 |
INTRODUCTION |
Eukaryotic cells employ multiple mechanisms to ensure accurate
transmission of genetic information between generations. Critical surveillance of this transmission is provided by the DNA damage response, which may arrest the cell cycle to allow damage repair or
direct cells to apoptosis in situations of severe damage (1-4).
Our understanding of how cells sense DNA damage remains incomplete, but
it is clear that two members of the phosphatidylinositol 3-kinase
(PI3K)1 family, ATM
and ATR, are major DNA damage signal transducers (4). Downstream of
ATM/ATR lies p53, a central mediator of the response, which in turn
induces cell cycle arrest by up-regulation of p21CIP1 and
14-3-3
, activates DNA damage repair pathways, and induces apoptosis (5, 6). p53 is normally rapidly degraded by Mdm2-mediated ubiquitin-dependent proteolysis (7, 8). In response to DNA damage, p53 is stabilized by inhibition of this proteolytic process, partly through post-translational phosphorylation. Stabilization of p53
by phosphorylation on residues Ser-15 and Ser-21 by ATM and Chk2
kinase in response to DNA damage is well described (9-11). Members of
the mitogen-activated protein kinase (MAPK) family have also been
demonstrated to phosphorylate and stabilize p53 (12-14). To wit,
activation of JNK leads to p53 phosphorylation, thus interfering with
the association of Mdm2 and p53 (12). Furthermore, activation of
p38 MAPK by UV was shown to phosphorylate p53 on S389 (13,
14).
The third member of the canonical MAPK family, ERK (extracellular
signal-regulated kinase), is centered on multiple signal transduction
pathways to accomplish a variety of functions. Activation of ERK
through different pathways leads to fundamentally different cellular
responses, including proliferation, differentiation, survival, and
memory consolidation (15-19). ERK activation by extracellular growth
signals is mediated through growth factor tyrosine phosphorylation and
activation of a small G protein, Ras, and promotes cell proliferation (20, 21). The importance of ERK in transduction of mitogenic signals is
illustrated by the demonstration that activation of ERK is sufficient
to transform NIH3T3 cells or MEF lacking either p53 or p16 (22, 23). In
some circumstances, ligand interactions with growth factor receptors
may result in ERK-mediated cell cycle exit. Indeed, ERK activation by
nerve growth factor drives PC12 cell differentiation (22, 24).
ERK may function in the response to DNA damage. ERK activation was
observed in response to cisplatin in ovarian cancer cells (25). ERK can
phosphorylate p53 in vitro (25), but its role in
vivo is unclear (25-27). The role of DNA damage-induced ERK activation in mediating apoptosis has also not been clarified (28, 29).
In this regard, however, tumors with constitutive ERK activation
undergo apoptosis when ERK activity is blocked (30).
Consequently, we sought to explore ERK activation in response to DNA
damage, and to outline the mechanisms and consequences of such
activation. We report here that multiple DNA damage stimuli, including
etoposide (ETOP), adriamycin (ADR), ultraviolet irradiation (UV), and
ionizing radiation (IR) activate ERK in various cell lines. We show
that this process depends on ATM and is independent of p53.
Functionally, we demonstrate that ERK activation cooperates with p53 to
lead to apoptosis or cell cycle arrest. This represents the first
report to fully outline an ERK-mediated DNA damage pathway and its
functional importance.
| |
EXPERIMENTAL PROCEDURES |
Materials, Cell Lines, and Plasmids--
Hoechst 33258, propidium iodide, ETOP, ADR, hydroxyurea, and wortmannin were
from Sigma Chemical Co. The MEK1 inhibitors, PD98059 and U0126, were
from Calbiochem and Promega, respectively. Stocks of ETOP, ADR,
wortmannin, PD98059, and U0126 were made in Me2SO and used
at the concentrations indicated. Hygromycin B was from Invitrogen.
NIH3T3 and MCF7 cells were cultured in Dulbecco's modified Eagle's
medium supplemented with 10% fetal calf serum (Invitrogen) at 37 °C
in a tissue culture incubator. MEF were isolated and cultured in
Dulbecco's modified Eagle's medium, 10% FCS, 55 µM 
-mercaptoethanol, and 10 µg/ml gentamicin. IMR-90 was from ATCC and cultured in MEM and 10% FCS. The AT fibroblast line, GM05823, was
from the NIGMS Human Genetic Cell Repository and cultured in MEM, 10%
FCS, with 2× concentrated essential and non-essential amino
acids and vitamins.
