| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Physiology, Ludwig-Maximilians-University, Munich, Germany
1Correspondence: Institute of Physiology, Schillerstr. 44, 80336 Munich, Germany. E-mail: s.zahler{at}lrz.uni-muenchen.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(2.5 ng/ml, 4 h) were measured
(flow cytometry and immunoassay), as were nuclear translocation of the
transcription factor NF
B (Western blotting, confocal microscopy) and
redox status of HUVECs (quantification of glutathione by HPLC). TNF-
elevated IL-6 in the cell supernatant from 8.8 ± 1 to 41 ±
3 pg/ml and IL-8 from 0.5 ± 0.03 to 3 ± 0.2 ng/ml. ICAM-1
was increased threefold and E-selectin rose eightfold. Oxidative stress
(decrease of glutathione by 50%) reduced post-TNF- levels of IL-6
to 14 ± 3 and IL-8 to 1 ± 0.2; the rise of ICAM-1 was
completely blocked and E-selectin was only doubled. The
anti-inflammatory effects of preconditioning via oxidative stress were
paralleled by reduction of the translocation of NFB on stimulation
with TNF-, and antagonized by the intracellular radical scavenger
N-acetylcysteine. ‘Anti-inflammatory preconditioning’ of
endothelial cells by oxidative stress may account for the inhibitory
effects of preconditioning on leukocyte adhesion in
vivo.—Zahler, S., Kupatt, C., Becker, B. F. Endothelial
preconditioning by transient oxidative stress reduces inflammatory
responses of cultured endothelial cells to TNF-.
Key Words: cytokine • adhesion molecule • NFB • glutathione
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. In the light of yearly rising numbers of cardiac
interventions invoking transient ischemia (PTCA, thrombolysis, coronary
bypass grafting, valve replacement, transplantation, CPB), the
incidence of reperfusion, and thus reperfusion injury, gains
ever-increasing relevance. Reperfusion is restricted mostly to clinical
settings, and therefore is a planned intervention. Accordingly,
awareness has grown that patients might benefit substantially from a
deeper knowledge of the mechanisms underlying reperfusion injury, since
this should lead to subsequent development of protective strategies.
To date, the most powerful tool for experimentally reducing infarct
size after ischemia and reperfusion is the so-called ‘ischemic
preconditioning’. This phenomenon was first described by Murry and
co-workers (3)
, who have shown that brief, transient
episodes of ischemia ‘precondition’ myocardium to become more
resistant against a subsequent severe ischemic insult. Since then the
research interest in this effect has grown constantly, and a variety of
potential mechanisms have been proposed: adrenergic stimulation
(4)
, activation of adenosine receptors (5)
,
opening of KATP channels (6)
,
induction of heat shock proteins (7)
, and induction of
oxidative stress (8)
, just to mention those that are
established best. However, the effector cells of preconditioning are
only poorly defined. Apart from parenchymal cells, especially cells of
the vascular wall seem of interest, because reperfusion injury is, at
least in part, a misguided inflammatory response (9)
. In
this sense, ischemic preconditioning has previously been shown to
reduce adhesion of leukocytes to postischemically reperfused vessels
(10)
.
The proinflammatory mediator tumor necrosis factor (TNF-) plays
a key role during ischemia and reperfusion (11
, 12)
,
modulating cell adhesion molecules, cytokines, and chemokines. We have
therefore investigated whether the responses of endothelial cells to
TNF- can be modulated by a protocol mimicking ischemic
preconditioning. The stimulus chosen was that of brief redox stress
imposed by transient application of hydrogen peroxide. Oxidative stress
has been shown to occur during preconditioning (13)
and to
be mandatory for the success of preconditioning in some models
(14
, 15)
. This also seemed of broader interest because the
modulation of cellular redox status has recently turned out to be a
major signal for inflammatory reactions (16
, 17)
.
Preconditioning at first seemed to be a temporally limited phenomenon,
being acutely effective only within periods of less than 1 h.
Recently, however, a ‘second window of protection’ has been
identified, occurring some 24 h after the preconditioning period
(7
, 18)
. Thus, we also tested whether an anti-inflammatory
effect still pertains after a 24 h lag phase between
preconditioning and inflammatory stimulus still allows.
