Garnier P, Ying WH, Swanson RAIschemic preconditioning by caspase cleavage of poly(ADP-ribose) polymerase-1. J Neurosci 23:7967-7973
A transient, sublethal ischemic interval confers resistance to a subsequent, otherwise lethal ischemic insult, in a process termed ischemic preconditioning. Poly(ADP-ribose) polymerase-1 (PARP-1) normally functions in DNA repair, but extensive PARP-1 activation is a major cause of ischemic cell death. Because PARP-1 can be cleaved and inactivated by caspases, we investigated the possibility that caspase cleavage of PARP-1 could contribute to ischemic preconditioning. Murine cortical cultures were treated with glucose deprivation combined with 0.5 mm 2-deoxyglucose and 5 mm azide ("chemical ischemia") to model the reversible energy failure that occurs during transient ischemia in vivo. Cortical cultures preconditioned with 15 min of chemical ischemia showed increased resistance to subsequent, longer periods of chemical ischemia. These cultures were also more resistant to the PARP-1 activating agent, N-methyl-N'-nitro-N-nitrosoguanidine, suggesting reduced capacity for PARP-1 activation after preconditioning. Immunostaining for the 89 kDa PARP-1 cleavage fragment and for poly(ADP-ribose) formation confirmed that PARP-1 was cleaved and PARP-1 activity was attenuated in the preconditioned neurons. Preconditioning also produced an increase in activated caspase-3 peptide and an increase in caspase-3 activity in the cortical cultures. A cause-effect relationship between caspase activation, PARP-1 cleavage, and ischemic preconditioning was supported by studies using the caspase inhibitor Ac-Asp-Glu-Val-Asp-aldehyde (DEVD-CHO). Cultures treated with DEVD-CHO after preconditioning showed reduced PARP-1 cleavage and reduced resistance to subsequent ischemia. These findings suggest a novel interaction between the caspase- and PARP-1-mediated cell death pathways in which sublethal caspase activation leads to PARP-1 cleavage, thereby increasing resistance to subsequent ischemic stress.
Ischemic Preconditioning by Caspase Cleavage of
Philippe Garnier, Weihai Ying, and Raymond A. Swanson
Department of Neurology, University of California at San Francisco and Veterans Affairs Medical Center, San Francisco, California 94121
A transient, sublethal ischemic interval confers resistance to a subsequent, otherwise lethal ischemic insult, in a process termed ischemic
preconditioning. Poly(ADP-ribose) polymerase-1 (PARP-1) normally functions in DNA repair, but extensive PARP-1 activation is a
major cause of ischemic cell death. Because PARP-1 can be cleaved and inactivated by caspases, we investigated the possibility that
caspase cleavage ofPARP-1 could contribute to ischemic preconditioning. Murine cortical cultures were treated with glucose deprivation
combined with 0.5 mM 2-deoxyglucose and 5 mM azide (“chemical ischemia”) to model the reversible energy failure that occurs during
transient ischemia in vivo. Cortical cultures preconditioned with 15 min of chemical ischemia showed increased resistance to subsequent,
longer periods of chemical ischemia. These cultures were also more resistant to the PARP-1 activating agent, N-methyl-N⬘-nitro-N-
nitrosoguanidine, suggesting reduced capacity for PARP-1 activation after preconditioning. Immunostaining for the 89 kDa PARP-1
cleavage fragment and for poly(ADP-ribose) formation confirmed that PARP-1 was cleaved and PARP-1 activity was attenuated in the
preconditioned neurons. Preconditioning also produced an increase in activated caspase-3 peptide and an increase in caspase-3 activity
in the cortical cultures. A cause– effect relationship between caspase activation, PARP-1 cleavage, and ischemic preconditioning was
supported by studies using the caspase inhibitor Ac-Asp-Glu-Val-Asp-aldehyde (DEVD-CHO). Cultures treated with DEVD-CHO after
preconditioning showed reduced PARP-1 cleavage and reduced resistance to subsequent ischemia. These findings suggest a novel
interaction between the caspase- and PARP-1-mediated cell death pathways in which sublethal caspase activation leads to PARP-1
cleavage, thereby increasing resistance to subsequent ischemic stress.