The constitutively activated MEK1/Q56P cloned in the retroviral vector,
pBabe, was kindly provided by Dr. Scott Lowe of Cold Spring Harbor
Laboratories (23). pBabe/Bcl-2 and pBabe/Bcl-xL were
constructed by insertion of human Bcl-2 and Bcl-xL into
pBabe as we have published previously (31).
Retroviral Infection--
Retroviral infection was performed as
we have shown previously (31, 32). Constructs were subcloned into a
retroviral vector. Plasmid DNA was isolated and used to generate
retrovirus using phoenix packing cells. Briefly, 5 × 106 phoenix cells were seeded in a 10-cm plate overnight
before transfection with 15 µg of retroviral vector for 48 h.
The virus-containing medium was harvested after 48 h and filtered
through a 0.45-µm filter. After addition of 10 µg/ml Polybrene
(Sigma), the medium was used to infect NIH3T3 cells three times for
24 h at 8-h intervals to maximize transfection rate. Expression
was confirmed with Western blot.
Cell Lysis and Western Blot--
After exposure to DNA damage
agents as indicated, cells were lysed in a buffer containing 20 mM Tris (pH 7.4), 150 mM NaCl, 1 mM
EDTA, 1 mM EGTA, 1% Triton X-100, 25 mM sodium
pyrophosphate, 1 mM NaF, 1 mM
-glycerophosphate, 0.1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 2 µg/ml leupeptin, and
10 µg/ml aprotinin. 50 µg of total cell lysate was separated on
SDS-PAGE gel and transferred onto Immobilon-P membranes (Millipore).
Membranes were blocked with 5% skim milk and then incubated with the
indicated antibodies at room temperature for 1 h. Signals were
detected using an ECL Western blotting Kit (Amersham Biosciences,
Inc.). Primary antibodies and concentration used were: anti-ATM (Ab-3 at 2 µg/ml, Oncogene); anti-p53 (FL-353 at 1 µg/ml, Santa Cruz Biotechnology); anti-phospho-p53(S15) (Cell Signaling, 1:1000); anti-ERK (1:500, New England BioLabs); anti-phospho-ERK (1:500 New
England BioLabs); anti-Actin (Santa Cruz Biotechnology); and anti-p21CIP1 (1 µg/ml, Santa Cruz Biotechnology).
Apoptosis Assay, Viability Assay, and Determination of Cell Cycle
Distribution--
DNA fragmentation and TUNEL assay were performed as
we have previously published (31-33). For the cell viability assays,
2 × 104 cells were seeded in 96-well plates and
incubated at 37 °C for 4 h. Cells were then treated with DNA
damage stimuli as indicated with or without PD98059 (50 µM) or U0126 (10 or 50 µM) for the indicated times. WST-1 cell proliferation reagent (Roche Diagnostics, Mannheim, Germany) 10 µl/100 µl medium was added, and cells were incubated for 30 min at 37 °C. Viable cell numbers were determined at 450 nm with a microplate reader (Fisher Diagnostics). Cell cycle
distribution was determined by staining NIH3T3 cells with propidium
iodide solution in the presence of RNase A overnight at 4 °C
followed by detection of signals using fluorescent automated cell
sorting (FACS) as we have published previously (11).
Mitotic Index--
Cells were grown to 80% confluence on a
glass slip. After treatment with ETOP at the indicated concentrations
for 30 min, nocodazole (200 ng/ml) was added. After 12, 18, 24, and
30 h, cells were fixed in 3.7% formaldehyde and stored at 4 °C
for 24 h. Cells were then permeabilized in a Triton
X-100-containing solution and stained with Hoechst 33258 DNA stain (10 µg/ml, Molecular Probes). The mitotic index was scored under a
fluorescence microscope.
| |
RESULTS |
Multiple DNA Damage Stimuli Activate ERK in Several Cell
Lines--
To determine whether ERK is activated by DNA damage
signals, early passage MEF cells were treated with ETOP, which induces double-strand DNA breaks by interfering with the function of DNA topoisomerase II (34). As expected, ETOP induced stabilization of p53
and up-regulation of p21CIP1 (data not shown). ETOP induced
ERK activation (as measured by phosphorylation) with two phases (Fig.
