| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
signaling events leading to tissue factor up-regulation via EGR-1 in endothelial cells
* Department of Vascular Biology and Thrombosis Research, VIRCC, University of Vienna, A-1235 Vienna, Austria; and
Department of Molecular Cell Biology, Max-Planck-Institute of Physiological and Clinical Research, D-61231 Bad Nauheim, Germany
1Correspondence: Department of Vascular Biology and Thrombosis Research, VIRCC, University of Vienna, Brunnerstrasse 59, A-1235 Vienna, Austria. E-mail: erhard.hofer{at}univie.ac.at
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(TNF-)
as well as by the main angiogenic factor VEGF. Since both stimuli
induce the transcription factor EGR-1, which is critically involved in
TF gene regulation, we used EGR-1-dependent TF induction as a model to
identify potential cross-talks between the various signal transduction
cascades initiated by VEGF and TNF-. The data show that at the MAP
kinase level, VEGF mainly activates ERK1/2 and p38 MAP kinases in human
endothelial cells. TNF- is able to activate all three MAP kinase
cascades as well as the classical inflammatory I
B/NFB pathway.
Furthermore, the MEK/ERK module of MAP kinases appears to act as the
convergence point of VEGF- and TNF--initiated signaling cascades,
which lead to the activation of EGR-1 and subsequent TF expression,
whereas the upstream signals are distinct. We found that induction of
TF by VEGF via EGR-1 is strongly PKC dependent. The TNF--initiated
MEK/ERK cascade connected to EGR-1 and TF expression is clearly less
sensitive to PKC inhibition. TNF--mediated activation of MEK/ERK and
EGR-1 can be blocked by adenoviral expression of a dominant negative
mutant of IKK2, whereas the VEGF signaling pathway is unaffected. Thus,
our data demonstrate a new link between the classical inflammatory
IKK/IB and the MEK/ERK cascades triggered by TNF-. The additional
finding that EGF induces ERK and EGR-1 in a PKC-independent manner and
that this signal is not sufficient to up-regulate TF emphasizes the
importance of a VEGF-specific signaling pattern for the induction of
TF.—Mechtcheriakova, D., Schabbauer, G., Lucerna, M., Clauss,
M., de Martin, R., Binder, B. R., Hofer, E. Specificity,
diversity, and convergence in VEGF and TNF- signaling events leading
to tissue factor up-regulation via EGR-1 in endothelial cells.
Key Words: endothelium • tissue factor • MAP kinases • inflammatory cytokines • angiogenic growth factors
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Interest in this protein has gained
additional impetus with the identification of novel
properties/functions such as transmission of intracellular signals and
a significant participation in embryonic blood vessel development as
well as tumor-associated angiogenesis (3
4
5
6
7)
. Another
characteristic feature of TF is that its expression and functional
activity can be induced by both the classical inflammatory stimulus
tumor necrosis factor (TNF-) and the main angiogenic growth
factor, vascular endothelial growth factor (VEGF) (8
9
10
11)
.
Unlike growth factor receptors, which have large cytoplasmic domains
with intrinsic protein tyrosine kinase activity (12)
, or
some cytokine receptors, which are directly coupled to JAK tyrosine
kinase (13)
, the TNF- receptors themselves seem to lack
identifiable catalytic activity (14)
. Binding of TNF-
to TNFR1 or TNFR2 induces receptor oligomerization and recruitment of
several downstream adaptor and signaling proteins to their cytoplasmic
domains (15
, 16)
. Nevertheless, both classes of receptors
can activate distinct and overlapping sets of genes and biological
responses in target cells (14
, 17)
. Among the classical
transcription factors activated by TNF-, nuclear factor B
(NFB) plays the major specific role in regulation of inflammatory
response genes (18
, 19)
and does not seem to be activated
by any of the growth factors including VEGF (11)
. NFB
is regulated primarily by phosphorylation of inhibitory proteins, the
IBs, which retain it in the cytoplasm of unstimulated cells
(20)
. Upon cell stimulation, a signaling cascade leads to
phosphorylation of IB via a cytokine-activated protein kinase
complex called IB kinase (IKK) (21
, 22)
. In addition,
TNF- can induce the transcription factor EGR-1 (early growth
response-1), which belongs to the immediate-early (IE) gene products.
Regulation of the egr-1 gene by TNF- has so far been analyzed only
in human fibroblasts (23)
. The precise biological role of
this activation as well as the protein kinases involved in the pathway
have not been described to our knowledge.
On the other side, our recent data demonstrate that EGR-1 is
rapidly induced by VEGF in endothelial cells, thereby leading to TF
up-regulation (11
, 24)
. These findings led to the
intriguing hypothesis that there might be cross-talk between the
induction programs of gene expression responsible for angiogenic and
inflammatory responses of endothelial cells, possibly starting at the
level of activated kinases and leading further to certain transcription
factors, and the spectrum of induced genes. Thus, the activation of TF
via EGR-1 provides a model to 1) reconstitute a detailed
picture of downstream signaling events initiated by VEGF and TNF-,
2) identify the possible interactions between the various
signal transduction components, and 3) link an
immediate-early gene response to a subsequent secondary gene
expression.