Key words: apoptosis; brain; culture; cycloheximide; mouse; tolerance
A brief period of sublethal ischemia induces resistance to a sub-
sequent, otherwise lethal ischemic insult in a process termed isch-
emic preconditioning (Nandagopal et al., 2001; Kirino, 2002;
Dirnagl et al., 2003). Several mechanisms have been implicated in
ischemic preconditioning, and the relative importance of these
mechanisms differs among different cell types (Chen and Simon,
1997; Ishida et al., 1997; Nandagopal et al., 2001; Kirino, 2002).
Studies in brain suggest involvement of post-translational
changes such as Ras activation (Gonzalez-Zulueta et al., 2000),
altered phosphorylation state of ERK (extracellular signal-
regulated kinase) and NMDA receptor subunits (Shamloo and
Wieloch, 1999; Shamloo et al., 1999), and altered NMDA recep-
tor function (Aizenman et al., 2000). Other studies suggest im-
portant roles for gene transcription and increased expression of
proteins such as hsp70 (Chen et al., 1996; Currie et al., 2000),
anti-oxidant enzymes (Hoshida et al., 2002), or anti-apoptosis
factors (Shimazaki et al., 1994; Shimizu et al., 2001). In only a few
instances, however, has it been possible to establish a cause–effect
relationship between changes induced by ischemic precondition-
ing and increased resistance to subsequent stress.
Activation of poly(ADP-ribose) polymerase-1 (PARP-1) is a
major cause of neuronal death after brain ischemia (Szabo and
Dawson, 1998). Poly(ADP-ribose) polymerases are enzymes that
transfer the ADP-ribose groups from NAD
to form branched
ADP-ribose polymers on acceptor proteins in the vicinity of DNA
strand breaks or kinks. Several PARP family members are now
recognized, with PARP-1 accounting for ⬎80% of nuclear
poly(ADP-ribose) polymerase activity (Heller et al., 1995;
D’Amours et al., 1999). Although PARP-1 normally functions to
facilitate DNA repair, extensive PARP-1 activation promotes cell
death through processes involving energy depletion and the re-
lease of apoptosis-inducing factor (Ha and Snyder, 1999; Ying et
al., 2002; Yu et al., 2002). Genetic or pharmacological inhibition
of PARP-1 activity reduces infarct size by up to 80% in brains
subjected to transient or permanent ischemia (Eliasson et al.,
1997; Endres et al., 1997; Tokime et al., 1998; Takahashi et al.,
1999). Similarly, inhibition of PARP-1 in cultured neurons sub-
stantially increases resistance to oxygen– glucose deprivation and
to NMDA toxicity (Cosi et al., 1994; Ha and Snyder, 1999; Ying et
PARP-1 can be cleaved and inactivated by caspase-3,
caspase-7, caspase-8, and calpains (D’Amours et al., 1998; Ger-
main et al., 1999; Wang, 2000). Because caspases and calpains are
activated by cerebral ischemia (Lipton, 1999; Chen et al., 2002),
Received Feb. 4, 2003; revised June 24, 2003; accepted July 10, 2003.
This work was supported by the Department of Veterans Affairs and by National Institutes of Health Grants
NS41421 (R.A.S.)and NS11048(W.Y.). Wethank Drs.Z. Q.Wang andB. Zingarellifor thePARP-1
and Elizabeth Gum for technical assistance.
Correspondenceshould be addressedtoRaymondA.Swanson,(127)Neurology,Veterans AffairsMedicalCenter,
4150 Clement Street, San Francisco, CA 94121. E-mail: email@example.com.
Copyright © 2003 Society for Neuroscience 0270-6474/03/237967-07$15.00/0
The Journal of Neuroscience, September 3, 2003 • 23(22):7967–7973 • 7967
we hypothesized that cleavage and inactivation of PARP-1 could
be an effector mechanism of ischemic preconditioning.