1A). The first phase of ERK
activation starts at 1 h and ends at 8 h of ETOP exposure, whereas the second takes place after 15 h in response to ETOP (Fig. 1A). Increasing ETOP concentrations resulted in
increasing levels of ERK activation with maximal ERK activation
observed at or higher than 25 µM (data not shown),
indicating a correlation between intensity of DNA damage and level of
ERK activation. To determine whether ERK activation is a unique
property of ETOP or MEF cells, we used multiple DNA damage stimuli to
initiate ERK activation. ETOP, ADR, and UV led to ERK activation in
primary (MEF), immortalized (NIH3T3), and transformed (MCF7) cells
(Fig. 1B). Additionally, IR and hydroxyurea also
activated ERK in MEF and IMR90 (data not shown), confirming that DNA
damage indeed leads to ERK activation.
View larger version (35K):
[in this window]
[in a new window]
|
Fig. 1.
ERK activation is seen in response to various
DNA damage stimuli. A, MEF cells were exposed to ETOP
(100 µM) for the indicated times. ERK activation was
assessed by Western blot using an anti-phospho-ERK antibody. Total ERK
protein was assessed by Western blotting of the same membranes with an
anti-ERK antibody. B, 2 × 106 primary
(MEF), immortalized (NIH3T3), and transformed
(MCF7) were seeded for 24 h in 60-mm plates. Cells were
treated with ETOP at 100 µM for 2 h, ADR at 1.6 µM for 4 h, or UV at 100 mJ/cm2 for
4 h. ERK activation was assessed by Western blot using an
anti-phospho-ERK antibody. Densitometric data are shown with standard
error bars. All experiments were performed at least three times.
|
|
Because ERK activation is known to require dual phosphorylation on
Thr-183 and Tyr-185 by MEKs (35, 36), we sought to determine whether
DNA damage-induced ERK activation also requires MEK activity. Addition
of the MEK1 inhibitors, PD98059 (Fig. 3B) and U0126 (data
not shown), prevented ETOP-induced ERK activation. Complete abrogation
of ERK activation was also observed when cells incubated with PD98059
or U0126 were exposed to ADR, UV, or IR (data not shown), confirming
the role of MEK1 in mediating DNA damage-induced ERK activation. Taken
together, these results show that ERK activation is a component of
cellular DNA damage responses initiated by diverse stimuli.
ERK Activation in Response to ETOP Requires
ATM--
Because ATM plays a central role in the relay of DNA damage
signals (4) and ATM has been shown to play a role in the DNA damage-induced activation of JNK and p38 (37, 38), we sought to
determine if ATM played a role in DNA damage-induced ERK activation. As
a member of the PI3K family, the kinase activity of ATM is inhibited by
the PI3K inhibitor, wortmannin (39, 40). Addition of wortmannin
dose-dependently inhibited ETOP but not UV-induced ERK
activation (data not shown), consistent with the fact that wortmannin
inhibits the kinase activity of ATM but not ATR (40) and that
UV-induced DNA damage responses are generally
ATR-dependent. To confirm ATM dependence of ETOP-induced
ERK activation, both ATM
/ (GM05823) and IMR90 (wild type) human
fibroblasts were used. Expression of ATM in IMR90 but not in GM05823
was confirmed by Western blot using an anti-ATM (Ab-3) antibody (data
not shown). Addition of ADR or ETOP to IMR90 cells led to p53
stabilization, associated with phosphorylation on Ser-15 and
up-regulation of p21CIP1 (Fig.