The inflammatory parameters measured in cultures of human umbilical
vein endothelium were the expression of the adhesion molecules ICAM-1,
E-selectin, and P-selectin and secretion of the cytokines interleukin 6
(IL-6) and IL-8. The redox status of the cells was monitored by
separate measurement of the intracellular concentration of reduced
glutathione (GSH). Application of the GSH precursor N-acetylcysteine
served to manipulate the intracellular redox status. Potential effects
of the experimental protocols on cell apoptosis and necrosis were
investigated by measurement of annexin-V binding and of lactate
dehydrogenase (LDH) release. Involvement of the transcription factor
NFB (nuclear factor B), which has been shown to be both redox
sensitive (19)
and to participate in the regulation of the
inflammatory parameters listed above, was also studied.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Briefly, the umbilical vein was cannulated, rinsed
with phosphate buffer, and filled with collagenase A (1 mg/ml) for 5
min at 37°C. Endothelial cells were eluted with medium 199,
centrifuged, washed, pelleted again, and then seeded with endothelial
cell growth medium (Promocell, Heidelberg, Germany) into T 25 culture
flasks. After reaching confluence, the primary cultures were detached
from the flask with trypsin and transferred into 24-well culture dishes
coated with collagen A (Biochrom, Berlin, Germany). Experiments were
performed 2 days after the cells had reattained confluence.
Flow cytometry
The expression levels of the adhesion molecules ICAM-1,
E-selectin, and P-selectin on HUVECs, as well as the binding of
annexin-V to these cells, were quantified by flow cytometry. Cell
monolayers were washed with phosphate-buffered saline (PBS) without
calcium and incubated with a trypsin solution (in the case of
P-selectin, a collagenase solution, P-selectin being rapidly activated
in response to trypsin) until the cells were detached from the culture
dish. The cells were treated with CellFix (Becton Dickinson,
Heidelberg, Germany) and passed through a nylon net (70 µm mesh). The
cells were pelleted by centrifugation, resuspended in CellWash (Becton
Dickinson), labeled with the respective antibodies, washed, and
measured on a FACScan flow cytometer (Becton Dickinson). The antibodies
MCA675PE, MCA883F, and MCA796PE (all Serotec, Kidlington, U.K.) were
used to detect ICAM-1, E-selectin, and P-selectin, respectively. Data
analysis was performed with Lysis II software (Becton Dickinson). The
median of the specific fluorescence intensity was used as a marker for
expression of the respective epitope; nonspecific fluorescence was
detected by using isotype-matched nonbinding antibodies (Serotec) and
subtracted. Binding of annexin-V to HUVECs was determined with an
apoptosis detection kit (PharMingen, Heidelberg, Germany) containing
FITC-labeled annexin-V, propidium iodide, and a calcium-rich binding
buffer. The staining protocol complied with the manufacturer’s
instructions. Median fluorescence intensity of FITC on propidium iodide
negative (i.e., living) cells was used as a measure of apoptosis.
Photometric tests
Levels of the cytokines IL-6 and IL-8 were determined in cell
culture supernatants with commercially available ELISAs (Endogen,
Woburn, Mass.). To determine IL-8, the samples were diluted 1:100.
Sample aliquots of 50 µl were used per well of the test plate in all
cases; staining and measurement were performed as proscribed by the
manufacturer. Briefly, the ELISA plates were incubated with the
standards or samples for 60 min. Unbound cytokine was removed by
washing three times, and a second antibody (labeled enzymatically for
photometric detection) was added. After further incubation and removal
of unbound antibody by washing, the plates were analyzed
photometrically with a microplate reader (Dynatech, Guernsey, U.K.) at
450 nm. As an indicator of cellular death, release of the enzyme LDH
into the supernatant was also measured photometrically at 490 nm
(CytoTox96 assay, Promega, Mannheim, Germany).
Measurement of GSH
After removal of medium, HUVECs were lysed with 400 µl/well of
a stopping solution (64 mM N-ethylmaleimide in 0.5 M perchloric acid)
to prevent autoxidation of GSH. Concentrations in the lysate are termed
intracellular, no attempt having been made to correct this to the real
intracellular volume. GSH was analyzed directly by high-performance
liquid chromatography (HPLC) as the NEM adduct at a wavelength of 202
nm. Two hundred microliters of the respective sample were applied to a
5 µm C-18 nucleosil column (Macherey and Nagel, Düren,
Germany); 25 mM
NH4H2PO4
containing 1.2% methanol served as eluent. At a flow of 1 ml/min, the
retention time was ~13 min.