Caspase-3 can be activated by low levels of oxidative stress or
excitotoxicity (Bonfoco et al., 1995). Moreover, caspase-3 activa-
tion has recently been shown to be important for ischemic pre-
conditioning in brain (McLaughlin et al., 2003). Here we show
that sublethal activation of caspase-3 in cultured neurons by a
preconditioning ischemic stimulus leads to PARP-1 cleavage, and
this renders neurons resistant to insults, such as subsequent isch-
emia, that would otherwise induce PARP-1-mediated cell death.
Materials and Methods
Materials. Chemicals were purchased from Sigma-Aldrich (St. Louis,
MO), except where noted otherwise.
Astrocyte–neuron cocultures. The animal use protocol was approved by
the Animal Studies Committee of the San Francisco Veterans Affairs
Medical Center. Cortical astrocyte–neuron cocultures were prepared by
seeding neurons onto a preexisting astrocyte layer, as described previ-
ously (Ying et al., 1999). Wild-type (WT) neurons were prepared from
fetal Swiss–Webster mice (Simonsen, Gilroy, CA). PARP-1 gene defi-
) neurons were prepared from fetal mice of the inbred
strain developed by Wang et al. (1995) on a mixed 129/Sv x
C57BL/6 background. Cocultures were maintained in glial-conditioned
medium prepared by placing MEM with 2 m
M glutamine, 50
streptomycin, and 2.5% FBS supplemented with 100 n
M sodium selenate
and 200 n
-tocopherol (Leist et al., 1996) into a flask of confluent
cortical astrocytes for 72 hr. Experiments were conducted when neurons
were 14 –15 d in vitro.
Experimental procedures. Experiments were performed using a bal-
anced salt solution (BSS) composed of (in m
M): 3.1 KCl, 134 NaCl, 1.2
, 1.2 MgSO
, 0.25 KH
, 15.7 NaHCO
, as described previously
(Ying et al., 2001). The pH was adjusted to 7.2 while the solution was
equilibrated with 5% CO
at 37°C. Osmolarity was verified at 280 –310
mOsm with a Wescor vapor pressure osmometer (Logan, UT). Concen-
trated drug stocks were prepared in BSS and likewise adjusted to pH 7.2
and 280 –320 mOsm. Chemical ischemia (CI) was induced by incubating
the cultures in glucose-free BSS containing 0.5 m
M 2-deoxyglucose and 5
M sodium azide in a 37°C, 5% CO
incubator (Swanson and Benington,
1996) and terminated by washing in BSS and replacement of glial-
Assessment of neuronal injury. Neurons were distinguished from the
underlying astrocyte layer by their phase-bright, process-bearing mor-
phology (Ying et al., 1999). Dead neurons were identified 24, 48, or 72 hr
after chemical ischemia exposures by propidium iodide (PI) fluores-
cence. PI was added at 0.04 mg/ml to each well, and both the PI-
fluorescing dead neurons and nonfluorescing live neurons were counted
in four randomly selected optical fields using a Nikon fluorescence mi-
croscope. At least 300 neurons were counted in each well, and results
from each well were expressed as “% neuronal death.”