Mitogen-activated protein (MAP) kinases are common participants in
signal transduction pathways initiated by growth factors, cytokines,
stress stimuli, and various pharmacological compounds (25
, 26)
. The mammalian MAP kinase family members identified so far
form three major types of MAP kinase modules with a generally high
degree of specificity and functional separation: ERK1/2, SAPK/JNK, and
p38 kinase. Each module contains at least three protein kinases that
work sequentially and can be activated by dual phosphorylation on
tyrosine and threonine residues. How a primary signaling pathway is
selected to achieve the induction of a certain immediate-early gene by
a given ligand is far from being understood.
In the present study, we have focused on those of the multiple
signaling cascades triggered by VEGF and TNF- receptors that are
linked to EGR-1 activation and followed by TF expression on endothelial
cells. We used an approach with small-molecule inhibitors in
combination with inactive mutants of kinases for the selective blockade
of signal transduction pathways. We could demonstrate that induction of
TF by VEGF via EGR-1 is protein kinase C (PKC) dependent and mediated
directly by ERK. Furthermore, our data suggest that the MEK/ERK module
of MAP kinases functions as the convergence point of VEGF and TNF-
signaling that leads to the activation of EGR-1, whereas the upstream
signals are distinct. We discovered a new link between the classical
inflammatory IKK/IB and the MEK/ERK cascades induced by TNF-.
Although the MEK/ERK cascade can be activated by a broad spectrum of
growth factors, in contrast to VEGF, epidermal growth factor (EGF) does
not trigger TF induction. Our data show that EGF initiates a pathway
from MEK/ERK to EGR-1 in a PKC-independent manner, which suggests
specificity in VEGF vs. EGF signaling.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Recombinant
human EGF was purchased from PromoCell (Heidelberg, Germany), TNF-
from Genzyme Inc. (Cambridge, Mass.), phorbol 12-myristate 13-acetate
(PMA), wortmannin, citrated plasma, and thromboplastin from Sigma
Chemicals (St. Louis, Mo.). The MEK inhibitor PD098059, the p38 kinase
inhibitor SB203580 and the PKC inhibitor bis-indolylmaleimide I
(synonyms: GF109203X and Gö6850) were obtained from Calbiochem
(La Jolla, Calif.). Phospho-specific MEK1/2, ERK 1/2, p38 kinase
polyclonal antibodies, nonphospho ERK1/2 antibodies, and LumiGLO
chemiluminescent reagent were from New England BioLabs (Beverly,
Mass.); polyclonal EGR-1 antibody, NFB p65, and IB were from
Santa Cruz Biotechnology (Santa Cruz, Calif.). Polyclonal TF antibodies
were a gift from Dr. W. Ruf (Scripps Research Institute, La Jolla,
Calif.). Peroxidase-conjugated donkey anti-rabbit immunoglobulin G
(IgG) and sheep anti-mouse IgG were purchased from Amersham Life
Science (Amersham Place, England), fluorescein isothiocyanate-labeled
goat anti-rabbit IgG from Accurate Scientific (Westbury, N.Y.).
Immobilon-P transfer membranes were products of Millipore (Bedford,
Mass.).
Clotting assay
Cells were seeded in 6-well plates at 80–90% confluency and
grown overnight. Cells were scraped from the plates and analyzed for TF
activity as described (10
, 28)
. Briefly, after induction
for 4 h with VEGF, TNF-, PMA, or EGF, cells were washed twice
and scraped in 1 ml clotting buffer (12 mM sodium acetate, 7 mM
diethylbarbitate, and 130 mM sodium chloride; pH 7.4); 100 µl of
resuspended cells was mixed with 100 µl of citrated plasma and
clotting times were measured after recalcification with 100 µl of 20
mM CaCl2 solution at 37°C. TF equivalents were
determined by using a standard curve obtained from rabbit brain
thromboplastin.
Western blot analysis
After various treatments, the cells were washed twice with cold
phosphate-buffered saline (PBS), lysed in 100 µl of Laemmli buffer,
scraped, and heated for 5 min at 95°C. Total cell lysates were
separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis
(SDS-PAGE) and transferred to Immobilon-P membrane. The membrane was
blocked for 30 min with PBS containing 0.1% Tween-20 and 3% skim milk
and incubated for 1 h at room temperature with a primary antibody
diluted in blocking buffer. Then the membrane was washed three times
for 5 min with PBS containing 0.1% Tween-20 and incubated with
peroxidase-conjugated secondary antibody for 1 h at room
temperature. After a washing step, the membrane was incubated for 1 min
with ECL reagent and exposed to film as required. For reprobing with
another antibody, the membrane was washed twice in PBS, stripped for 30
min at 55°C with stripping buffer (62.5 mM Tris-HCL, pH 6.8, 2% SDS,
100 mM 2-mercaptoethanol), and washed three times for 5 min with PBS at
room temperature. The membrane was stored wet wrapped in Saran Wrap at
4°C after each immunodetection.
