretreated animals as they
were increased 24 h after the second LPS treatment when GSH
was decreased. Our data suggest that LPS has a biphasic effect on
tissue GSH levels. Whereas the first LPS injection led to an acute
decrease in serum and brain GSH, presumably because the organism
used GSH to cope with the LPS-induced oxidative stress, four days
after the first LPS injection, mice had increased liver GSH,
perhaps as a result of stimulated hepatic GSH production in
response to the initial demand. This additional GSH appeared to be
rapidly available when the mice were confronted with the second LPS
injection because in this situation they showed higher serum and
brain GSH levels when compared to the twice saline-injected control
mice. Thus, the stimulated production and rapid distribution of GSH
rather than elevated tissue levels of GSH alone may contribute to
LPS hyposensitivity. In line with this hypothesis, depletion of GSH
rendered LPS-pretreated mice susceptible to LPS anorexia,
suggesting that GSH can curb LPS-induced anorexia. Of course, we
can also not exclude that a difference in GSH turnover rates
between LPS hypo-responsive and LPS-naive mice contributed to the
observed differences in GSH levels because we did not measure GSH
turnover. Finally, we also demonstrate that altered intracellular
GSH repartitioning occurs in LPS hyposensitivity in both brain and
liver. Thus, we observed an increase in mitochondrial GSH in
LPS-pretreated mice compared to LPS-naïve mice. In other oxidative
stress models, increases in mitochondrial GSH are associated with
an increased resistance to the degree of pathology [6-9], and yet
other studies have shown that modification of mitochondrial
activity can result in resistance to LPS-induced anorexia [21-23].
Further studies using mice with genetically altered mitochondrial
GSH levels could be used to critically examine the role of
mitochondrial GSH in LPS anorexia by using 1) increased
mitochondrial GSH transport [24] in LPS-naïve mice and 2) decreased
mitochondrial GSH [25, 26] in LPS tolerance studies. A change in
the redox state can alter the immune response due to changes in the
production of cytokines and/or acute phase proteins. It has been
shown that acute phase proteins, in particular an increase in
lipoproteins (found in the beta-globulin faction), can attenuate
some of the responses to LPS and their severity [16, 17]. Our
electrophoretograms revealed that depletion of GSH per se did not
result in a marked, qualitative decrease in the beta-globulins. The
lack of major changes between ss[-db] and ss[+db] serum protein
levels and the bands at 24 hours after treatment suggests that the
difference between ll[-db] and ll[+db] mice was due to specific
effects of GSH depletion on the response to LPS. In addition, we
can exclude a general liver toxicity effect of db treatment at this
time point because albumin levels were not reduced in db treated
mice. This suggests that the reduction of GSH by db was not due to
general liver toxicity/failure. In summary, these results suggest
that LPS affects GSH levels in a tissue-specific manner and that
stimulation of GSH production and its rapid distribution in the
organism rather than absolute GSH levels confer hyposensitivity to
LPS anorexia. The intracellular repartition of GSH may also be
important because hypo-responsive animals had increased
mitochondrial GSH and did not display anorexia 24 hours after the
second LPS treatment. Further support for a role of GSH in the
feeding response to LPS is derived from the fact that mice that are
hypo-responsive to the anorexic effect of LPS because of LPS
pretreatment can be rendered sensitive again by pharmacological GSH
depletion. The generality of this phenomenon, i.e., whether for
instance IFNγ-KO mice or other proinflammatory cytokine-deficient
mice can also be made more responsive to LPS-induced anorexia by
pharmacological GSH depletion remains to be examined.
References
1 Haddad JJ, Harb HL.
L-gamma-Glutamyl-L-cysteinyl-glycine (glutathione; GSH) and
GSH-related enzymes in the regulation of pro- and anti-inflammatory
cytokines: signalling transcriptional scenario for redox(y)
immunological sensor(s). Mol Immunol 2005; 42: 987.
2 Bizzozero OA, Ziegler JL, de Jesus G,
Bolognani F. Acute depletion of reduced glutathione causes
extensive carbonylation in rat brain proteins. J Neurosci Res 2006;
83: 656.
3 Reid M, Jahoor F. Glutathione in disease. Curr Opin
Clin Nutr Metab Care 2001; 4: 65.
4 Villa P, Saccani A, Sica A, Ghezzi P.
(2002) Glutathione protects mice from lethal sepsis by limiting
inflammation and potentiating host defense. J Infect Dis 2002; 185:
1115.
5 Haddad JJ. Redox regulation of pro-inflammatory cytokines
and IKB-a/NF-KB nuclear translocation and activation. Biochem
Biophys Res Commun 2003; 296: 847.
6 Xu F, Putt DA, Matherly LH, Lash LH.
Modulation of expression of rat mitochondrial 2-oxoglutarate
carrier in NRK-52E cells alters mitochondrial transport and
accumulation of glutathione and susceptibility to chemically
induced apoptosis. J Pharmacol Exp Ther 2006; 316: 1175.
7 Martensson J, Jain A, Stole E, Frayer W,
Auld PAM, Meister A. Inhibition of glutathione synthesis
in the new born rat: A model for endogenously produced oxidative
stress. Proc Natl Acad Sci USA 1991; 88: 9360.