Western blot analysis
HUVECs were lysed in a reducing Triton lysis buffer (Tris 1 mM,
NaCl 50 mM, Triton 1x, sodium vanadate 5 mM, NaF 50 mM, Na
pyrophosphate 30 mM, DTT 1 mM, PMSF 1 mM, leupeptin 10 µM, pepstatin
10 µM). Nuclear and cytosolic fractions were separated by
centrifugation (10,000 g for 5 s), the membrane
fractions were removed by 15 min centrifugation at 10,000 g
(4°C). The protein concentrations of the lysates (nuclear and
cytosolic) were determined with a detergent compatible assay (BCA,
Pierce, Rockford, Ill.). Aliquots (40 µg) of protein were separated
on a 12% polyacylamide gel, transferred to a nitrocellulose membrane
(Hybond-enhanced chemiluminescence, Amersham, Braunschweig, Germany),
blocked with buffer containing 5% nonfat dry milk, and incubated with
2 µg/ml of primary antibody (anti-p65 subunit of NFB or
anti-IB, both from Santa Cruz Biotechnology, Santa Cruz, Calif.).
After washing four times with buffer containing 1% nonfat dry milk,
the membranes were exposed to 0.2 µg/ml of secondary antibody,
labeled with horseradish peroxidase (Santa Cruz). The membranes were
then incubated with an enhanced chemiluminescence kit (Pierce) for 1
min and exposed to an X-ray film. The films were analyzed with a video
system (GelDoc 1000, Bio-Rad, Hercules, Calif.) and optical density of
the bands was calculated with MolecularAnalyst software (Bio-Rad). Due
to considerable interindividual variability, values are expressed as
percentual changes vs. controls (100%). Analogous experiments were
performed with the cytoplasmatic fractions to control the purity of the
nuclear and cytoplasma preparation, respectively.
Confocal microscopy
To determine the subcellular localization of NFB, stimulated
HUVECs were fixed with buffered formaldehyde (3%) and subsequently
permeabilized by submersion in 0.2% Triton X-100 in PBS for 2 min. The
samples were rinsed three times with PBS, blocked with 0.2% BSA in PBS
for 15 min, and incubated with the primary antibody against p65 (Santa
Cruz) for 45 min. After four washings, the secondary antibody (FITC
labeled, Santa Cruz) was added and samples were incubated for 30 min at
room temperature. After four final washing steps, the cells were
covered with mounting medium (Sigma, Eching, Germany) and a coverslip
before inspection with a confocal microscope (LSM 410 Invert, Zeiss,
Jena, Germany).
Experimental protocols
Flow charts of the experimental protocols A–C are
depicted in Fig. 1
. In preconditioning experiments, medium was removed from confluent
HUVECs and replaced by PBS containing 1 mM
H2O2 for 5 min at 37°C.
Afterward, the supernatant was removed and the cells were covered with
standard medium again. The cells were then incubated with TNF- (2.5
ng/ml) for 4 h, after which supernatant was sampled for
measurement of IL-6, IL-8, and LDH. Cells were detached and treated for
flow cytometric analysis of ICAM-1, E-selectin, P-selectin, and
annexin-V binding (Fig. 1A
). Time matched controls were
treated with neither H2O2
nor TNF-. Further experimental groups consisted of cells that had
been treated with H2O2 (5
min) or TNF- (4 h) alone. In protocol B, these
experiments were repeated in cells that had been pretreated with the
intracellular radical scavenger N-acetylcysteine (NAC, 1 mM) for 30
min. After this incubation, the cells were washed to remove all
extracellular NAC and then subjected to the aforementioned procedures.