Recombinant adenoviral constructs and infection
HUVEC grown in 6-well plates were washed twice with PBS and
incubated at a multiplicity of infection of 100 with recombinant
adenovirus that express dominant negative mutant of IKK2 (K44A)
(AdIKK2dn) (W. Oitzinger et al., unpublished results) or control
adenovirus (AdGFP) (29)
in PBS for 30 min. Thereafter,
cells were washed with PBS and then cultured in normal HUVEC medium.
Twenty-four hours postinfection, cells were short-starved in 1% SCS
medium for 5 h, followed by appropriate stimulation. For
immunofluorescence assay, cells were grown and infected in LabTek
tissue culture chamber slides.
Immunofluorescence
Immunofluorescence assay was performed mainly as described
previously (30)
. Briefly, cells were grown in LabTek
tissue culture chamber slides for at least 24 h prior to fixation.
After appropriate stimulation, cells were washed twice with PBS, fixed
for 10 min at room temperature with 3.7% formaldehyde, 2% sucrose in
PBS, and permeabilized for 5 min with 0.5% Triton X-100 in PBS.
Primary antibodies were diluted in PBS, 1% bovine serum albumin and
incubated with the cells for 1 h at room temperature. Cells were
washed in PBS and incubated with Texas Red-labeled goat anti-rabbit IgG
for 1 h at room temperature. To visualize cell nuclei, a
fluorescent groove binding probe for DNA, 4',6-diamidino-2-phenylindole
(DAPI; Sigma), was added to the secondary antibody solution at 100
ng/ml. Slides were washed in PBS and mounted with mounting fluid. The
immunofluorescence results were analyzed with a Bio-Rad MRC 600
confocal laser scanning microscope.
Statistical analysis
The results obtained were analyzed by one-way analysis of
variance and the Student’s paired t test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
in
endothelial cells
. To examine whether
EGR-1 can also be induced in endothelial cells by the inflammatory
stimulus TNF-, HUVEC were treated for various time periods with
either TNF- (100 U/ml) or VEGF (1.25 nM). As shown previously, at
these concentrations the two agents caused maximal induction of TF
procoagulant activity when monitored by a one-stage clotting assay
(10
, 11)
. Cell lysates were prepared and subjected to
Western blot analysis. A low level of endogenous EGR-1 was detectable
in short-starved unstimulated endothelial cells, whereas a transient
induction of EGR-1 was observed upon treatment of these cells with VEGF
as well as TNF- (Fig. 1
). The maximum induction of EGR-1 expression by both agents was at 
60
min after treatment of the cells and dropped to basal levels within
4 h. To determine the level of TF up-regulation in these cells,
the membranes were reprobed with anti-TF antibodies. TF was almost
undetectable in unstimulated HUVEC and, as expected, was strongly
induced by VEGF and TNF-, reaching maximal expression at 4 h.
Thus, our data clearly demonstrate that EGR-1 expression is induced in
endothelial cells by the inflammatory stimulus TNF- to an extent
comparable to the angiogenic growth factor VEGF.
|
VEGF- and TNF--induced MAP kinase ‘signature’ in endothelial
cells
Next we tested whether there is a convergence point of VEGF- and
TNF--initiated signaling pathways leading to induction of EGR-1.
Stimulation of all three MAP kinase cascades by VEGF and TNF- in
human endothelial cells was measured under the same conditions and
compared (Fig. 2
). Since members of the MAP kinase families are known to be activated by
dual phosphorylation on Tyr/Thr (25)
, we examined kinase
activation using a Western blot technique with phospho-specific
antibodies. These reagents selectively recognize only active forms of
the kinases and do not cross-react with other related family members.
Compared to unstimulated cells, VEGF strongly activated ERK1/2, with a
peak at 10 min and values still above the basal level by 60 min (Fig. 2)
. The much weaker and more transient activation of p38 kinase
observed after VEGF treatment returned to baseline by 20 min, whereas
no activation of JNK by VEGF was detectable in HUVEC. In agreement with
our previous data (11)
, IB levels were not affected by
VEGF treatment. Antibodies to the nonphosphorylated form of ERK1/2 were
used to monitor sample loading. In the case of TNF--treated cells,
activation of ERK1/2 was transient, with a maximum at 10–20 min and
decreasing to the baseline at 30 min, whereas strong activation of p38
kinase was sustained up to 60 min. Activation of JNK was transient,
with maximum levels detected at 10 min. As expected, TNF- treatment
of cells resulted in rapid degradation of IB within 10 min, followed
by resynthesis of the protein starting after 30 min. These results show
that in endothelial cells, VEGF mainly activates ERK1/2 and p38, but no
detectable levels of JNK. TNF- is able to activate all three MAP
kinase pathways in addition to the cascade leading to IB
degradation.