8 Meister A. Glutathione deficiency produced by inhibition
of its synthesis and its reversal; applications in research and
therapy. Pharmacol Ther 1991; 51: 155.
9 Shertzer HG, Bannenberg GL, Zhu H, Liu RM,
Moldeus P. The role of thiols in mitochondrial susceptibility
to iron and tert-butyl hydroperoixde-mediated toxicity in cultured
mouse hepatocytes. Chem Res Toxicol 1994; 7: 358.
10 Faggioni R, Fantuzzi G, Villa P,
Buurman W, van Tits LJH, Ghezzi P. Independent
down-regulation of central and peripheral tumor necrosis factor
production as a result of lipopolysaccharide tolerance in mice.
Infect Immun 1995; 62: 1473.
11 de Bilbao F, Arsenijevic D, Vallet P,
Hjelle OP, Ottersen OP, Bouras C, Raffin Y,
Abou K, Langhans W, Collins S, Plamondon J,
Alves-Guerra MC, Haguenauer A, Garcia I,
Richard D, Ricquier D, Giannakopoulos P. Resistance
to cerebral ischemic injury in UCP2 knockout mice: evidence for a
role of UCP2 as a regulator of mitochondrial glutathione levels. J
Neurochem 2004; 89: 1283.
12 Pileblad E, Magnusson T. Effective depletion of
glutathione in rat striatum and substantia nigra by L-buthionine
sulfoximine in combination with 2-cyclohexene-1-one. Life Sci 1990;
47: 2333.
13 Arsenijevic D, Garcia I, Vesin C,
Vesin D, Arsenijevic Y, Seydoux J, Girardier L,
Ryfell B, Richard D. Differential roles of tumor necrosis
factor-α and interferon-γ in mouse hypermetabolic and anorectic
responses induced by LPS. Eur Cytokine Netw 2000; 11: 662.
14 Langhans W, Balkowski G, Savoldelli D.
Differential feeding responses to bacterial lipopolysaccharide and
muramyl dipeptide. Am J Physiol 1991; 261: R659.
15 Faggioni R, Fantuzzi G, Villa P,
Buurman W, van Tits LJ, Ghezzi P. Independent
down-regulation of central and peripheral tumor necrosis factor
production as a result of lipopolysaccharide tolerance in mice.
Infect Immun 1995; 63: 1473.
16 Feingold KR, Staprans I, Memon RA,
Moser AH, Shigenaga JK, Doerrier W,
Dinarello CA, Grunfeld C. Endotoxin rapidly induces
changes in lipid metabolism that produces hypertiglyceridemia: low
doses stimulate hepatic triglyceride production while high doses
inhibit clearance. J Lipid Res 1992; 33: 1765.
17 Feingold KR, Funk JL, Moser AH,
Shigenaga JK, Rapp JH, Grunfeld C. Role of
circulating lipoproteins in protection from endotoxin toxicity.
Infect Imm 1995; 63: 2041.
18 Wang F, Wang LY, Wright D, Parmely MJ.
Redox balance differentially inhibits lipopolysaccharide-induced
macrophage activation in the mouse liver. Infect Immun 1999; 67:
5409.
19 Jafari B, Ouyang B, Li LF, Hales CA,
Quinn DA. Intracellular glutathione in stretch-induced
cytokine release from alveolar type-2 like cells. Respirology 2004;
9: 43.
20 Victor VM, Guayerbas N, Garrote D, Del
Rio M, De la Fuente M. Modulation of murine macrophage
function by N-acetylcysteine in model of endotoxic shock.
Biofactors 1999; 10: 347.
21 Baud O, Haynes RF, Wang H, Folkerth RD,
Li J, Volpe JJ, Rosenberg PA. Developmental
up-regulation of MnSOD in rat oligodendrocytes confers protection
against oxidative injury. Eur J Neurosci 2004; 20: 29.
22 Johnson BJ, Le TTT, Dobbin CA, Banovic T,
Howard CB, de Maria Leon Flores F, Vanagas D,
Naylor DJ, Hill GR, Suhrbier A. Heat shock protein
10 inhibits 14 lipopolysaccharide induced inflammatory mediator
production. J Chem Biol 2005; 280: 4037.
23 Guidot DM. Endotoxin pre-treatment in vivo increases the
mitochondrial respiratory capacity in rat hepatocytes. Arch Biochem
Biophys 1998; 354: 9.
24 Buetler TM. Identification of glutathione S-transferase
isoenzymes and γ-glutamylcysteine synthetase as negative
acute-phase proteins in rat liver. Hepatology 1998; 28: 1551.
25 Will Y, Fischer KA, Horton RA,
Kaetzel RS, Brown MK, Hedstrom O, Lieberman MW,
Reed DL. γ-Glutamyltranspeptidase-deficient knockout mice as a
model to study the relationship between glutathione status,
mitochondrial function and cellular function. Hepatology 2000; 32:
740.
26 Esposito LA, Kokoszka JE, Waymire KG,
Cottrell B, MacGregor GR, Wallace DC. Mitochondrial
oxidative stress in mice lacking the glutathione peroxidase-1 gene.
Free Radic Biol Med 2000; 28: 754.
|
Printable version
Version PDF