Immunostaining. Immunostaining for poly(ADP-ribose) (PAR) was
performed according to the method of Burkle et al. (1993) with modifi-
cations. After one wash in cold PBS, the cultures were fixed in 10%
trichloroacetic acid on ice for 15 min, dehydrated by sequential washes in
70, 90, and 100% ethanol, and air dried. The fixed cultures were prein-
cubated with blocking buffer (PBS), 10% goat serum, 0.1% Triton
X-100) for 60 min at room temperature, and anti-PAR monoclonal an-
tibody (Trevigen, Gaithersburg, MD) was added at a 1:2000 dilution for
incubation overnight at 4°C. After washing with PBS containing 0.1%
Triton X-100, the cells were incubated with Alexa-fluor 488-conjugated
goat anti-mouse IgG (Molecular Probes, Eugene, OR) at 1:500 dilution
for 1 hr at room temperature and then washed again to remove excess
antibody. Immunostaining for the 89 kDa PARP-1 cleavage product was
performed in cultures fixed in 4% paraformaldehyde for 45 min at room
temperature. After preincubation in blocking buffer, rabbit polyclonal
anti-cleaved PARP-1 antibody (#9544, Cell Signaling Technology, Bev-
erly, MA) was added at a 1:500 dilution and incubated overnight at 4°C.
Antibody visualization was achieved with Alexa-fluor 488-conjugated
goat anti-mouse IgG.
Western blots. Cells were lysed at 4°C in buffer containing 1% SDS, 1
M Na-vanadate, 0.1 mM phenylmethylsulfonylfluoride, 2.5
g/ml aprotinin, 5
g/ml leupeptin, and 10 mM Tris-HCl, final
pH 7.4. Viscosity of the samples was reduced by brief sonication. Aliquots
g of protein were mixed with Laemmli buffer, boiled for
5 min, and electrophoresed on a SDS-polyacrylamide gel. Proteins were
transferred to a polyvinylidene difluoride membrane (Amersham Bio-
sciences, Piscataway, NJ), incubated overnight at 4°C in Tris-buffered
saline containing 5% nonfat dry milk and 0.1% Tween 20, and then
incubated for 4 hr at room temperature with the primary antibody. After
three washes, the membrane was incubated for 90 min with a 1:10,000
dilution of anti-mouse or anti-rabbit IgG horseradish peroxidase anti-
body (Vector Laboratories, Burlingame, CA). The membrane was
washed three additional times, and the signal was detected with the Am-
ersham ECL system (Amersham Biosciences). After antibody stripping,
membranes were immunostained for
-actin to confirm consistent pro-
tein loading in each lane. The primary antibody dilutions were as follows:
1:1500 for rabbit polyclonal antibody to PARP-1 (#9544, Cell Signaling
Technology); 1:1000 for rabbit polyclonal antibody to cleaved caspase-3
(#9661, Cell Signaling Technology); and 1:10,000 for the mouse mono-
clonal antibody to
-actin (Sigma-Aldrich). Bands were quantified using
the SCION Image system. The analysis of
-actin expression showed
negligible variation in protein loading.
Caspase-3 activity. Caspase-3 activity was measured as described by
Sordet et al. (2002), with minor modifications. After lysing, cell samples
were centrifuged at 10,000 ⫻ g for 20 min. The protein concentration of
the resulting supernatant was measured by the bicinchonic acid method
(Smith et al., 1985), and 25
g protein aliquots were loaded into 96-well
plates. A reaction mixture containing 100
M fluorogenic peptide sub-
strate Ac-Asp-Glu-Val-Asp-7-amino-4-methyl coumarin (Calbiochem,
San Diego, CA), 1 m
M EDTA, 0.1% 3-[(3-cholamidopropyl)D:methyl
amino]-1-propane sulfonate, 10% glycerol, 20 m
M dithiothreitol, and
M HEPES, final pH 7.0 was added to the wells. After 1 hr at 37°C,
fluorescence was measured using an F
fluorescence plate reader with
Softmax Pro software (Molecular Devices, Sunnyvale, CA) at an excita-
tion wavelength of 355 nm and an emission wavelength of 460 nm.
Fluorescence values were expressed as increase over control and normal-
ized to the increase induced by staurosporine.
Caspase inhibition. Effects of caspase inhibition on PARP-1 cleavage
and ischemic preconditioning were tested by adding 25
M of Ac-Asp-
Glu-Val-Asp-aldehyde (DEVD-CHO) (Calbiochem) to the culture me-
dium 2 hr before the preconditioning and during the subsequent 24 hr.