Some of the experiments of protocols A and B were
terminated 5 min or 30 min after the
H2O2 stimulus to allow for
determination of intracellular GSH or for analysis of the subcellular
localization of NFB or IB (Fig. 1)
. To investigate whether
transient redox stress causes a second window of protection against
stimulation with TNF-, a 24 h interval was allowed between
stimulation with H2O2 and
TNF-. During this time the cells were kept in the incubator with
standard medium. Afterward, TNF- was given for 4 h and the
experiments were continued as described above (Fig. 1
, protocol
C).
|
Statistical procedures
Data are usually expressed as mean ± SE;n values are shown in Results. Statistical analysis
was performed with 2-way ANOVA. One factor for group comparisons was
treatment with TNF- or not; the other factor was treatment with
H2O2 or not. Pairwise
multiple comparison tests were carried out using the
Student-Newman-Keuls test. Differences between groups were considered
significant for P 
0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Short incubation with
H2O2 (1 mM for 5 min) did
not alter ICAM-1 expression within 4 h. In contrast, stimulation
with TNF- (2.5 ng/ml for 4 h) elevated ICAM-1 ~threefold
(10.6±0.8, Fig. 2
). Pretreatment with
H2O2 strongly reduced the
effect of TNF- (5.9±0.7, Fig. 2
). This action was not observed when
stimulation with TNF- occurred 24 h after treatment with
H2O2 (Fig. 2)
.
Preincubation of the HUVECs with the intracellular radical scavenger
NAC (1 mM, 30 min) did not alter constitutive ICAM-1 levels or the
response to TNF- (4.0±0.7 and 12.9±1.1, respectively, Fig. 4
),
whereas the inhibitory effect of
H2O2 on TNF--dependent
ICAM-1 expression was abolished (9.9±2.5).
|
|
Expression of the adhesion molecule E-selectin was regulated in an
analogous way. Basal expression (4.0±0.2, Fig. 2
) was unaltered by
H2O2 alone, whereas TNF-
led to an eightfold increase (31±3, Fig. 2
). This was dramatically
infringed by pretreatment with
H2O2 (10.8±2, Fig. 2
).
Again, there was no second window after 24 h. NAC did not
influence the constitutive E-selectin level (3.5±0.3) or the effect of
TNF- (32.2±4.8, Fig. 4
). However, the blocking action of
H2O2 was nearly neutralized
by NAC (27.5±3.7).
The presence of P-selectin behaved in a different way:
H2O2, as well as TNF- or
the combination of both stimuli, elevated basal expression from
4.15 ± 0.13 to 5.38 ± 0.17, 5.14 ± 0.16, and
5.88 ± 0.34, respectively (Table 1
).
|
Secretion of cytokines
Cytokines IL-6 and IL-8 were both constitutively secreted by
HUVECs (8.9±1.4 and 450.4±36 pg/ml, respectively, Fig. 3
). H2O2 alone caused no
changes, whereas TNF- increased the levels of both cytokines
substantially to 40.6 ± 3.3 for IL-6 and 2960 ± 192 pg/ml
for IL-8 (Fig. 3)
. These increases were mitigated by application of
H2O2 prior to TNF-
(14.4±2.9 and 1015±159 pg/ml, resp., not significant vs. control).
When TNF- was applied with a latency of 24 h after
H2O2 treatment, this
inhibitory effect was only marginal and not significant. Preincubation
of the cells with NAC decreased constitutive levels of IL-6 release to
3.2 ± 1.5 pg/ml, but did not affect the increase mediated by
TNF- (45.6±2 pg/ml, Fig. 4
). The loss of efficacy of TNF- after
H2O2 was less pronounced
under the influence of NAC (25±4 pg/ml, Fig. 4
). IL-8 release was not
determined in this series.
|
Levels of GSH in cell lysates
Intracellular GSH levels, as determined in cell lysates, are
illustrated in Fig. 5
. Control levels of GSH in HUVEC lysates were 0.48 ± 0.03 µM.
This value was reduced by half (0.25±0.01 µM) 5 min after
stimulation with H2O2,
whereas TNF- (5 min) had no effect (0.53±0.03 µM). The loss of
GSH due to H2O2 was not
aggravated or reduced by additional application of TNF- (0.26±0.03
µM). Pretreatment with NAC doubled the constitutive GSH level
(0.85±0.04 µM). Though
H2O2 also caused a
reduction of GSH in this protocol (0,44±0.03 µM), the resulting
concentration did not differ from the resting level in untreated cells.
