 |
INTRODUCTION |
Apoptosis is a form of "cell suicide" in which cells fragment
their cytoplasm, nucleus, and DNA for orderly disposal (1, 2). It is
ubiquitous and essential to normal homeostasis in higher animals and is
important in modulating immune cell populations for defense against
cancer and virus-infected cells (3, 4). Apoptosis is induced through
activation of caspases, which activate one another in a proteolytic
cascade (5), leading to cleavage of cytoskeletal and signaling
molecules that result in irreversible changes in cell morphology and
the plasma membrane (6-10).
The Rho family of small GTPases plays a role in the terminal
morphological changes of apoptosis (11), but previous reports present
conflicting data about its role in apoptosis induction. Apoptotic
signaling pathways can be grouped into two distinct categories (12):
mechanisms that induce rapid apoptosis through specific molecules that
have evolved to produce cell death (i.e. Fas ligand or
cytotoxic T cell granzymes) (12) and pathways that lead to apoptosis
when normal homeostatic signaling is disrupted (13-15). Examples of
the latter type of perturbation include serum withdrawal (13),
detachment from extracellular matrix (14), and an imbalance of survival
homeostatic signaling pathways (15). Numerous studies of Rho family
signaling in such pathways have revealed that the GTPases play a
complex role, participating in both stimulatory and inhibitory
paradigms (16-18). Downstream mediators for positive (i.e.
c-Jun N-terminal kinase
(JNK)1) and negative
(i.e. Akt) effects have been identified. In contrast, much
less is known about Rho family signaling in apoptosis produced by CTLs,
Fas, and other molecules directly inducing apoptosis. Here we show that
Rho family GTPases play a stimulatory role in CTL- and Fas-induced apoptosis.
The Rho family of small GTPases consists of three proteins (Cdc42, Rac,
and Rho) that interact with downstream effectors upon binding to GTP.
Their nucleotide state and localization are controlled by >40
currently identified regulatory factors, including guanine nucleotide
exchange factors, GDP dissociation inhibitors, and GTPase-activating
proteins (19-23). Rho family proteins have been implicated in
controlling both transcriptional activation and a wide range of cell
morphological changes, including mitosis, nerve outgrowth, and immune
cell motility (24-30). Their effects on cell shape are mediated by
regulation of the actin-myosin cytoskeleton, and each of the family
members has been shown to have different specific effects on actin in
studies of fibroblast morphology (31-33). Rho stimulates formation of
stress fibers and impacts on focal adhesion formation; Rac promotes
cell ruffles and lamellipodia; and Cdc42 controls the extension of
filopodia (34). The Rho GTPases also influence each other's activity:
Cdc42 activates Rac and can antagonize Rho activation, and in more
limited cases, Rac can activate Rho (33, 34).
Rho family proteins can coordinate multiple signaling pathways through
their ability to regulate both the cytoskeleton and transcription. They
induce rearrangements of the actin scaffolding to control the
organization and interactions of signaling molecules (35, 36), whereas
they directly stimulate transcriptional cascades (24-26). They also
control the formation of "signaling organelles" such as focal
adhesion complexes that affect both extracellular matrix signaling and
anchor actin fibers (37). Proteins that modulate Rho family nucleotide
state contain both catalytic and localization domains (19-23),
indicating that control of cell morphology through localization of Rho
family activity is coordinated with activation of downstream pathways.
This information suggested to us that the Rho proteins might also
regulate the signaling events controlling rapid apoptosis through
control of the actin cytoskeleton. We show here that, although Rho
family GTPases do not themselves induce rapid cell death, their
activity is critical for Fas and CTL death signaling. We also show that Rho proteins modulate CTL and Fas signaling through their effects on
the actin cytoskeleton.
| |
EXPERIMENTAL PROCEDURES |
Cell Lines and Antibodies--
Chinese Hamster Ovary (CHO) cells
stably transfected with the murine Fc
II receptor (38) were kindly
provided by Dr. Ira Mellman. This cell line was grown in Dulbecco's
modified essential medium (Life Technologies, Inc.) supplemented with
10% fetal bovine serum, 100 units/ml penicillin, and 100 µg/ml
streptomycin. CHO cells stably transfected with the extracellular
portion of the human CD4 protein fused to the cytoplasmic domain of the
murine Fas receptor (39) were grown in Ham's F-12 medium (Life
Technologies, Inc.) supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 µg/ml streptomycin. 145-2C11 hybridoma
cells producing anti-CD3 monoclonal antibody were obtained from the
American Type Culture Collection (Manassas, VA). This cell line was
grown in Iscove's Dulbecco's modified essential medium supplemented
with 20% fetal bovine serum, 100 units/ml penicillin, and 100 µg/ml streptomycin. SV40-transformed African green monkey kidney cells (COS-7) and Swiss 3T3 fibroblasts were obtained from the American Type
Culture Collection. Both cell lines were grown in Dulbecco's modified
essential medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 µg/ml streptomycin. Anti-human CD4
monoclonal antibody was obtained from Becton Dickinson (San Jose, CA).