Control experiments were performed in parallel by adding only the ve-
hicle, 0.25% dimethylsulfoxide, to the culture medium.
Statistical analyses. Data are presented as means ⫾ SE. Statistical sig-
nificance was assessed using ANOVA and the Student–Newman–Keuls
post hoc test to compare the indicated experimental groups.
A model of ischemic preconditioning in mouse cortical cultures
was established using inhibitors of energy metabolism (Rajdev
and Reynolds, 1994; Aizenman et al., 2000). Oxidative ATP pro-
duction was blocked with 5 m
M azide, and glycolytic ATP pro-
duction was blocked by removing glucose and adding 0.5 m
2-deoxyglucose to prevent metabolism of retained intracellular
glucose (Swanson and Benington, 1996). A 15 min period of CI
was sublethal, resulting in neuronal death no greater than that
caused by medium exchanges alone. In contrast, CI of ⱖ30 min
killed ⬎75% of the neurons (Fig. 1A,B). Cell death was assessed
24 hr after CI (Fig. 1B). The neuronal death resulting from 30 or
45 min of CI was substantially reduced in cultures that had been
preconditioned with a sublethal, 15 min interval of CI 24 hr
previously. Assessment of neuronal death at 48 or 72 hr after CI
gave nearly identical results. This model of ischemic precondi-
tioning was used for all subsequent studies in this report.
Previous studies have reported that ischemic preconditioning
requires several hours or days to become fully manifest (Kirino,
7968 • J. Neurosci., September 3, 2003 • 23(22):7967–7973 Garnier et al. •Ischemic Preconditioning by Caspase Cleavage of Poly(ADP-Ribose) Polymerase-1
2002; Dirnagl et al., 2003). Consistent with these reports, we
found that significant resistance to subsequent ischemia in this
model first developed between 3 and 6 hr after the precondition-
ing ischemia and increased through at least 24 hr (Fig. 1C). As
shown in Figure 1D, the preconditioning effect was substantially
reduced in cultures treated with the protein synthesis inhibitor
cycloheximide (CHX). Cycloheximide alone (in the absence of
preconditioning) also had a modest neuroprotective effect, as
reported previously (Lobner and Choi, 1996).
Ischemic preconditioning attenuates PARP-1-mediated
Neuronal death caused by CI was substantially reduced by the
PARP inhibitor 3,4-dihydro-5-[4-(1-piperidinyl)butoxy]-1(2H)-
isoquinolinone (DPQ) and by PARP-1 genetic deletion (Fig. 2), con-
sistent with a major role for PARP-1 activation in ischemic neuronal
death. These results suggested the possibility that the resistance to
ischemia conferred by ischemic preconditioning might likewise re-
sult from a block in PARP-1 activation. We tested this possibility by
examining the effect of ischemic preconditioning on PARP-1 activa-
tion and neuronal death induced by N-methyl-N⬘-nitro-N-
nitrosoguanidine (MNNG), a DNA alkylating agent widely used to
trigger PARP-1 activation and PARP-1-mediated cell death (Virag
and Szabo, 2002; Ying et al., 2002; Yu et al., 2002). MNNG-induced
neuronal death was substantially reduced by ischemic precondition-
ing (Fig. 3A), as well as by the PARP inhibitor
DPQ (data not shown). PARP-1 activation
was also assessed by immunostaining for
poly(ADP-ribose), the enzymatic product of
PARP-1. In the absence of preconditioning,
poly(ADP-ribose) immunoreactivity was
evident in most neurons after 40 min of ei-
ther CI or MNNG. These treatments pro-
duced almost no detectable poly(ADP-
ribose) in cultures that had been treated with
ischemic preconditioning, suggesting a re-
duced capacity for PARP-1 activation in the
preconditioned neurons. The caspase inhib-
itor DEVD-CHO negated this effect of pre-
conditioning, whereas CHX had no appre-
ciable effect (Fig. 3B).