Again, TNF- had no effect, either alone (0.75±0.03 µM) or in
combination with H2O2
(0.38±0.02 µM). At 30 min after stimulation, GSH levels were still
slightly reduced in all the
H2O2-treated cells (Fig. 5)
, but this decrease did not reach statistical significance vs.
controls and the TNF- group.
|
Quantification of necrosis and apoptosis
To establish whether the observed effects on inflammatory
parameters were perhaps caused by modulation of cell necrosis or
apoptosis, LDH levels in the supernatant and annexin-V binding on the
cells were quantified (Fig. 6
). LDH in the supernatant of control cells amounted to 22.5 ± 1.1
mU/ml and did not change significantly in any group. Annexin-V binding
was not altered by H2O2
either alone or in combination with TNF- (Fig. 6)
. TNF-
concentration dependently increased annexin-V-specific fluorescence
from 4.1 ± 0.2 units (control) to 7.4 ± 0.9 at 300 ng/ml
(Fig. 6)
.
|
IB and NFB
IB was consistently found in the cytoplasm of control HUVECs
(defined as 100%). Addition of
H2O2 or TNF- caused
severe reductions in IB levels (31±9% and 53±14%, resp.,
Table 2
). A combination of both stimuli did not show any interference or
augmentation. IB was absent from all nucleus preparations, which
suggests good separation quality (data not shown).The p65 subunit of
NFB was detected in the cytoplasm of resting cells and, to a much
lower extent, in the nuclei. This basal presence of p65 in the nucleus
(set at 100%) was also to be seen in the immunohistochemical confocal
images (Fig. 7A
). Both H2O2 and
TNF-, separately caused increases of nuclear NFB (190±52% and
253±115%, resp., Table 2
and Figs. 7A
, B
). However, the
effect of TNF- on translocation of NFB was totally inhibited by
pretreatment of cells with
H2O2, as shown by Western
blotting and confocal microscopy (Table 2
and Figs. 7A
, B
).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, the phenomenon
of ‘ischemic preconditioning’ has aroused considerable interest. On
the one hand, this is due to the fact that preconditioning is the most
powerful tool known at present to reduce postischemic infarct size
(21)
. On the other hand, the attention results from the
growing clinical relevance and frequency of reperfusion scenarios.
Knowledge of the underlying principles might help to design
pharmacological interventions, keeping tissues at risk in a ‘permanent
state of preconditioning’ (21)
. Accordingly, numerous
studies have been performed to elucidate the mechanisms behind
preconditioning, but the results have suggested nearly as many
different mediators, including adenosine (5)
,
acetylcholine (22)
, catecholamines (4)
,
angiotensin II (23)
, bradykinin (24)
, nitric
oxide (25)
, and reactive oxygen species (8)
.
To make things even more complicated, the cellular targets of
preconditioning seem to be heterogeneous. Though saline perfused
isolated hearts or isolated cardiomyocytes can be preconditioned
(15
, 26)
, other reports suggest a microvascular component
of preconditioning (27
28
29)
. This mainly concerns the
reduction of postischemic inflammation or, more precisely, the
mitigation of leukocyte adhesion in the reperfused vascular bed that
has been described after preconditioning (10
, 30)
.
Accordingly, it is tempting to speculate that the vascular endothelium,
the interface between parenchymal tissue and blood cells, is involved
in this phenomenon.
Oxidative stress has been shown to occur after preconditioning
(13)
and to be a mandatory factor of preconditioning at
least in some models (14)
. Moreover, oxidative stress and
changes in cellular redox status have turned out to be powerful
modulators of intracellular signaling (19)
. For example,
the expression of adhesion molecules and cytokines by endothelial cells
is demonstrably regulated by changes in intracellular redox status
(16)
. Since leukocyte recruitment at the site of
reperfusion is caused by exactly these endothelial factors, we
endeavored to establish whether oxidative stress might elicit a kind of
anti-inflammatory ‘endothelial preconditioning’.
To this end, we subjected HUVECs to 5 min incubation with
H2O2, a reagent that has
turned out to be a reproducible and easily controllable tool for
generating reversible intracellular redox stress (31)
.
TNF- was used as a subsequent inflammatory stimulus because this
cytokine has been shown by us (12)
and others
(11)
to be released during postischemic reperfusion and to
be a key mediator or reperfusion injury. The inflammatory parameters we
measured were chosen in order to cover the whole cascade of leukocyte
adhesion to the vessel wall and subsequent transmigration: P-selectin,
which is stored in endothelial vesicles and can be rapidly transferred
into the cell membrane on stimulation, enabling rapid recruitment of
leukocytes and rolling along the vessel wall (32)
.