|
A specific inhibitor of MEK blocks VEGF- and TNF--induced EGR-1
expression
To determine which of the MAP kinase cascades, if any, mediates
the induction of EGR-1 by VEGF and TNF-, a specific inhibitor of
MEK1/2 activity, PD098059 (31)
, and a specific inhibitor
of p38 kinase, SB203580 (32)
, were used. Cells were
preincubated with the inhibitors and then induced for 10 min (for
maximal kinase activation) or 60 min (for maximal EGR-1 expression)
(Fig. 3
). The activation of ERK1/2 by VEGF and TNF- was prevented by
PD098059, but not by SB203580, as predicted, since to date ERK1/2 are
the only known substrates for MEK1/2. Furthermore, induction of EGR-1
by both stimuli was also inhibited by PD098059. In contrast, the p38
kinase inhibitor did not block VEGF- or TNF--induced EGR-1
expression. Moreover, in the presence of SB203580 the activation of
ERK1/2 by VEGF and TNF- was slightly enhanced, showed prolonged
kinetics, and was followed by increased expression of EGR-1. Thus, the
data obtained suggest that the MEK/ERK cascade is required for EGR-1
induction in response to both VEGF and TNF-. At the same time,
reprobing with anti-IB antibodies clearly demonstrated that the MEK
inhibitor PD098059 had no effect on TNF--induced IB degradation
(Fig. 3
, IB) and p65 nuclear translocation analyzed by
immunofluorescence (data not shown). These data indicate that PD098059
selectively blocks the ERK pathway triggered by VEGF or TNF-, but is
not inhibitory in general to other kinase cascades induced by TNF-
such as the NIK/IKK or p38 kinase cascades. Furthermore, these data
linked the MEK/ERK module of MAP kinases to EGR-1 induction whether
VEGF or TNF- initiated this cascade.
|
VEGF and TNF- regulate TF activity by mechanisms involving EGR-1
in a PD098059-sensitive fashion
As we reported earlier (11
, 24)
, EGR-1 plays a
significant role in TF activation by VEGF. In the case of inflammatory
stimuli, induction of the TF promoter was previously analyzed in
monocytic cells and porcine aortic endothelial cells and found to be
mediated mainly by NFB in combination with AP-1 (8
, 10
, 24
, 33)
. At the same time, it was not determined whether
TNF--induced EGR-1 plays a significant role in TF up-regulation in
human endothelial cells. Based on the above results, we next tested to
what extent inhibition of EGR-1 expression by PD098059 would lead to TF
inhibition. Cells were preincubated with the inhibitor and then induced
by TNF- or VEGF for 4 h. Stimulated TF activity was measured in
a one-stage clotting assay and TF expression was analyzed in parallel
by Western blots (Fig. 4
). VEGF-induced TF activity was strongly suppressed in a dose-dependent
manner by the MEK inhibitor PD098059, showing 80% inhibition at 25
µM (Fig. 4)
. TNF--stimulated TF activity was clearly less
affected, but was still reduced by 50–60% at 25 µM. These results
were confirmed at the level of protein expression by Western blot
analysis (Fig. 4
, insert), illustrating good correlation between the
data obtained by both methods. To test whether the residual activities
detected were in part due to an incomplete inhibitory potency of
PD098059, another potent MEK inhibitor, U0126 (34)
, was
used at 50 µM and gave similar results (data not shown). These data
show that induction of the MEK/ERK/EGR-1 cascade is required to a large
degree for TF activation by VEGF, but also plays a significant role in
TNF--triggered TF.
|
Effect of a PKC inhibitor on EGR-1 and TF induced by PMA, VEGF, or
TNF-
The mechanisms involved in the activation events for MEK and ERK
have been studied extensively. In both cases, two phosphorylations
within the activation loop of the kinases are required (35
, 36)
. In contrast to these activation mechanisms, the regulation
of Raf, the immediate upstream kinase of MEK, is substantially more
complex. Among the mechanisms involved, both PKC-dependent and
-independent pathways of Raf activation in response to different
agonists may occur (37)
. We used a broad spectrum
inhibitor of PKC isozymes, bis-indolylmaleimide I (also known as
GF109203X or Gö6850), to assay the contribution of PKC in the
signaling chain leading to EGR-1 activation by VEGF and TNF-.
Bis-indolylmaleimide I is known to inhibit membrane and cytosolic PKC
with similar potencies. IC50 values of inhibition
increase from conventional to novel and atypical PKC isoforms
(38
, 39)
.
Since the biological properties of phorbol esters are based on their
ability to mimic the responses of diacylglycerol, allowing direct
activation of PKC (40)
, we used PMA as a
positive control to activate the PKC-dependent MEK/ERK module. The data
in Fig. 5A
show that PMA treatment resulted in strong induction of
EGR-1. As expected, the PKC inhibitor at 5 µM inhibited over 90% of
EGR-1 induction, and at 20 µM the expression of EGR-1 was blocked
completely (data not shown). Among the properties of phorbol esters is
the PKC-mediated activation of NFB (41)
. As shown in
Fig. 5A
, PMA treatment resulted in IB degradation, but
with slower kinetics when compared to TNF-. Moreover,
bis-indolylmaleimide I at 5 µM completely abolished PMA-induced IB
degradation, thus clearly demonstrating the potency of the inhibitor
used.
|
VEGF-initiated signaling events leading to EGR-1 were very similar to
PMA with respect to sensitivity to the PKC inhibitor. Pretreatment of
the cells with bis-indolylmaleimide I before VEGF stimulation repressed
expression of EGR-1 (Fig. 5A
). Taken together, our findings
indicate that induction of EGR-1 by VEGF is mediated by a signaling
cascade involving PKC and ERK.