Ischemic preconditioning causes
caspase activation and PARP-1 cleavage
Because PARP-1 can be irreversibly inacti-
vated by caspase cleavage, we examined
whether this process could account for the
reduced PARP-1 activity and PARP-
mediated cell death observed in precondi-
tioned cultures. Cultures were harvested
24 hr after ischemic preconditioning and
assessed by Western blots and immunocy-
tochemistry for the 89 kDa PARP-1 frag-
ment produced by caspase cleavage (La-
zebnik et al., 1994; Le et al., 2002). As
shown in Figure 4 A, ischemic precondi-
tioning caused a significant increase in the
89 kDa fragment, and this increase was at-
tenuated in cultures treated with the
caspase inhibitor, DEVD-CHO. Cultures
immunostained with the same antibody
used in the Western blots showed the pres-
ence of PARP-1 cleavage product in most
neurons (and no astrocytes) after preconditioning (Fig. 4 B).
Consistent with the Western blots, appearance of the 89 kDa
PARP-1 cleavage product was significantly reduced in cultures
treated with DEVD-CHO.
Caspase-3 is the major effector of PARP-1 cleavage during
apoptosis (Lazebnik et al., 1994; Le et al., 2002). Procaspase-3
becomes activated by cleavage to a 17 kDa fragment (Cohen,
1997). To assess procaspase-3 cleavage in preconditioning, West-
ern blots were prepared at serial time points after precondition-
ing and probed with an antibody specific to the active 17 kDa
caspase-3 fragment. These studies showed an increase in active
caspase-3 fragment after an interval of 3–6 hr (Fig. 5A). We also
assessed the time course of PARP-1 cleavage. Western blots for
the 89 kDa PARP-1 cleavage product showed a significant in-
crease by 12 hr after preconditioning ischemia (Fig. 5A). Direct
measures of caspase-3 enzymatic activity in lysed cultures showed
a pattern similar to that observed in the Western blots (Fig. 5B).
An increase in caspase-3 enzymatic activation was significant by 3
hr and maximal by 6 hr after preconditioning ischemia. The
preconditioning-induced caspase-3 activity was blocked by the
caspase inhibitor DEVD-CHO, whereas CHX had no effect. Of
note, the magnitude of caspase-3 activation was modest relative
to that induced by staurosporine, a classic inducer of apoptosis
(Fig. 5B). Staurosporine caused extensive neuronal death at 24 hr
(data not shown).
Figure 1. Preconditioning (PC) increases neuronal resistance to subsequentchemical ischemia. Cultures werepreconditioned
with15 minof chemicalischemia (15min PC)or shamwashesonly (0min PC).Twenty-fourhours later,the cultureswere exposed
to 0, 30, or 45 min of CI. Cell death was assessed by propidium iodide fluorescence 24 hr after the CI. A, Representative cell fields
photographed with phase-contrast and epifluorescence optics show normal, phase-bright neurons and dead neurons with nu-
clear propidium iodide fluorescence. Scale bar, 100
m. B, Effect of preconditioning on neuronal resistance to CI. C, Neuronal
resistance to CI at sequential time points after a 15 min ischemic preconditioning. D, Effects of CHX on ischemic preconditioning.
Cultures wereincubated with 10
g/ml CHXfor the 24hr interval betweenthe 15 minischemic preconditioning andthe onset of
chemical ischemia. *p ⬍ 0.05; **p ⬍ 0.01; representative of five independent experiments each with n ⫽ 7–11.
Garnier et al. •Ischemic Preconditioning by Caspase Cleavage of Poly(ADP-Ribose) Polymerase-1 J. Neurosci., September 3, 2003 • 23(22):7967–7973 • 7969
Caspase inhibition attenuates ischemic preconditioning
To further evaluate the role of caspase activation in ischemic
preconditioning, cultures were incubated with DEVD-CHO for a
24 hr interval after the preconditioning stimulus. As shown in
Figure 6, cultures treated with 25
M DEVD-CHO acquired less
resistance to a second, longer ischemic challenge, whereas the 24
hr incubation with DEVD-CHO did not itself affect cell viability.