E-selectin and ICAM-1 allow for sustained rolling at later time points
and for firm adhesion, respectively. Finally, by chemotactically
activating the leukocytes, IL-6 and IL-8 mediate the transition from
rolling to firm adhesion (sticking), and cause transmigration (for a
review, see ref 33
). P-selectin has previously been shown
to be rapidly expressed (within minutes) on the surface of HUVECs
after treatment with H2O2
(31)
. Four hours later, P-selectin on HUVECs was still
elevated (Table 1)
, but to an extent that does not necessarily involve
de novo synthesis. The same was true for TNF- or the
combination of both stimuli (Table 1)
. The small extent of these
effects might be due to the fact that the promoter region of human
P-selectin lacks B sites, necessary to cause the pronounced
up-regulation of this adhesion molecule, e.g., in mice after
stimulation with TNF- (34)
. In contrast, as was to be
expected, TNF- caused dramatic increases of the expression of ICAM-1
and E-selectin (Fig. 2)
and of the secretion of IL-6 and IL-8 (Fig. 3)
.
Transient brief pretreatment with
H2O2 inhibited all these
inflammatory effects (Figs. 2
and 3)
, an action that could easily be
interpreted as an endothelial preconditioning.
However, alternative explanations are also feasible. For instance,
H2O2 can be a cytotoxic and
proapoptotic stimulus, depending on concentration and incubation time.
This is also true of TNF-. So, if most of the cells were to be
killed or made apoptotic when both stimuli are combined in the
preconditioning protocol, a reduction in protein synthesis by these
cells would not be surprising. Such relatively nonspecific effects,
however, were ruled out: neither LDH release (a sign of necrosis) nor
annexin-V binding (a marker of apoptosis) was elevated beyond the
extent observed after stimulation with TNF- alone (Fig. 6)
. The
sensitivity of the assay seems appropriate, as TNF- did cause a
concentration-dependent increase of annexin-V binding when given alone
(Fig. 6)
.
It could further be argued that the relatively high concentration of
H2O2 (1 mM) might have
caused nonspecific extracellular alterations of cellular structures
(e.g., the TNF- receptor). Thus, to differentiate between possible
extracellular oxidative effects and intracellular actions, the
intracellular scavenger NAC, a precursor of GSH was used. NAC has been
reported to prevent the redox mediated part of TNF- action
(19)
. In the present paper, however, we chose a
concentration and incubation time for NAC that did not influence the
effects of TNF- on ICAM-1, E-selectin, or IL-6 (Fig. 4)
.
Nevertheless, the increase in intracellular GSH levels to ~200% of
control levels found after treatment with NAC (Fig. 5)
demonstrated
that NAC had been taken up and metabolized by the endothelial cells. In
contrast to the indifferent behavior of NAC toward TNF, the
protective effect of transient
H2O2 application was nearly
abolished by pretreatment with NAC (Fig. 4)
. This finding suggests that
the protective effect of preconditioning with
H2O2 is caused by changes
of intracellular redox status.
Indeed, using GSH as a marker of intracellular redox stress showed that
application of H2O2 reduced
GSH levels to ~50% of the control level (Fig. 5)
. This oxidative
stress was transient, as it was detectable 5 min (but no longer 30 min)
after treatment. After pretreatment with NAC, GSH levels also decreased
on incubation with H2O2,
but they did not fall below the concentrations seen in untreated
control cells (Fig. 5)
. Thus, these cells presumably did not experience
critical oxidative stress in the course of the experiment. Accordingly,
it may be concluded that the preconditioning effect in our model is
caused by brief, nonlethal intracellular redox stress.
The classical protocols of ischemic preconditioning (short periods of
ischemia immediately followed by a more severe ischemic insult) have
revealed the transient nature of this phenomenon. Some years ago,
however, an additional time course of myocardial protection had been
detected: ~24 h after the preconditioning process, there is a second
phase of myocardial protection, which has been termed second window of
protection (18)
. This kind of preconditioning has been
ascribed to the delayed induction of cardioprotective mechanisms, like
production of heat shock proteins (7)
. In this case, too,
oxidative stress has been postulated as a key stimulus
(26)
. However, when we extended our observations to a
later time point, no influence on the inflammatory reactions of the
endothelial cells was seen when 24 h lay between the applications
of hydrogen peroxide and TNF- (Figs. 2
and 3)
. The phenomenon of the
second window of protection thus seems to be based on different
mechanisms or it may not develop in HUVECs.