TNF--induced EGR-1 expression was not altered or only slightly
affected in some experiments by 5 µM of bis-indolylmaleimide I (Fig. 5A
) and was partially inhibited by 20 µM of the compound
(data not shown). Thus, the TNF- signaling chain leading to EGR-1
induction is less sensitive to alterations in PKC activity than the
cascades induced by PMA or VEGF. bis-Indolylmaleimide I did not inhibit
TNF--mediated activation of p38 kinase (data not shown) and, in
contrast to the results with PMA described above, did not alter
TNF--initiated IB degradation (Fig. 5A
). This
illustrates the specific action of the PKC inhibitor as well as the
important differences in the multiple signaling pathways leading to
NFB activation.
Our next step was to evaluate whether the alterations in signaling
observed in the presence of the PKC inhibitor would affect induction of
TF activity. From three stimuli used, PMA caused the strongest
up-regulation of TF activity whereas VEGF and TNF- induced TF to a
comparable extent (Fig. 5B
). Bis-indolylmaleimide I caused a
dose-dependent inhibition of TF activity induced by all three stimuli.
In PMA- and VEGF-stimulated cells, the level of TF activity was
repressed over 90% by 5 µM bis-indolylmaleimide I, whereas only 20%
inhibition was observed in TNF--stimulated cells. Thus, the
TNF--initiated signaling cascade connected to TF expression was
clearly less sensitive to the PKC inhibitor.
A dominant negative IB kinase 2 (IKK2) mutant selectively blocks
the induction of EGR-1 by TNF-
Our findings clearly show that TNF--induced EGR-1 expression in
endothelial cells is MEK/ERK dependent. On the other hand, it is well
established that induction of the major specific transcriptional
programs by TNF- is primarily mediated by the activation of NFB
via the NIK/IKK pathway. To test whether this NFB-activating cascade
interferes with a MAP kinase-mediated induction of EGR-1, we used a
recombinant adenovirus expressing a dominant negative IKK2 mutant
(AdIKK2dn). This mutant has been described to block TNF-- and
interleukin 1 (IL-1) -induced NFB activation (21
, 42)
.
In parallel, a GFP-expressing virus (AdGFP) was used as a negative
control. HUVEC infected with virus were stimulated with VEGF or TNF-
for 1 h. Immunofluorescence analysis showed that the
IKK2 mutant, as expected, inhibited p65 nuclear
translocation in TNF--stimulated cells and had no effect
on the subcellular localization of p65 in unstimulated
HUVEC (Fig. 6A
). We have previously analyzed the subcellular
localization and expression of EGR-1 protein in unstimulated and
VEGF-induced endothelial cells and demonstrated a rapid accumulation of
EGR-1 in the nucleus on induction (11)
. HUVEC
infected with control AdGFP showed a diffuse cytoplasmic staining for
EGR-1 in unstimulated cells and predominant nuclear localization on
stimulation with VEGF or TNF- (Fig. 6B
). The IKK2 mutant
blocked TNF--induced nuclear accumulation of EGR-1. In contrast,
VEGF-triggered EGR-1 induction was almost unaffected (Fig. 6B
).
|
We next tested whether the dominant negative IKK2 mutant interferes
with the TNF--initiated activation of the MEK/ERK cascade. Cells
were infected with AdIKK2dn or AdGFP, stimulated by VEGF or TNF- for
10 and 30 min, and subjected to Western blot analysis. Staining with
phospho-specific MAP kinase antibodies showed that the activation of
MEK, ERK, and p38 kinase typically observed in VEGF- or TNF--treated
HUVEC (compare Fig. 2
) was unaltered in the AdGFP-infected cells
(Fig. 7
). In marked contrast, TNF--induced MEK as well as ERK activation was
inhibited by AdIKK2dn, whereas activation of the p38 kinase was not
affected. In comparison, VEGF-initiated signals were not altered
significantly. Reprobing with anti-IB antibodies was used to
demonstrate the inhibitory potency of the dominant negative IKK2 mutant
on the NFB pathway. Indeed, overexpression of the IKK2 mutant
completely prevented TNF--induced IB degradation (Fig. 7
; compare
10 min time points for AdGFP and AdIKK2dn). Taken together, these data
indicate that the IKK2 mutant interferes specifically with the
TNF--induced pathway connected to NFB activation as well as the
TNF--induced MEK/ERK cascade leading to EGR-1 induction. These
results lend further support to our hypothesis that from the multiple
pathways triggered by TNF- and VEGF receptors, those that lead to
EGR-1 induction converge on the level of MEK, whereas the initial
upstream events are distinct.