2). Ser-15 phosphorylation-associated p53
stabilization was significantly reduced, and p21CIP1
induction was abolished in GM05823 cells in response to ADR or ETOP
(Fig. 2), further confirming the lack of ATM function in the ATM/
fibroblasts. Importantly, ETOP-induced ERK activation was absent in
GM05823 but present in IMR90 (Fig. 2), confirming a role of ATM in
ETOP-induced ERK activation. However, ADR-induced ERK activation was
not attenuated in the ATM/ fibroblast line (Fig. 2), suggesting
that ATR might also play an important role in this setting. Caffeine (2 mM), which inhibits both ATM and ATR (41), did prevent
ADR-induced ERK activation in this context, supporting the supposition
that ATR plays a role in the transmission of ADR-mediated DNA damage to
ERK (data not shown).
View larger version (46K):
[in this window]
[in a new window]
|
Fig. 2.
ERK activation by DNA damage signals requires
ATM. Wild type fibroblast IMR90 and ATM/ fibroblast cells,
GM05823, were seeded and grown to confluency in 60-mm plates. Cells
were treated with ETOP at 100 µM for 2 h or ADR at
1.6 µM for 4 h as indicated. Western blot was
carried out to determine p53 stabilization with an anti-p53 antibody,
phosphorylation on the Ser-15 of p53 with anti-phospho-p53(S15),
up-regulation of P21CIP1 with anti-p21CIP1, and
ERK activation with an anti-phospho-ERK antibody. Blots were routinely
re-probed with an anti-actin antibody to control for loading. All
experiments were performed at least in triplicate, and representative
radiographs are shown.
|
|
DNA Damage-induced ERK Activation Is Independent of
p53--
Because p53 is the major effector protein downstream of ATM
mediating the DNA damage response, we sought to investigate if ERK
activation played a role in the p53 pathway, as has been suggested by
others (25-27). Kinetically, p53 stabilization by ETOP (Fig. 3A) and ADR and IR (data not
shown) preceded ERK activation in all cell lines studied. Fig.
3A shows MCF-7 cells. MEK1 inhibition with PD98059 inhibited
ERK activation but was without effect on p53 stabilization in response
to DNA damage in MEF, NIH3T3, and MCF-7 cells (Fig. 3B,
NIH3T3), suggesting that ERK activation was not upstream of p53.
Conversely, pretreatment of cells with cycloheximide for 30 min to
destroy endogenous p53 prevented ETOP (data not shown)- and ADR (Fig.
3C, NIH3T3)-induced p53 stabilization but not ERK
activation, indicating that ERK activation is not downstream of p53.
However, given conflicting data suggesting that ERK activation may be
either upstream (27) or downstream (26) of p53 after DNA damage, we
wished to confirm our observations. In support of our results, a
maximal level of ERK activation was not attenuated in
MEFp53
/ compared with wild type MEF (Fig.
3D, inset), although the -fold increase of ERK
activation is slightly lower in MEFp53/ due to higher
basal level of ERK activity in these p53-negative cells (Fig.
3D). ADR and UV also induced ERK activation in
MEFp53/ (data not shown). Taken together, these data
indicate quite conclusively that under physiologic conditions, DNA
damage-induced ERK activation is independent of p53.
View larger version (39K):
[in this window]
[in a new window]
|
Fig. 3.
ERK activation by DNA damage occurs
independently of p53. A, MCF-7 cells were seeded for
24 h in 60-mm plates and then treated with ETOP at 100 µM for the indicated times. Stabilization of p53 with an
anti-p53 antibody and ERK activation with an anti-phospho-ERK antibody
were determined. B, NIH3T3 cells were pretreated with
PD98059 (100 µM) for 30 min before addition of 100 µM ETOP for 2 h. ERK activation, p53 stabilization,
and p21CIP1 up-regulation were then examined by Western
blot with their respective antibodies. Blots were routinely re-probed
with anti-actin to ensure equivalence of loading (data not shown).
C, NIH3T3 cells were pretreated with cycloheximide (10 µg/ml) for 30 min before addition of 1.6 µM ADR for
4 h. ERK activation and p53 stabilization were determined by
Western blotting with their respective antibodies. Blots were routinely
re-probed with anti-actin to ensure equivalence of loading (data not
shown). D, wild type and p53/ MEFs were
treated with ETOP at 100 µM for 2 h, ADR at 1.6 µM for 4 h, or UV at 100 mJ/cm2 for
4 h. ERK activation was determined by Western blot using
anti-phospho-ERK antibody (inset gel) and quantified.