Chromium Release Assay and Induction of CTL
Killing--
Specificities and culture conditions of the CTL clones
(NP18 and HD8) and procedures for 51Cr release assays of
CTL killing were as described previously (40, 41). After 45 min of
incubation with 51Cr, target cells were washed and combined
with CTLs in 96-well plates at an effector/target ratio of 5:1 and
incubated for 5-6 h. The radioactivity released by cells in each assay
was compared with detergent-lysed targets as a control for maximal
release of 51Cr and with target cells incubated without
CTLs as a control for spontaneous release. All CTL assays were
performed using triplicate wells with a variation always <15%, and
all assays except the dose-response curve in Fig. 1C were
performed at least twice. For induction of CTL killing through
lymphocytic choriomeningitis virus infection of target cells, target
cells were infected with the virus 48 h prior to assay at a
multiplicity of infection of 1. Target cells were MC57 fibroblasts
derived from C57 Bl/6 mice (H-2b haplotype) and Balb/clone 7 fibroblasts derived from Balb/c mice (H-2d haplotype).
Antibody-redirected killing was also carried out using previously
described methods (42). In these assays, CHO cells stably transfected
with the FcII receptor were used as targets, and killing was induced
by adding culture supernatant from 145-2C11 hybridoma cells producing
anti-CD3 monoclonal antibody. Clostridium toxin B was added
to target cells at the indicated concentration 2 h prior to
addition of CTLs, and the targets were washed three times with medium
prior to CTL addition. Results are reported as a percentage of the
maximum possible chromium release, determined by lysing control cells
with detergent. Background from spontaneous release in the absence of
CTLs was subtracted from each value and was in all cases <5% of that
induced by CTLs.
Plasmids and Transfection--
For the construction of mammalian
expression vectors for green fluorescent protein (GFP) S65T-GTPase
fusion proteins, a GFP S65T cDNA fragment (kindly provided by Dr.
Roger Tsien) was subcloned at the BamHI and EcoRI
sites into a pcDNA3 expression vector (Invitrogen, Carlsbad, CA).
This cDNA fragment was generated by polymerase chain reaction
amplification with two oligonucleotide primers based on the N- and
C-terminal sequences of the GFP cDNA, but with the stop codon
eliminated from the C-terminal primer. Constitutively active mutants of
Rho family GTPases (Cdc42 Q61L, Rac1 Q61L, and RhoA Q63L) and
dominant-negative mutants (Cdc42 T17N, Rac1 T17N, and RhoA T19N) were
polymerase chain reaction-amplified from pcDNA3 vectors (43) and
fused into the GFP vector at the EcoRI and XhoI
sites. All of the constructed cDNA vectors were isolated from
transformed Escherichia coli cells with QIAGEN cDNA
isolation kits. Cells were transfected through nuclear microinjection
of log-phase cells with 50 ng/ml DNA in water at 20-30 h prior to the
experiment. Rac1 vectors containing F37A and Y40C point mutations were
kindly provided by A. Hall (60). Cells were cotransfected through
nuclear microinjection with each of these mutants and GFP at
concentrations of 50 and 10 ng/ml DNA, respectively. The GTPase-binding
domain from human p21-activated kinase (amino acids 67-150) was cloned
into the bacterial expression vector pGEX-4T3, expressed in E. coli as a fusion protein with glutathione
S-transferase, and purified as described previously
(61).
JNK Assay--
JNK activity was determined using GST-c-Jun as
substrate for a solid-phase assay as described (43). HeLa cells were
transfected with the GFP-Rho family fusion constructs, and at the
indicated times after selection, 8 × 105 cells were
lysed in 100 µl of cell lysis buffer (25 mM HEPES (pH
7.5), 300 mM NaCl, 1.5 mM MgCl2,
0.2 mM EDTA, 0.1% Triton X-100, 0.5 mM
dithiothreitol, 20 mM 
-glycerophosphate, 1 mM sodium orthovanadate, 0.5 mM
phenylmethylsulfonyl fluoride, 5 µg/ml aprotinin, and 0.5 µg/ml
leupeptin). 6 µg of GST-c-Jun-(1-79) coupled to glutathione-Sepharose beads was added to clarified lysates and incubated for 3 h at 4 °C. Bead pellets were washed, and kinase assays were performed as described (43). Quantitation was done using a
PhosphorImager (Molecular Dynamics, Inc.).
Fas-induced Apoptosis--
These experiments were carried out
using CHO cells transfected with a fusion protein in which the
intracellular domain of the Fas receptor is coupled to the
extracellular domain of CD4. Antibodies against CD4 and cross-linking
secondary antibodies were used to induce apoptosis as described
previously (39).