|
TF is activated by VEGF but not EGF in endothelial cells
Since the MEK/ERK module of MAP kinases is triggered by a wide
range of growth factors, one would expect that TF expression should be
generally induced by all growth factors for which receptors are present
on endothelial cells. However, in contrast to VEGF, treatment of
endothelial cells with EGF did not increase TF expression and
procoagulant activity, whereas EGR-1 was strongly induced (Fig. 8A
, insert). On the other hand, though by itself noninducing,
EGF was able to enhance TF activity induced by VEGF and TNF- (Fig. 8A
). These data suggest that EGF does not display an
inhibitory potency, but rather an important step in the signaling
cascade necessary for TF induction is missing. Therefore, we compared
the signaling events triggered by VEGF and EGF. The EGF-induced pattern
of the different MAP kinase modules was similar to VEGF and the ERK
cascade was predominantly induced. Furthermore, addition of EGF to
cells led to a sustained high level of phospho-ERK; thus, no principal
differences were detected in the kinetics of ERK activation by VEGF and
EGF. In addition, EGF stimulation of EGR-1 could be blocked by the MEK
inhibitor PD98059 but not by the p38 kinase inhibitor (data not shown),
similar to the effects seen in VEGF-treated cells. Despite of these
clear similarities in signaling events induced by these growth factors,
one principal difference was observed: the PKC inhibitor did not
prevent EGR-1 activation by EGF (Fig. 8B
), which suggests
that the signaling pathway leading from the EGF receptor to EGR-1 is
MEK/ERK mediated but, in contrast to VEGF, PKC independent.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In this study, we have used the induction of EGR-1 and subsequent TF
expression in endothelial cells to 1) characterize the MAP
kinase profiles induced by VEGF and TNF- and 2) define
common routes that mediate similar cellular responses to these agents.
The findings of the present study demonstrate that VEGF activates
mainly ERK1/2 and p38 kinase, but no or only trace amounts of JNK.
TNF- was able to strongly activate all three MAP kinases in addition
to the classical inflammatory cascade leading to IB degradation and
NFB activation. The ERK cascade plays the major role in controlling
EGR-1 by both agents.
In principle, activation of the egr-1 gene, similar to other IE genes,
is a complex process and can receive input from several pathways.
Different MAP kinase subtypes may lead to phosphorylation and
subsequent activation of transcription factors, such as Elk-1 and the
related ternary complex factor, serum response factor, and CREB. From
the current work, it is now clear that a single MAP kinase cascade,
MEK/ERK, is absolutely required for EGR-1 induction by the inflammatory
TNF- and the angiogenic VEGF. This is in contrast to stress-induced
EGR-1, which has been shown to be mediated by p38 kinase and JNK
(43)
.
Furthermore, it appears that VEGF and TNF- signaling converges at
the level of MEK, whereas initial upstream events leading to MEK
activation are distinct. In the case of TNF- signaling, a new link
was identified between IKK2 and MEK activation. When overexpressed, a
dominant negative IKK2 mutant, as expected, inhibited TNF--initiated
IB degradation and p65 nuclear translocation, confirming the
critical contribution of IKK2 to NFB activation. The kinase inactive
mutant of IKK2 also blocked TNF--triggered MEK
phosphorylation/activation and EGR-1 induction. One possible
explanation would be that the two pathways use common upstream signal
transduction component(s). Several lines of evidence suggest that MEKK1
(MAPK kinase kinase 1) could be an appropriate candidate for
coordinated regulation of both NFB and MAP kinase pathways. MEKK1
was shown to be able to induce IB phosphorylation/NFB activation
and the JNK pathway (44
, 45)
. At the same time, previous
studies clearly indicate that MEKK1 can regulate two different MAP
kinase modules: ERK and JNK (46
, 47)
. However, these data
do not explain why a dominant negative IKK2 mutant would block the
activity of an upstream kinase for MEK such as MEKK1. Recently,
researchers described IKK as one of the essential kinases in the
multiprotein signaling complex, called signalsome, that regulates
NFB activation in response to proinflammatory cytokines. MEKK1 has
also been identified to be a component of the same signalsome
(21)
. Thus, it seems possible that incorporating an
inactive kinase in the complex may potentially lead to an inactive
state of the signalsome as a whole. Indeed, data demonstrating that
dominant negative MEKK1 inhibited TNF--induced NFB activation
(44
, 48)
would be consistent with this interpretation.
However, the potential contribution of MEKK1 to the activation of
MEK/ERK by TNF- reflects only part of the possibilities, since a
number of different stimuli can use alternative pathways of activation,
such as Raf family members, which also regulate MEK/ERK
(49)
.