Densitometric results from three independent experiments are shown with
error bars.
|
|
ERK Activation Maximizes p21CIP1 Induction upon DNA
Damage--
The two paradigmatic cellular responses to DNA damage are
cell cycle arrest (allowing the damage to be repaired) or apoptosis in
response to low or high intensities of DNA damage, respectively. The
Cdk inhibitor, p21CIP1, is up-regulated in this process and
plays a major role in arresting cells in either G1 or at
G2/M in response to DNA damage (42, 43). We have
observed that wortmannin affected ERK activation and
p21CIP1 induction in a more clearly
dose-dependent manner than p53 stabilization after DNA
damage (data not shown), suggesting that ERK activation might
up-regulate p21CIP1. This is consistent with the
observations that enforced ERK activation by overexpression of a
constitutively activated Raf1 led to p21CIP1 induction (44,
45).
Consequently, we examined the effect of inhibition of ERK activation on
p21CIP1 induction after DNA damage. Inhibition of ERK
activation by PD98059 reduced p21CIP1 induction by half
(Figs. 3B and 4) but had no effect on p53 stabilization by
ETOP (Fig. 3B). PD98059 also attenuated ADR-induced
p21CIP1 induction (Fig. 4).
As expected, overexpression of a constitutively activated MEK1/Q56P
(23) strongly activated ERK in NIH3T3 and MEF cells (data not shown)
and led to p21CIP1 induction that was also inhibited by
PD98059 (Fig. 4). Furthermore, expression of MEK1/Q56P in
MEFp53/ cells also resulted in p21CIP1
induction (data not shown), indicating that ERK activation is capable
of inducing p21CIP1 expression, consistent with prior
observations with forcibly expressed Raf1 (44, 45). Our data, however,
is the first to demonstrate p21CIP1 induction by ERK under
physiologic conditions.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 4.
ERK activation contributes to DNA
damage-induced p21CIP1 up-regulation. NIH3T3 cells
were seeded for 24 h in 60-mm plates and then treated with ETOP at
100 µM for 2 h or ADR at 1.6 µM for
4 h with or without PD98059 (100 µM) as indicated.
Similarly, cells infected with retrovirus expressing MEK1/Q56P were
grown for 24 h and then treated with PD98059 (100 µM) as indicated. The abundance of p21CIP1
protein was determined by Western blot and quantified. Densitometric
results from three independent experiments are shown with error
bars.
|
|
ERK Activation Is Required for Cell Cycle Arrest upon
DNA Damage--
Because ERK activation partially contributes to
p21CIP1 induction (Figs. 3B and 4), and
p21CIP1 plays an important role in both induction of
G1 arrest (43) and maintenance of arrest at the
G2/M checkpoint (42), we sought to determine whether the
magnitude of p21CIP1 induction induced by ERK activation
played a functional role in cell cycle arrest in response to DNA
damage. Although PD98059 100 µM alone results a degree of
G1 arrest (Fig.
5B), consistent with the
notion that ERK activity facilitates G1 progression (46), addition of PD98059 to ETOP 0.5 µM (Fig. 5D)
led to a marked release from the G2/M arrest seen with
etoposide 0.5 µM alone (Fig. 5C). We confirmed
that G2/M release was indeed occurring by mitotic labeling
with nocodazole trapping and showed an increase of mitotic cells from 5 to 11% at 24 h when PD98059 is added to ETOP 0.5 µM
(data not shown). To further confirm a role of ERK in maintaining G2/M arrest in response to ETOP, a more specific MEK1
inhibitor, U0126, was also used. Although 50 µM U0126
alone has a minimal effect on G1 phase (Fig.
5E), 10 and 50 µM U0126 again dramatically released ETOP (0.5 µM)-induced G2/M
arrest (Fig. 5E). Similar results were observed with ADR
(data not shown), although the effect of PD98059 on cell cycle arrest
induced by ADR is less prominent than that induced by ETOP, consistent
with the observation that inhibition of ERK activation was more
effective in preventing ETOP- than ADR-mediated
p21CIP1 induction (Fig. 4).
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 5.