Rho GTPase Activation Assay--
The level of Cdc42-GTP was
determined using a recently described assay that measures the formation
of Cdc42-GTP based on its specific binding to the GTPase-binding domain
of p21-activated kinase (61). Target cells (107 cells/ml)
were stimulated with Fas for the times indicated. At the appropriate
time, cell activation was stopped by addition of ice-cold 2× lysis
buffer (50 mM Tris-HCl (pH 7.5), 10 mM
MgCl2, 200 mM NaCl, 2% Nonidet P-40, 10%
glycerol, 2 mM phenylmethylsulfonyl fluoride, 2 µg/ml
leupeptin, and 2 µg/ml aprotinin). Cells were vortexed and clarified
by low speed centrifugation at 4 °C. Guanine nucleotide loading of
lysates and affinity precipitation using GST-p21-binding domain were
carried out as described (61). The quantity of total protein assayed
from the lysate was determined using the BCA protein assay (Pierce).
Live Cell Microscopy for Apoptosis and Cytoskeletal
Assays--
CHO/Fc cells were microinjected with vectors expressing
GFP fusion proteins and maintained in an incubator for 24 h prior to the experiment. Imaging was performed on subconfluent log-phase CHO
cells used 1-2 days after plating. Immediately prior to the experiment, the medium was changed to complete Dulbecco's modified essential medium free of phenol red, but supplemented with 10% fetal
bovine serum. Coverslips with adherent cells were mounted in a sealed
Dvorak chamber (Nicholson Instruments) held in a temperature-controlled stage (20/20 Technologies). Microscopy was performed using an Axiovert
100 TV microscope (Carl Zeiss, Inc.) modified with automated stage and
filter wheels (LEP Ltd.) and a 40× NA 1.3 objective with differential
interference contrast optics (Carl Zeiss, Inc.). Control of microscope
automation and image analysis was performed using Inovision ISEE
software. Fluorescence images were obtained with a Photometrics PXL
cooled charge-coupled device camera with 2 × 2 binning as 12-bit,
658 × 517-pixel arrays using GFP or 4,6-diamidino-2-phenylindole filter sets (Chromatech Co.). Fluorescence images were
contrast-stretched and sharpened for clarity of display using ISEE and
Adobe Photoshop software (44). Unless otherwise mentioned, the extent
of apoptosis was assessed 3 h after targets were first exposed to
CTLs. In the live cell studies shown in Fig. 9, 30 h after
injection with DNA, the cells were preincubated with toxin B (12.5 ng/ml) for 2 h, washed thoroughly, and finally treated using the
antibody-redirected CTL killing method (see above). Hoechst staining of
nuclei was performed as described previously (45). Latrunculin A was
used at 2 µg/ml to produce specific inhibition of actin
polymerization (46). For CTL-induced apoptosis, target cells were
incubated in latrunculin for 1 h and then washed prior to addition
of CTLs. In Fas killing experiments, latrunculin was added to the cells for 1 h, followed by addition of antibody to the
latrunculin-containing medium for an additional 1 h. Cells were
then washed, and antibody was added again. Error bars represent S.E.
values calculated for three independent experiments. Each experiment
was repeated a minimum of three times. Background cell death in the
absence of apoptotic stimuli was subtracted from each value in Fig. 3
(B and C). These background levels were 3.62-6.45% (mean
of 4.79%).
Caspase-3 Activity Assay--
The synthetic tetrapeptide
fluorogenic substrate
N-acetyl-Asp-Glu-Val-Asp-7-amino-4-methylcoumarin (Enzyme
Systems Products) was used as described previously (47) to identify and
quantify caspase-3 activity in apoptotic cells. Release of fluorescent 7-amino-4-methylcoumarin was quantified in cell lysates by
spectrofluorometry using excitation at 380 nm and emission at 430-460 nm.
| |
RESULTS |
CTL-induced Apoptosis Is Inhibited by Clostridium difficile Toxin
B--
To probe whether Rho family proteins play a role in CTL-induced
apoptosis, we treated fibroblasts with C. difficile toxin B,
an inhibitor of all family members (48), and then assayed the cells'
susceptibility to CTL-induced killing. Clostridium toxin has
been shown to specifically glucosylate Cdc42, Rac, and Rho (48). The
experiments were performed using a well established assay of CTL
killing in which target cells that had internalized radioactive
chromium were exposed to CTLs, and the amount of radioactive chromium
released into the medium was determined (40, 41). Only target cells
were treated with toxin, eliminating the possibility that CTLs were
being affected by the toxin.
As shown in Fig. 1A, CTL
killing was inhibited at least 60% by the toxin, indicating that the
activity of at least one of the Rho family proteins was necessary for
CTL-induced apoptosis. The CTL clones are very specific for a peptide
of precise sequence complexed to a specific allele of the major
histocompatibility complex (MHC) I molecule on the target cell surface
(40). When we tested target cells presenting either the wrong peptide
or the wrong MHC-I molecule, no killing was observed. This showed that
death was not caused by the toxin or by a manipulation of the cells
during the experiment, but that death was due to CTL-induced apoptosis.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 1.