The results of the present study show that in contrast to the
TNF--mediated MEK/ERK/EGR-1 cascade, the VEGF-initiated
MEK/ERK/EGR-1 pathway was not significantly altered by the dominant
negative mutant of IKK2. Selective or preferential activation of either
MEKK1 or Raf by the two different stimuli is therefore one possible
interpretation of these data. There exists an alternative, nonexclusive
explanation for the TNF- specific effect of the IKK2 mutant. Like
MEK, IKK contains a canonical MAP kinase kinase (MAPKK) activation loop
motif (Ser-X-X-X-Ser, where X is any amino acid:
SLCTSFVGT for IKK2 vs.
SMANSFVGT for MEK). Thus, according to the
current model of sequential kinase cascade organization, MEK and IKK
belong to the same level of hierarchy within the MAP kinase families.
It is tempting to speculate that a hypothetical, not-yet-identified
scaffold protein could coordinate these two signal transducers in
signalsomes triggered by TNF-. Furthermore, the fact that the
catalytically inactive mutant of IKK2 did not block MEK/ERK activation
triggered by VEGF may point to the existence of signalsome complexes
specific for the individual receptors and containing signal-determined
components.
The demonstrated activation of common cascades by TNF- and
VEGF could account for the ability to elicit in part similar responses
from their target cells in certain instances. One example of common
responses of these factors based on MEK/ERK activation in endothelial
cells is the induction of TF. Among additional described common
bioactivities elicited by TNF- and VEGF are the release of soluble
tie-1 from endothelial cells (50)
and the ability to
increase vascular permeability (28
, 51
, 52)
. The exact
pathways leading to these responses have not been defined yet.
EGR-1 and TF induction: dependence on PKC
The transcriptional induction of EGR-1 has previously been
described to be mediated by PKC-dependent and -independent mechanisms.
In fact, the PKC pathway can connect extracellular stimuli to EGR-1
induction as it can be shown that direct activation of PKCs by PMA
induces EGR-1 (53)
(see Fig. 5
). A PKC-dependent EGR-1
induction has been shown in cells activated by hypoxia
(54)
. On the other hand, pathways independent of PKC can
also play a role in EGR-1 induction. In fibroblasts deficient in PKC
signaling, EGR-1 was strongly induced by serum and EGF
(53)
. The results presented in this study demonstrate
that, in the case of VEGF stimulation of endothelial cells, PKC
activation is an essential step in the chain of events connecting the
VEGF receptor (VEGFR) with TF activation via EGR-1. The fact that, in
comparison to the other stimuli used, PMA treatment of the cells
resulted in the strongest induction of TF further emphasizes that PKC
can play an important role for TF expression. In line with the findings
concerning TF expression inhibitors of PKC have been reported to
inhibit angiogenesis induced by VEGF (55)
. Taking into
consideration the fact that the amino acid sequence of EGR-1 includes
several potential PKC phosphorylation sites (D. Mechtcheriakova et al.,
unpublished data), it is tempting to speculate that PKC is involved not
only in the activation of MEK/ERK/EGR cascade, but also in additional
direct phosphorylation of EGR-1 to modulate and/or specify its
transcriptional activity.
In contrast to PMA or VEGF, the TNF--initiated signaling
cascade connected to TF expression was clearly less sensitive to the
PKC inhibitor. This may be due to only a partial involvement of this
pathway or a difference in PKC isozymes. Indeed, previous studies have
demonstrated that TNF- is a potent activator of the atypical PKCs
(56)
. The important contribution of these isoforms in the
activation of the MEK/ERK pathway (57)
as well as in an
NFB activation through the IKK pathway (58)
has been
described. On the other side, in contrast to conventional PKCs, much
higher concentrations of bis-indolylmaleimide I have to be used to
block the activity of atypical PKCs (38)
.
EGR-1 and TF induction: VEGF vs. EGF
In several instances, treatment of a single cell type with
different factors can activate similar signaling pathways but result in
different biological outcomes. In this respect it is an open question
to what extent the VEGF signals are distinct from other growth factors
with growth-promoting activity on endothelial cells and lead to a VEGF
characteristic response pattern. One classical example of differential
responses to growth factors are the responses of PC12 cells to NGF and
EGF. In these cells EGF promotes mitogenesis, but NGF promotes
differentiation (59)
. A possible explanation is that the
initial signals triggered by the stimuli can be diverted to different
outcomes by differences in the strength or kinetics of the signals, or
both (59
, 60)
. EGF induces transient activation of ERK and
stimulates proliferation of PC12 cells, whereas NGF stimulates
prolonged ERK activation and induces cellular differentiation. An
alternative possibility is that all growth factors may induce ERK, but
differences exist in the upstream signals that couple receptor tyrosine
kinases to ERK. Indeed, supporting data for this possibility have been
published recently showing that PKC
is required for the activation
of MEK/ERK by differentiating factors in neuronal cells, but not by the
growth-promoting EGF (61)
.
The data presented here demonstrate a differential response of endothelial cells to EGF and VEGF. In contrast to VEGF, EGF treatment of the cells did not lead to up-regulation of TF, although both growth factors triggered sustained ERK activation resulting in EGR-1 induction. One significant difference is that despite the dominant role of the MEK/ERK cascade in EGF- and VEGF-induced EGR-1, the EGF activation seems to be PKC independent. This suggests that distinct upstream signals initiated by VEGF and EGF lead to MEK activation. We believe that this could result in a qualitative difference of EGF- vs. VEGF-induced EGR-1 such as a differential phosphorylation/dephosphorylation and a different response of the TF gene.