ERK activation functions in DNA
damage-induced cell cycle arrest. Experiments shown were done
after 24-h exposure to the indicated compounds and were performed in at
least triplicate. The percentages of cells in each phase were
determined by propidium iodide staining followed by fluorescence
automated cell sorting (63). A, NIH3T3 cells exposed to
vehicle (Me2SO). B, NIH3T3 cells were exposed to
PD98059 (100 µM). C, NIH3T3 cells were exposed
to ETOP (0.5 µM). D, PD98059 (100 µM) was added to ETOP (0.5 µM).
E, U0126 was used in the indicated concentrations instead of
PD98059 with or without ETOP as indicated. The percentage of cells in
each phase is shown graphically. Error bars are based on the
results of four separate experiments.
|
|
ERK Activation Facilitates DNA Damage-induced Apoptosis--
Given
our data for a pro-cell cycle arrest function of ERK in response to DNA
damage, we wished to determine if DNA damage-induced ERK activation
also mediated apoptotic responses. Addition of ETOP 10 µM
for 24 h led to DNA fragmentation in NIH3T3 cells, which was
essentially prevented by ERK inhibition (Fig.
6A). Overexpression of the
anti-apoptotic proteins Bcl-2 or Bcl-xL also prevented apoptosis in response to ETOP, indicating that the mitochondrial cellular execution pathway plays a role in DNA damage-induced apoptosis
(Fig. 6A). Quantification of apoptosis using terminal deoxyuridine end-nick labeling (TUNEL) revealed that 65% of cells were
apoptotic when exposed to 10 µM ETOP for 24 h, and
that ERK inhibition with PD 98059 dose- dependently reduced this to
less than 10%, suggesting a major role for ERK in DNA damage-induced apoptosis in response to ETOP (Fig. 6B). Similarly, Bcl-2 or
Bcl-xL overexpression also reduced apoptosis to ~10%,
indicating again the importance of the mitochondrial execution pathway
(Fig. 6B). Identical results were also obtained when U0126
was used. Although U0126 alone at 50 µM induced no
apoptosis, addition of 10 and 50 µM strongly inhibited 10 µM ETOP-induced apoptosis (Fig.
7). When TUNEL was used to quantify
apoptotic cells in Fig. 7, ETOP alone resulted in 70% apoptosis and
U0126 at 10 or 50 µM reduced this to 12 and 4%,
respectively (data not shown), confirming the importance of ERK
activity in ETOP-induced apoptosis.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 6.
ERK activation mediates DNA damage-induced
apoptosis. NIH3T3 cells were infected with either a control
retrovirus, pBabe, or retrovirus expressing Bcl-2 or Bcl-xL
and seeded for 24 h in 60-mm plates. A, cells were then
treated with ETOP (10 µM) with or without PD98059 (100 µM), and DNA fragmentation was determined. B,
cells were treated with ETOP (10 µM) with or without
PD98059 at the indicated concentrations, and apoptosis was determined
by TUNEL. Results from three independent experiments are shown with
error bars.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 7.
ERK activation mediates DNA damage-induced
apoptosis. NIH3T3 cells were treated with Me2SO
(1), 50 µM U0126 (2), 10 µM ETOP (3), 10 µM U0126 plus 10 µM ETOP (4), or 50 µM U0126 plus
10 µM ETOP (5) for 24 h. Cells were then
stained with propidium iodide, followed by detection of apoptotic cells
using FACS. Cells with sub-G1 DNA content are apoptotic,
and the percentage of total cells in this area is given. Results shown
are representative of four separate experiments.
|
|
We reasoned that if ERK is central in induction of apoptosis in
response to DNA damage, then overexpression of ERK should further
sensitize cells to undergo programmed death when exposed to ETOP.
Indeed, TUNEL revealed marked (15%) apoptosis in NIH3T3 cells infected
with retrovirus for MEK1/Q56P, a constitutively active MEK (23), in
response to 5 µM ETOP for 18 h, a condition usually
associated with <5% apoptosis in NIH3T3 cells infected with control
retrovirus (Fig. 8). Inhibition of MEK-1
with PD98059 again essentially prevented this (Fig. 8).
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 8.