Most inflammatory parameters measured by us on HUVECs have been
reported to be regulated by the transcription factor NFB. Upon
stimulation, NFB is translocated from the cytosol into the nucleus.
This process is enabled by previous phosphorylation, ubiquitination,
and proteasomal degradation of the inhibiting factor IB. Since
translocation of NFB has also been shown to be redox sensitive
(19)
and to occur early after cardiac ischemia and
reperfusion (12
, 35)
, we investigated whether alterations
of NFB-mediated signaling might be causal for the reduced
inflammatory reactions of HUVECs subsequent to preconditioning by
oxidative stress. The translocation of NFB into the nuclear fraction
differed between treatments: whereas
H2O2 and TNF- alone both
caused translocation, sequential treatment (TNF- after
preconditioning with H2O2)
was not additive and instead completely abolished the effects of the
single stimuli (Figs. 7A
, B
and Table 2
). The fact that
P-selectin, an adhesion molecule that is not regulated by NFB in
human cells (34)
—in contrast to the NFB-dependent
parameters ICAM-1, E-selectin, IL-6, and IL-8—did not respond to the
preconditioning protocol (Table 1)
further strengthens the hypothesis
that NFB is a selective target for the inhibitory effects of brief
oxidative stress.
The identical decrease of IB after treatment of HUVECs with either
H2O2 alone, TNF- alone,
or TNF- after H2O2 (Fig. 7B
and Table 2
) suggests that the inhibition of
translocation in the latter intervention group is not due to a
different degree of phosphorylation of IB (Table 2)
. Alternative
explanations for the decreased level of nuclear NFB could be an
infringed transport of NFB to and into the nucleus, altered
permeability of nuclear pores for NFB, or a lower binding affinity
between NFB and the DNA. To our knowledge, the first two mechanisms
have not been validated for NFB. In contrast, a reduction of NFB
affinity toward its binding sites on DNA due to oxidative stress has
been reported (36)
. However, in our model, no differences
were observed in global oxidative stress (mirrored as GSH
concentration) between treatment with
H2O2 alone and with TNF-
after H2O2. This was the
case both after 5 min or 30 min (Fig. 5)
, i.e., the time, when NFB
translocation was measured. Despite the absence of differences in
global redox status between these two groups, it cannot be excluded
that the redox state of specific thiols, which modulate TNF- effects
(17)
, might have been affected for this duration. A
partial explanation of the anti-inflammatory actions of brief oxidative
stress could be that NFB down-regulates its own expression via a
negative feedback loop (37)
. Indeed, we have found that
overall content of NFB in HUVECs pretreated with
H2O2 (1 mM, 5 min) was
reduced to ~60% of control cells 4 h after the treatment (data
not shown). However, this change in NFB expression cannot account
for the rapid infringement of NFB translocation 30 min after
preconditioning.
The presence of NFB in the nucleus is important but far from
sufficient for the expression of ICAM-1, E-selectin, IL-6, and IL-8.
Our data show, for instance, that redox stress in form of a brief pulse
of H2O2 can cause
translocation of NFB without subsequent up-regulation of adhesion
molecules or cytokines (Table 2
; Figs. 2
and 3
). TNF-, in contrast,
elicited NFB translocation and higher expression levels of the
proinflammatory parameters, but without having induced oxidative
stress. Thus, other (e.g., posttranscriptional) influences of the
preconditioning protocol on inflammatory responses might occur in
addition to the reduction of NFB translocation. One feasible
modulator of inflammatory reactions in this context might be nitric
oxide, which on the one hand has been demonstrated to be involved in
preconditioning in some models (25)
, and on the other has
elicited protective actions during reperfusion (38)
.