C. difficile toxin B inhibits
CTL-induced apoptosis. The effects of C. difficile
toxin B (100 ng/ml), an inhibitor of all three Rho family GTPases, on
CTL-induced apoptosis are shown in A. Cell death was
assessed by 51Cr release assay. Target cells were incubated
with 51Cr, washed, and then combined with CTLs. The
radioactivity released by target cells in each assay was compared with
that released by detergent-lysed target cells as a control for maximal
release of 51Cr and with that released by target cells
incubated without CTLs as a control of spontaneous release. The
first and last sets of bars show experiments in
which the CTL clone and target cell MHC allele were appropriately
matched. The second and third sets of bars are
controls in which CTLs were specific for MHC alleles not expressed on
the targets or in which targets were not infected with virus to induce
killing. In B, CTLs were caused to induce apoptosis using an
antibody (Ab)-redirected killing method. With this method,
the antigen processing and presentation system was bypassed to show
that the toxin B effects on CTL-induced apoptosis were not due to
alteration of antigen presentation. Treatment with antibody alone did
not increase apoptosis. C shows that toxin B inhibited
antibody-redirected killing in a dose-dependent manner,
with 12.5 ng/ml being the lowest dose that could significantly inhibit
cell death. All CTL assays were performed using triplicate wells with a
variation always <15%, and all assays except the dose-response curve
in C were performed at least twice. LCMV,
lymphocytic choriomeningitis virus.
|
|
We were concerned that the toxin was affecting not the apoptosis of the
target cells, but instead cell behaviors involved in activation of the
CTLs for killing, including antigen processing or specific interactions
between the T cell receptor and the target. To bypass the requirement
for antigen presentation and MHC-T cell receptor interactions, the
assays were repeated using an antibody-induced CTL activation procedure
(38). CHO cells stably transfected with the FcII receptor were
incubated with antibody against CD3
, a component of the T cell
receptor, and CTLs were then added. As described previously (42), the
FcII receptor on the target cell binds to the antibody, which binds
to the T cell receptor, cross-linking CTLs and the target while
activating killing. As shown in Fig. 1B,
Clostridium toxin profoundly blocked apoptosis in these
experiments, just as it had in CTL killing activated by peptide or
virus. The antibody in the absence of CTLs had no effect, showing that
CTL-induced apoptosis was responsible for the measured cell death. All
CTL killing experiments in the remainder of this study were induced
using this antibody-redirected procedure to bypass possible effects on
CTL activation and targeting.
Fig. 1C shows that toxin B inhibited antibody-directed
killing in a dose-dependent manner. Our studies were
carried out at 100 ng/ml, a concentration established in previous
reports to be specific for Rho proteins (1000 ng/ml and lower) (48).
Transfection of the target cells with constitutively active mutants of
Rho family proteins overcame the effects of the toxin, confirming that
the toxin acts upon Rho family proteins within cells (data not shown).
These studies established that C. difficile toxin B, a
specific inhibitor of all Rho GTPases, strongly inhibited the ability
of target cells to undergo CTL-stimulated apoptosis.
Rho Proteins Modulate the Induction of Apoptosis by Both CTLs and
Fas, but by Themselves, Do Not Cause Apoptosis--
We next examined
the effects of transfecting dominant-negative or activated Rho GTPase
mutants in the target cells. This enabled us to determine which
specific Rho proteins were involved and to examine whether the proteins
alone were sufficient to produce apoptosis in the absence of apoptotic
stimuli. CHO/Fc cells were transfected through microinjection with GFP
fusion proteins of activated or dominant-negative GTPase mutants and
maintained alive on the stage of an automated microscope. After
addition of CTLs, the behavior of each fluorescent cell was recorded. A
cell was considered to have undergone apoptosis when it exhibited a set of clearly discernible apoptotic morphological changes, including blebbing, contraction, and fragmentation (42). This was verified as
apoptosis by staining the nuclei with Hoechst dye to confirm that
nuclear fragmentation had occurred. As described below, results of
these assays were also found to correlate with caspase-3 activation.
It was important to demonstrate that the fusion of GFP to the Rho
family proteins did not block their biological activity. As shown in
Fig. 2A, Jun kinase activation
was substantially enhanced in cells transfected with constitutively
active GFP-Rac and GFP-Cdc42, but was only modestly affected by
GFP-Rho, consistent with the activity of proteins not fused to GFP
(43). In addition, the images shown in Fig. 2B demonstrate
that each of the constitutively active mutants produced the previously
described characteristic effects on cell morphology (31-33). In three
different cell types, GFP-Rho led to stress fiber formation; GFP-Rac
produced extensive ruffling; and GFP-Cdc42 led to filopodia
formation.
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 2.