Enforced ERK activation sensitizes to DNA
damage-induced apoptosis. NIH3T3 cells were infected with either a
control retrovirus, pBabe, or retrovirus expressing MEK1/Q56P and
seeded for 24 h in 60-mm plates. Cells were treated with ETOP at
the indicated concentrations with or without PD98059 (100 µM) and apoptosis determined by TUNEL staining. Results
were from three independent experiments and are shown with error
bars.
|
|
Because our data indicated that ERK was activated by a number of DNA
damage stimuli (Fig. 1B), we sought to ensure that
pro-apoptotic signaling through ERK was a general response to DNA
damage. Consequently, the effects of PD98059 on UV- and ADR-induced
cell death were also examined. Inhibition of ERK activation by PD98059
increased cell survival in response to both UV and ADR treatments (Fig. 9, A and B). Taken
together, these results show that ERK activation facilitates DNA
damage-induced apoptosis.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 9.
ERK activation contributes to UV and
ADR-induced cell death. NIH3T3 cells were seeded for 24 h in
60-mm plates. A, cells were treated with UV at 100 mJ/cm2 with or without PD89059 (50 µM) for
the indicated time periods and viability determined as described under
"Experimental Procedures." Results were from three independent
experiments and are shown with error bars. B,
cells were treated with ADR for 19 h at the indicated
concentrations with or without PD89059 (50 µM), and
viability was determined. Results were from three independent
experiments and are shown with error bars.
|
|
| |
DISCUSSION |
DNA damage surveillance systems play an important role in
maintaining genomic integrity. Inactivations in or losses of components of these pathways results in genetic instability and may lead to the
accumulation of multiple genetic alterations and development of
neoplasia (47-49). All human tumors display genetic instability (47),
and functional inactivation of p53, a major player in DNA damage
checkpoint pathways, occurs in more than 50% of cancers (50). Lack of
a functional ATM kinase, the best characterized transducer of DNA
damage, is also characterized by chromosomal instability in response to
DNA damage and the development of cancer (51, 52).
In this report, we add novel information to an already very complex DNA
damage response. We initially observed that DNA damage agents were able
to initiate ERK activation as measured by ERK phosphorylation in
several cell lines (Fig. 1), and the degree of ERK activation
correlated with the intensity of DNA damage. This is consistent with
recent reports that treatment of cells with UV, mitomycin-C, and
cisplatin led to ERK activation (25, 26, 29). Our new observations
result from detailed characterization of the mechanisms and
consequences of ERK activation after ETOP-induced DNA damage. We
have mapped ERK activation to be downstream of ATM in response to ETOP,
but, importantly, independent of p53 (Figs. 2 and 3). Together with
p53, such ERK activation contributed to DNA damage-induced cell cycle
arrest and apoptosis.
DNA damage activates the kinase activity of ATM, which
subsequently modifies a number of downstream targets (4), including the
phosphorylation of Ser-15 at the N terminus of p53 (9, 10). This
modification prevents ubiquitin-dependent p53 degradation by interfering with the interaction between p53 and Mdm2 (53, 54). We
report here that ETOP is unable to activate ERK in ATM/ fibroblasts
but does so in wild type IMR90 fibroblasts, indicating that ERK is a
downstream target of ATM in response to DNA damage. Supporting this is
our observation that wortmannin, a PI3K inhibitor, also blocked
ETOP-induced ERK activation. Interestingly, it has been reported
recently that ATM kinase functions in the insulin signaling pathway to
stimulate protein synthesis by phosphorylating 4E-BP1 (55), supporting
the hypothesis that the role of ATM in cell signaling may not be
restricted to the maintenance of genomic stability. This raises the
possibility that ERK might also be a target of ATM normally, and
therefore, the lack of a functional ATM kinase might interfere with ERK
activation and cell proliferation. This would be consistent with the
observations that ATM/ cells proliferate poorly and require high
levels of growth factors (56). To further support this scenario, we
have observed that wortmannin at concentrations of or higher than 100 nM also inhibits the basal level of ERK
activation.2 Because the
IC50 values of wortmannin for lipid PI3K, DNA-PK, and ATM
kinase are less than 5, 16, and approximately100 nM,
respectively (39, 40, 57), these observations indicate a potential role for ATM in broader cell signaling by affecting ERK activation.