In conclusion, we have shown for the first time that endothelial cells
can be preconditioned by transient intracellular redox stress to reduce
responsiveness to an inflammatory mediator. The finding that this
phenomenon is related to infringed translocation of the transcription
factor NFB may enable novel approaches directed toward alleviating
reperfusion damage. Though the underlying mechanisms are not resolved,
endothelial preconditioning may well account for the anti-adhesive
effects of ischemic preconditioning (10
, 30)
and its
protective role in the microvasculature (27
, 28
, 29)
in vivo. In a wider context, the intriguing finding that
sequential stimuli can blunt each other with respect to translocation
of NFB might aid the further clarification of mechanisms involved in
redox signaling.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
contributes to ischemia- and reperfusion-induced endothelial activation in isolated hearts. Circ. Res. 84,392-400- and interleukin-1-induced endothelial E-selectin expression by thiol-modifying agents. Arterioscler. Thromb. Vasc. Biol. 18,1829-1837B transcription factor and HIV-1. EMBO J 10,2247-2258[Medline]
B during ischemia in perfused rat heart. Am. J. Physiol. 276,H543-H552 B. J. Biol. Chem. 268,11380-11388B2 gene. Nucleic Acids Res 23,2328-2336This article has been cited by other articles:
|
|
 |
|
  T. P. Obrenovitch Molecular Physiology of Preconditioning-Induced Brain Tolerance to Ischemia Physiol Rev, January 1, 2008; 88(1): 211 - 247. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Zhang, T. S. Park, and J. M. Gidday Hypoxic preconditioning protects human brain endothelium from ischemic apoptosis by Akt-dependent survivin activation Am J Physiol Heart Circ Physiol, June 1, 2007; 292(6): H2573 - H2581. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Postea, F. Krotz, A. Henger, C. Keller, and N. Weiss Stereospecific and Redox-Sensitive Increase in Monocyte Adhesion to Endothelial Cells by Homocysteine Arterioscler. Thromb. Vasc. Biol., March 1, 2006; 26(3): 508 - 513. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. A. Zager, A. C. M. Johnson, S. Y. Hanson, and S. Lund Ischemic proximal tubular injury primes mice to endotoxin-induced TNF-{alpha} generation and systemic release Am J Physiol Renal Physiol, August 1, 2005; 289(2): F289 - F297. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. C. Schafer, D. N. Sehrt, M. Kamler, H. Jakob, and H.-A. Lehr Paradoxical attenuation of leukocyte rolling in response to ischemia- reperfusion and extracorporeal blood circulation in inflamed tissue Am J Physiol Heart Circ Physiol, July 1, 2005; 289(1): H330 - H335. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Carrasco-Martin, S. Alonso-Orgaz, J. C De la Pinta, M. Marques, C. Macaya, A. Barrientos, M. M Gonzalez, A. Garcia-Mendez, P. J. Mateos-Caceres, J. C Porres, et al. Endothelial hypoxic preconditioning in rat hypoxic isolated aortic segments Exp Physiol, July 1, 2005; 90(4): 557 - 569. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Tang, Q. Hao, S. A. Rutherford, B. Low, and Z. J. Zhao Inactivation of Src Family Tyrosine Kinases by Reactive Oxygen Species in Vivo J. Biol. Chem., June 24, 2005; 280(25): 23918 - 23925. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Furst, C. Brueckl, W. M. Kuebler, S. Zahler, F. Krotz, A. Gorlach, A. M. Vollmar, and A. K. Kiemer Atrial Natriuretic Peptide Induces Mitogen-Activated Protein Kinase Phosphatase-1 in Human Endothelial Cells via Rac1 and NAD(P)H Oxidase/Nox2-Activation Circ. Res., January 7, 2005; 96(1): 43 - 53. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. P. Sorescu, M. Sykes, D. Weiss, M. O. Platt, A. Saha, J. Hwang, N. Boyd, Y. C. Boo, J. D. Vega, W. R. Taylor, et al. Bone Morphogenic Protein 4 Produced in Endothelial Cells by Oscillatory Shear Stress Stimulates an Inflammatory Response J. Biol. Chem., August 15, 2003; 278(33): 31128 - 31135. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z.-Q. Zhao, J. S. Corvera, M. E. Halkos, F. Kerendi, N.-P. Wang, R. A. Guyton, and J. Vinten-Johansen Inhibition of myocardial injury by ischemic postconditioning during reperfusion: comparison with ischemic preconditioning Am J Physiol Heart Circ Physiol, August 1, 2003; 285(2): H579 - H588. [Abstract] [Full Text] [PDF]
|
|