GFP-Rho protein fusions have intact
biological activity. JNK activity was determined using a GST-c-Jun
solid-phase assay with the c-Jun amino terminus as substrate. As shown
in A, Jun phosphorylation was detected in cells transfected
with constitutively active Rac and Cdc42, but not with Rho, consistent
with the biological activity of GTPase proteins not fused to GFP. The
images shown in B demonstrate that each constitutively
active mutant produced the previously described characteristic effects
on cell morphology. The left panels show the effect of
constitutively active Rho. The upper left panel shows stress
fibers induced by Rho in the only cell in the field expressing
constitutively active GFP-Rho. Expression of Rho in that cell is seen
in the lower left panel. The upper middle panel
shows active ruffling in a COS-7 cell expressing GFP-fused
dominant-positive Rac. No ruffling is seen in the control COS cell
expressing GFP alone (lower middle panel). The
upper right panel shows filopodial formation in a CHO cell
expressing the activated GFP-Cdc42 fusion (see arrows). The
lower right panel shows control CHO cells expressing GFP
alone. wt, wild-type.
|
|
Transfection of CTL target cells with constitutively active or
dominant-negative mutants of each Rho family protein alone had little
effect, resulting in essentially the same amount of apoptosis as
transfection with GFP (Fig.
3A). However, when CTLs were
added to the cells 1 h prior to quantitation of apoptosis, both
constitutively active and dominant-negative mutants significantly altered the rate of cell death (Fig. 3A). The constitutively
active GTPases increased the ability of CTLs to induce apoptosis,
whereas dominant-negative proteins reduced killing by CTLs. Thus,
although each Rho family mutant was incapable of activating apoptosis
by itself at 24 h after transfection, each was able to strongly
modulate the induction of apoptotic cell death by CTLs.
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 3.
Effects of constitutively active and
dominant-negative Rho family GTPases on CTL- and Fas-induced
apoptosis. In A, target cells were microinjected with
vectors expressing constitutively active and dominant-negative
GFP-fused constructs of Rho GTPase mutants. The extent of apoptosis was
assessed 1 h after CTLs were added. In the absence of CTLs, the
GTPases produced only background levels of cell death, but the mutants
strongly affected the ability of CTLs to cause apoptosis. Error
bars represent S.E. values calculated for three independent
experiments. 21-285 cells (mean of 85) were assayed for each
experiment. B and C show effects of Rho family
GTPase mutants on Fas-induced death. These cells were transfected with
constitutively active and dominant-negative GFP fusion constructs of
Cdc42 (B) and Rho
(C). The extent of apoptosis at different
times after Fas stimulation is shown in B. C
shows the extent of apoptosis at the 2-h time point. Error
bars represent S.E. values calculated for three independent
experiments. 20-190 cells (mean of 69) were assayed for each
experiment. CHO cells were stably transfected with the extracellular
portion of CD4 fused to the cytoplasmic domain of murine Fas. Efficient
Fas-induced death was triggered upon incubation with anti-CD4
monoclonal antibody (39). dn, dominant-negative;
dp, dominant-positive.
|
|
We carried out similar experiments for apoptosis induced by Fas. Here
we used CHO cells transfected with a fusion protein consisting of the
cytoplasmic portion of the Fas receptor fused to the extracellular
domain of CD4. As shown previously, antibody cross-linking of the
extracellular portion of this molecule produces efficient killing by
the intracellular Fas death domain (39). As in CTL killing experiments,
transfection with activated and dominant-negative mutants of Rho and
Cdc42 alone had little effect, but the GTPases significantly altered
the rate of death produced by Fas (Fig. 3, B and
C).
Cytotoxic T cells kill targets using both Fas ligation and cytotoxic
granzyme protease injection (49). In our experiments using CTLs, it was
very unlikely that death was induced substantially by Fas ligation, as
cell death was assessed 1 h after addition of CTLs. Fas requires
3-4 h to begin producing appreciable amounts of apoptotic
morphological changes. A comparison of Fig. 3A, which shows
the extent of apoptosis 1 h after CTL addition, and Fig. 3B, which shows the kinetics of Fas-induced death, shows
this to be true. Previous work has also shown that the granule
exocytosis pathway predominates in 4-h CTL killing assays (49).
After determining that Rho protein activity was required for Fas- and
CTL-induced apoptosis, we asked whether Cdc42 was activated by Fas
ligand or whether the required level of activity was present prior to
stimulation. This was addressed by quantifying Fas-induced Cdc42
activation at various times following Fas stimulation
(61). Fig. 4 shows
that there was a significant increase in the formation of Cdc42-GTP 5 min after Fas stimulation, which then returned to base-line levels
within 1 h. This is consistent with active stimulation of Rho
family activity by Fas ligation. Activation of Rac with a similar time
course has been shown by Gulbins et al. (52).
View larger version (48K):
[in this window]
[in a new window]
|
Fig. 4.
Cdc42 activation in CHO cells after
stimulation of the Fas receptor. Cells were treated with Fas and
then lysed and incubated with a GST-tagged fragment of p21-activated
kinase that has been shown to bind Cdc42-GTP, but not Cdc42-GDP (61).