Because p53 protein is a major effector in the DNA damage response, it
is essential to determine any linkage between p53 and ERK activation
after DNA damage. Like other oncogenic signals, ERK activation leads to
p14ARF-dependent p53 activation in primary but
not immortalized MEF (23, 58-61). In the case of DNA damage response,
however, we have found that ERK activation occurs independently of p53.
This is based on following observations: 1) the MEK1 inhibitor,
PD98059, prevented ERK activation but not p53 stabilization in response to DNA damage (Fig. 3B); 2) preincubation with cycloheximide
at 10 µg/ml for 30 min to destroy endogenous p53 completely blocked DNA damage-induced p53 stabilization but not ERK activation (Fig. 3C); and 3) the maximal level of ERK activation by UV, ETOP,
and ADR was not attenuated in MEFp53/ (Fig.
3D). This is consistent with the observation that enforced activation of ERK leads to cell cycle arrest in a p53-independent manner (44, 45).2
Our results are different from reports that ERK activation lies either
upstream or downstream of p53 in response to various DNA damage agents.
Inhibition of ERK activation by PD98059 was found to prevent
cisplatin-induced p53 stabilization by reducing phosphorylation of
Ser-15 in the N-terminal region of p53 (25). Although it was not
clearly demonstrated, the authors indicated that, by interaction with
p53, ERK phosphorylated p53 on Ser-15. It is unclear how ERK, a
proline-directed kinase, could phosphorylate Ser-15 of p53, because
this residue is not followed by a proline, but rather a Gln residue.
Furthermore, cisplatin-induced ERK activation could be inhibited by
suramin, a growth factor receptor antagonist, suggesting that ERK
activation might not be related to cisplatin-induced DNA damage (29).
However, the possibility remains that these observations may be due to
the different DNA damage stimuli used. This is supported by data
indicating that mitomycin C- and actinomycin D-induced ERK activation
was abolished in a p53-defective tumor cell line (26), although in this
case p53 would be placed upstream of ERK activation.
In support of the notion that p53 is required for some forms of DNA
damage-induced ERK activation, Lee et al. (26) showed that
overexpression of wild type p53 itself led to ERK activation. An
alternative explanation for this is that overexpression of p53 leads to
feedback activation of a DNA damage response pathway, resulting in ATM
kinase activation. ATM kinase in turn leads to ERK activation,
consistent with our observation that ERK activation depends on ATM
after DNA damage. Supporting this inference was the observation that
overexpression of wild type p53 led to activation of p53 (26). However,
it is known that simply increasing the abundance of p53 is insufficient
to activate p53 pathway (62).
Although independent of p53, our data show that ERK activation makes
important contributions to functional outcomes, cell cycle arrest and
apoptosis, after DNA damage. Inhibition of ERK activation by PD98059 or
U0126 (data not shown) diminished p21CIP1 induction,
explaining the partial release by these inhibitors from cell cycle
arrest in response to low concentrations of either ETOP (Fig. 5) or ADR
(data not shown), because p21CIP1 is required to maintain
G2/M arrest (42). Although enforced activation of ERK is
known to lead to p21CIP1 induction in a p53-independent
manner (44, 45), our demonstration here is the first that ERK
activation contributes to the induction of p21CIP1 under
physiologic conditions. In response to higher doses of ETOP, ADR, and
UV, inhibition of ERK activation by PD98059 and U0126 attenuated
apoptosis (Figs. 6 and 7). Because DNA damage-induced cell cycle arrest
and apoptosis requires p53, our results indicate the existence of a
parallel ERK-dependent pathway also functional in DNA
damage-induced cell cycle arrest and apoptosis.
The data presented here demonstrate that ERK activation after DNA
damage plays a functional role in surveillance and response to DNA
damage, independently of p53. Because many malignant cells display high
levels of ERK activation, these findings may help explain the
heightened sensitivity of some transformed cells to DNA-damaging
chemotherapeutic agents when compared with primary cells. It will be
instructive to examine such lines for amplifications of components of
the ERK signaling pathway.
§,
,
TOP
ABSTRACT
INTRODUCTION
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