Alternatively, it is possible that a second factor(s) is
important for the full response of the TF promoter that may not be
active in EGF-treated cells. The transcription factor NFAT could be an
appropriate candidate since, in addition to EGR-1 (11)
,
NFAT has also been implicated in transcriptional activation of the TF
gene by VEGF (62)
. It is noteworthy that the NFAT binding
site within the TF promoter overlaps with the previously identified
NFB-like site participating in TF activation in response to
lipopolysaccharides or TNF- (10
, 33)
. Therefore, it is
possible that full transcriptional activity of the TF gene would
require the presence and cooperation of both EGR-1 and NFB for
TNF- responsiveness or EGR-1 and NFAT, occupying the overlapping
NFB/NFAT binding site, for VEGF responsiveness.
EGR-1 and TF induction: correlation with VEGFR-2 activation
VEGF is considered to be a main mediator of angiogenesis. Numerous
proteins closely related in primary structure to VEGF have been grouped
in the VEGF family. The angiogenic activity of various family members
is mediated in turn by several VEGF receptors (63)
. The
VEGF-A165 isoform binds with high affinity to
both VEGFR-1 (flt-1) and VEGFR-2 (flk-1), whereas
VEGF-A121 acts solely via VEGFR-2
(64)
. Recent data demonstrate that VEGF-E, a novel member
of the VEGF family, possesses an angiogenic activity as potent as
VEGF-A. In contrast to VEGF-A, VEGF-E binds selectively to VEGFR-2 but
not to VEGFR-1 (65)
. This finding is in accordance with
previous studies suggesting that the angiogenic signal by VEGF-A itself
may be predominantly mediated by VEGFR-2 (66
, 67)
. These
data indicate that VEGFR-2 activation can mediate angiogenesis
efficiently without concomitant VEGFR-1 activation. In addition,
placental growth factor (PlGF) binds to VEGFR-1, but not to VEGFR-2,
and its role in angiogenesis is currently not well understood. The
different VEGFs also have different capacities to modulate TF activity
on endothelial cells. Both isoforms of VEGF-A,
VEGF165 and VEGF121, were
equipotent to induce TF (11)
. VEGF-E was found to have a
similar bioactivity to VEGF-A with respect to TF induction
(65)
. In contrast, TF was only slightly induced by PlGF
(twofold) in comparison to a significantly higher up-regulation by
similar concentrations of VEGF-A (25- to 80-fold) (9
, 11)
.
These findings apparently reflect clear parallels between VEGFR-2
activation and TF expression and thereby suggest that activation of
VEGFR-2 seems to be sufficient for TF induction in endothelial cells.
The biological relevance of VEGF-induced TF expression in the
complex process of new vessel formation is still unresolved. Since in
the present study we did not investigate this aspect directly, it is
only possible to speculate based on two experimental observations.
First, it appears that there are clear differences in the abilities of
growth factors to induce TF expression: VEGF is a strong stimulus
whereas EGF does not activate TF in endothelial cells. Second,
VEGF-triggered signaling via VEGFR-2 seems to be sufficient to induce
TF. It thus appears that TF induction parallels the property of VEGF as
a strong inducer of angiogenesis via VEGFR-2. It is widely accepted
that active endothelial cell proliferation, migration, and matrix
remodeling are all essential components of angiogenesis, and it is
conceivable that a small amount of fibrin deposition is necessary for
angiogenesis to proceed. For the process of vasculogenesis during
development, one of the initial characteristic steps involves
differentiation of the mesoderm-derived angioblasts into endothelial
cells. It is not completely clear to what extent the proposed
recruitment of endothelial precursor cells from the circulation may
also play a role in angiogenesis in parallel to endothelial cell
migration and sprouting from preexisting mature endothelial cells. It
is possible that the pathway delineated here, which connects MEK/ERK
activation to EGR-1, will also be important for the process of
endothelial cell differentiation. The involvement of this route in
differentiation processes has been described for other cell types. The
sustained ERK activation, followed by EGR-1 induction, which is
required for NGF-induced differentiation of PC12 cells (68
, 69)
, is one example. As with VEGF induction, this process is PKC
dependent (61)
. Other examples are the
granulocyte-macrophage colony-stimulating factor (GM-CSF) inducing the
differentiated functions of mature granulocytes and
monocytes/macrophages (70)
and IL-3 resulting in
proliferation and maturation of early bone marrow progenitor cells
(71)
. Pathways involving MEK/ERK are activated by both
factors (72)
. Furthermore, signals from the GM-CSF and
IL-3 receptors lead to rapid induction of the egr-1 gene
(73)
. Given the close relationship of endothelial cell
precursors with the hemopoietic lineage, a common theme in the
induction of differentiated functions seems possible.
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 19, 2000.
Revision received June 30, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