Horseradish peroxidase-conjugated anti-GST antibodies were used in the
Western blot shown to quantify the extent of Cdc42 activation at 5 and
60 min after Fas treatment. The first lane is a control
showing maximal activation, in which cell lysates were preloaded with
GTPS. Loading with GDP as a negative control is shown in the
second lane, followed by a time point immediately prior to
Fas addition (third lane) and at 5 (fourth lane)
and 60 (fifth lane) min after addition. It is clear that
Cdc42-GTP was elevated at the 5-min time point and that levels return
to prestimulation values by 60 min. The experiment was carried out
twice with similar results.
|
|
Rho Family Effects Are Upstream of Caspase-3 Cleavage--
The
apoptosis assays used in the experiments described above relied on
morphological changes and nuclear fragmentation as indicators of
apoptosis. It was possible that Rho proteins were altering only
terminal apoptotic events, rather than affecting the upstream
caspase-mediated cleavage cascades. Although it was unlikely that both
the nuclear and morphological changes we assayed required activation of
Rho GTPases, we sought to verify that the Rho GTPases were involved in
the induction phase of cell death. We therefore assayed the ability of
Clostridium toxin B to inhibit Fas-induced activation of
caspase-3. Activation of this caspase is a critical "commitment"
step that triggers multiple terminal pathways leading to morphological,
nuclear, and membrane changes in the apoptotic cell. Caspase-3
activation in cell lysates was determined using a fluorogenic substrate
assay of caspase-3 (47). Clostridium toxin B blocked
Fas-induced caspase-3 activation by >75% (Fig.
5), showing that Rho proteins were
involved in steps at or upstream of caspase activation. These
experiments clearly demonstrated a role for Rho family GTPases in
cellular commitment to apoptosis.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 5.
Rho family effects occur upstream of caspase
cleavage. The fluorogenic substrate
N-acetyl-Asp-Glu-Val-Asp-7-amino-4-methylcoumarin was used
to quantify caspase-3 activity in cell lysates 2 h after
stimulation of Fas-induced killing. Upon preincubation of cells with
C. difficile toxin B (100 ng/ml), there was a strong
inhibition of caspase-3 activity. Error bars represent S.E.
values calculated for three independent experiments.
|
|
We furthermore reasoned that if Rho proteins controlled only terminal
apoptotic changes, it was likely that cells could not be completely
rescued from apoptosis by inhibiting Rho protein activity. Given
sufficient time, terminal events not controlled by Rho GTPases would
likely kill the cells. We therefore extended the time of our CTL
killing assay from 1 to 12 h to see if Rho GTPase inhibition could
decrease cell death even at extended time periods. As shown in Fig.
6, 100% of the cells were killed by 12 h when they were not treated with toxin B. In contrast, for cells preincubated with toxin B, at 12 h, 40% of the cells
remained attached to the coverslip and showed no evidence of nuclear
fragmentation when stained with Hoechst dye. This further supported the
involvement of Rho proteins in apoptosis induction.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 6.
Effect of toxin B on the extent of
CTL-induced apoptosis at late time points. Toxin B strongly
affected the kinetics of CTL killing. For cells preincubated with toxin
B (100 ng/ml), at 12 h, 40% remained attached to the coverslip,
showing no sign of apoptosis. Apoptosis was assessed based on
morphological criteria and examination of nuclear fragmentation using
Hoechst staining. Error bars represent S.E. values
calculated for three independent experiments.
|
|
Regulation of Fas and CTL Killing by Rho Family Proteins Requires
the Actin Cytoskeleton--
Because the Rho proteins are potent
modulators of the actin cytoskeleton, we examined the possible role of
the cytoskeleton in Fas- and CTL-induced apoptosis. Target cells
were incubated with latrunculin A, a specific inhibitor of actin
polymerization (46), and the effect of this treatment on Fas-induced
caspase-3 activation was quantified using the fluorogenic substrate
assay described above. Latrunculin inhibited the ability of Fas to
activate caspase-3 nearly 60%, indicating that polymerized actin was
required for apoptosis induction (Fig.
7A). Perturbation of actin
with latrunculin also affected the ability of activated Rho family mutants to potentiate Fas- and CTL-induced apoptosis. This was seen in
fluorogenic substrate assays of Fas-induced caspase activation (Fig.
7A) and demonstrated for both CTL and Fas killing using microscope tracking of cells transfected with GFP-Rho mutants (as
described above) (Fig. 7B). Taken together, these
experiments show that apoptosis stimulated by CTLs and Fas requires
polymerized actin structures and that effects of Rho proteins on these
signaling pathways do not occur independently of GTPase signaling to
the actin cytoskeleton.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 7.
Induction of apoptosis by the Rho GTPases is
mediated by the actin cytoskeleton. The actin polymerization
inhibitor latrunculin A (2 µg/ml) (A) led
to a significant inhibition of Fas-induced caspase-3 activation. These
caspase activity assays were done 2 h after Fas induction.
Error bars represent S.E. values calculated for three
independent experiments. Latrunculin A blocked the ability of activated
Rho GTPase mutants to enhance CTL- and Fas-induced death. The rate of
apoptosis in the presence or absence of latrunculin A
(Latrunc; 2 µg/ml) is shown in B for cells
transfected with the constitutively active mutants. Error
bars represent S.E. values calculated for three independent
experiments. 20-150 cells (mean of 74) were assayed for each
experiment.
|
|
To examine whether the Rho proteins affected apoptosis by altering the
actin cytoskeleton, cells were transfected with dominant-positive Rac1
Q61L containing additional point mutations that affected interactions
with downstream targets mediating effects on actin. As shown in Fig.
8, the Y40C mutation, previously shown to
affect interactions with targets containing the Cdc42/Rac-interactive binding domain (60), had little effect on the action of
dominant-positive Rac to promote CTL killing. However,
dominant-positive Rac containing the additional F37A mutation, which
alters the ability of Rac to regulate cortical actin assembly
(60), completely eliminated the ability of dominant-positive
Rac to increase apoptotic response.
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 8.
Effector site mutations disrupt the ability
of Rac to increase actin polymerization. The ability of
dominant-positive (dp) Rac Q61L to increase apoptosis
induced by Fas was affected by some effector site mutations. Rac1 Q61L
with a Y40C (40C) substitution, previously shown to affect
the interaction of Rac with all targets containing
Cdc42/Rac-interactive binding domain motifs (60), did not alter the
extent of cell death produced by Rac1 Q61L. On the other hand, Rac1
Q61L with an F37A substitution (37A), which is unable to
induce actin polymerization at the plasma membrane (60), reduced
Fas-induced apoptosis to levels seen in the absence of activated
GTPase. These experiments were done 2 h after Fas induction.
Error bars represent S.E. values calculated for three
independent experiments.
|
|
Additional evidence that Rho family proteins influence apoptosis by
modulating the cytoskeleton was provided by the strong correlation
between Rho family effects on cell shape and susceptibility to
apoptosis. Cells treated with Clostridium toxin B retracted their edges and took on an altered morphology, with long cell extensions and a condensed cell body (Fig.
9A). This morphological change
was blocked by transfection with the constitutively active GTPases
(Fig. 9B). We examined the correlation between apoptosis and
cell morphology controlled by the competing actions of each GTPase and
toxin B. Individual cells were transfected with constitutively active
GTPase mutants and then treated with toxin B. Cells in which the
constitutively active protein caused well spread morphology were three
times as likely to undergo apoptosis, as is shown in Fig.
9C. Thus, the effects of Rho proteins on morphology strongly correlated with susceptibility to CTL-induced killing.
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 9.
Rho GTPase effects on apoptosis correlate
with effects on cell morphology. The fate of individual target
cells expressing the constitutively active Rho family GTPase mutants
correlated with their morphological features. The block in apoptosis
induced by toxin B was overcome by dominant-positive Rho family
mutants, but only when these mutants were expressed at sufficiently
high levels to prevent the altered cell morphology produced by the
toxin. A shows an example of a cell with the retracted
morphology typical of toxin B treatment. B shows a cell in
which this morphological change was blocked through prior transfection
with constitutively active Rac Q61L. A population of cells on the same
coverslip was transfected with dominant-positive GTPase protein and
then treated with toxin. Cells with or without the retracted morphology
were compared for susceptibility to apoptosis. C shows that
the well spread cells had higher levels of CTL-induced apoptosis than
those that had the altered morphology typical of toxin B
treatment.
|
|
| |
DISCUSSION |
Our results demonstrate that the Rho family proteins play a role
in the induction of apoptosis by CTLs and Fas. C. difficile toxin B, a specific inhibitor of all Rho proteins (48), blocks both
Fas- and CTL-induced apoptosis. Furthermore, transfection with mutants
of each family member affects the ability of Fas and CTLs to induce
apoptosis; activated mutants stimulate killing, and dominant-negative
mutants inhibit it. We also demonstrated that Fas stimulation increases
the level of Cdc42 activation within 5 min. Rho proteins are known to
control some aspects of apoptotic morphological changes (11), so we
were careful to establish that the GTPases were not affecting only late
stage apoptotic morphological events, but were in fact regulating the
induction of apoptosis. This was most conclusively demonstrated by
inhibiting Fas activation of caspase-3 using Clostridium
toxin B. Caspase-3 cleavage and activation are considered to be
critical steps in the commitment to apoptosis in Fas-induced cell
death (50). More important, we observed that both activated and
inhibitory GTPase mutants had no effect on cell death in the absence of
apoptotic stimuli, despite their strong influence on CTL- and
Fas-induced killing. In summary, these results demonstrate that the Rho
family GTPases regulate the induction of apoptosis produced by CTLs and Fas, but that activation alone is not sufficient to produce cell death
during the time required for killing by CTLs and Fas.
There have been previous reports indicating that Rac plays a role in
Fas-induced apoptosis. Rac antisense oligonucleotides and
dominant-negative Rac mutants have been shown to inhibit Fas-induced killing of T cells after 3 h (52). Rac activation by Fas ligand has been demonstrated (52). Dominant-negative Rac inhibited apoptosis
produced by tumor necrosis factor, whereas dominant-negative Rho had no
effect (53). In contrast to these and the present data, other studies
have shown that transfection with activated Cdc42 and Rac alone is
sufficient to induce apoptosis in some cells (16, 51). Ashkenazi and
Dixit (12) and Evan and Littlewood (12, 15) have pointed out that
apoptotic stimuli can be divided into two groups: perturbations of
normal homeostasis or stimuli that interact with cellular molecules
specifically evolved to induce apoptosis. Perturbations of homeostasis
such as growth factor withdrawal (13), loss of cell-matrix attachment
(14), and DNA damage (15) produce apoptosis much more slowly than direct interaction with apoptosis-inducing molecules such as CTL granzymes, Fas ligand, or tumor necrosis factor. This direct
stimulation of apoptotic pathways occurs in minutes to hours. Our
measurements of apoptosis due to activation of Rho proteins alone were
made at the short times required for CTL and Fas killing. Previous reports examined effects of activated GTPase mutants after 24 h or
longer. Thus, our data do not contradict earlier work, but instead
indicate that Rho family proteins affect direct induction of apoptosis
by mechanisms different from those in apoptosis produced by
perturbation of homeostasis.
In studies where apoptosis was produced through perturbation of normal
homeostasis, both inhibitory and stimulatory activities of Rho GTPases
have been reported. Activation of Rac and Rho inhibits apoptosis
produced by growth factor deprivation (17, 54), and Rho activation
inhibits apoptosis induced by DNA damage (55) and loss of cell adhesion
(18). Previously published experiments using activated or inhibitory
mutants of Rac, Rho, or Cdc42 indicate that these proteins potentiate
apoptosis produced by withdrawal of serum or nerve growth factor (16,
53). Mechanistic studies have revealed that Cdc42 and Rac induce
apoptosis through activation of JNK and/or p38 mitogen-activated
protein kinase signaling cascades (16, 51) and that inhibition by Rac
proceeds through activation of phosphatidylinositide 3'-hydroxykinase
and Akt (17).
The Rho family proteins control the actin cytoskeleton via multiple
downstream effectors, resulting in complex localized changes in actin
structures and focal adhesions (23, 31, 37). Our data show that the Rho
family proteins alter apoptotic signaling, but cannot themselves induce
apoptosis. This is consistent with a mechanism in which Rho proteins
alter the cytoskeletal scaffolding required for apoptosis induction
and/or affect pathways by which cytoskeletal changes induce apoptosis
(35, 36). This is in agreement with recent studies showing that
caspases assemble into filamentous cytoplasmic structures during
apoptosis (56). Furthermore, the extracellular matrix controls
apoptotic susceptibility by interacting with focal adhesion components.
Rho influences the number and nature of focal adhesions, and both Cdc42
and Rho can induce focal complexes and alter the filamentous actin
network anchored to focal adhesions.
We tested the involvement of actin in Rho family modulation of
apoptotic signaling. Latrunculin, an inhibitor of actin polymerization, strongly affected caspase activation by Fas. Recently published studies
have shown that jasplakinolide, a molecule that stabilizes polymerized
actin structures, increases apoptosis (57). Latrunculin blocked the
ability of constitutively active Rho family members to enhance CTL- and
Fas-induced killing. These experiments indicate that polymerized actin
structures are required for apoptosis induction and that Rho family
proteins do not operate on a pathway independent of their ability to
regulate the actin cytoskeleton. This evidence is consistent with a
mechanism in which Rho modulation of actin affects Fas and CTL
signaling, possibly through rearrangement of actin scaffolds,
alterations in focal adhesion signaling, or cytoskeletal changes
accompanying Fas receptor capping. These models are supported by our
experiments in which point mutations affecting the ability of Rac to
alter cortical actin interfered with its ability to increase apoptosis.
The facts that Cdc42 is stimulated rapidly, within 5 min of Fas
addition, and that simulation recedes well before caspase activation
suggest that Rho proteins participate in early events of
receptor-mediated signaling. Previous reports also show rapid
activation of Rac by Fas (52).
Our data show a strong correlation between morphological features
produced by Rho proteins and susceptibility to apoptosis. This
further supports a mechanism in which the Rho GTPases affect apoptosis
by modulating the cytoskeleton. Our observations are consistent with
previous studies demonstrating a connection between cell shape and
susceptibility to apoptosis. Ingber and co-workers (58, 59) have
elucidated mechanisms by which mechanical stresses control cell death.
It was shown that cell shape strongly influences the extent of
apoptosis regardless of the total area of contact with cell-surface
molecules. Using a system distinct from that utilized in our studies,
these workers observed that spread cells were less susceptible to
apoptosis. This apparent contradiction suggests that apoptosis requires
specific cytoskeletal structures whose formation does not correlate
with gross morphological features like the extent of cell spreading.
The work described here further dissects the complex multifunctional
role of the Rho proteins in apoptosis induction. The mechanisms by
which cytoskeletal dynamics impact upon cell signaling remain poorly
understood. In future studies, we hope to identify specific actin
structures and dynamics involved in apoptosis induction and elucidate
their control by Rho family signaling.
,
,
TOP
ABSTRACT
INTRODUCTION
CiteULike 
Complore 
Connotea 
Del.icio.us 