DEVD-CHO reduced the extent of the preconditioning-induced
resistance by ⬃40% when tested with a 30 min ischemic challenge
and by ⬃100% when tested with a 45 min ischemic challenge.
These differences cannot be compared directly because the de-
gree of preconditioning-induced resistance also differed at the
two ischemic time points in the absence of DEVD-CHO, but they
suggest that the relative contribution of caspase-mediated pro-
cesses to ischemic tolerance may vary with the severity of the
A major role for PARP-1 activation in ischemic neuronal death is
well established (Virag and Szabo, 2002). The increased resis-
tance to ischemic insults resulting from genetic or pharmacolog-
ical inhibition of PARP-1 activity, as confirmed in the cell culture
model used here, led us to propose that a block in PARP-1 acti-
vation might similarly contribute to the protective effect of isch-
emic preconditioning. This proposal was supported by the find-
ing that preconditioning blocks the poly(ADP-ribose) formation
and cell death that otherwise result from exposure to the PARP-1
activating agent MNNG. PARP-1 can be irreversibly inactivated
by cleavage of the catalytic site from the DNA binding domain,
through the actions of caspase-3 and other proteolytic enzymes
(Pieper et al., 1999). The present studies provide evidence for
both PARP-1 cleavage and caspase-3 activation in precondi-
tioned neurons. Additionally, treatment with the caspase inhibi-
tor DEVD-CHO blocked PARP-1 cleavage and reduced the de-
velopment of ischemic tolerance. Taken together, these results
suggest that preconditioning provides increased resistance to
ischemia at least in part by inducing caspase-mediated PARP-1
cleavage and thereby blocking the PARP-1-mediated cell death
Ischemic tolerance developed over a period of 6 –24 hr after
the preconditioning stimulus, similar to observations made in
vivo and in other cell culture systems (Nandagopal et al., 2001;
Kirino, 2002; Dirnagl et al., 2003). This time course can be com-
pared with the time course of procaspase-3 cleavage, increased
caspase enzymatic activity, and PARP-1 cleavage presented in
Figure 5. Procaspase-3 cleavage and caspase activity both reached
maximum values at ⬃6 hr after ischemic preconditioning, and
PARP-1 cleavage reached a maximum ⬃12 hr after precondi-
tioning. These events correlate well with the development of isch-
emic tolerance, which first became apparent between 3 and 6 hr
after preconditioning and neared a maximum at 12 hr after precon-
ditioning. These time course studies suggest that procaspase-3 cleav-
age, caspase activation, and PARP-1 cleavage occur sequentially and
precede the development of ischemic tolerance. A causative link be-
tween these events is supported by the effects of the caspase inhibitor
DEVD-CHO. Cultures incubated with DEVD-CHO after ischemic
preconditioning showed no inactivation of PARP-1 activity, as
judged by lack of poly(ADP-ribose) formation in response to CI or
Figure 2. Ischemic neuronal deathis mediated by PARP-1 activation.PARP-1
and WT neurons in cultures pretreated with the PARP inhibitor DPQ showed markedly reduced
neuronal death after CI. **p ⬍ 0.01; representative of three independent experiments each
with n ⫽ 3– 4.
Figure 3. Preconditioning increasesneuronal resistanceto PARP-1-mediatedcell deathand
blocks PARP-1 activation. Cultures were treated with 15 min of ischemic preconditioning (15
min PC)or sham washes only(0 min PC).Twenty-four hours later, the cultures wereexposed to
MNNG to induce PARP-1 activation. A, Neuronal death assessed 24 hr after MNNG exposures.
**p ⬍ 0.01; representativeof threeindependent experimentseach withn ⫽ 3– 4.B, Effectsof
PC on immunostaining for poly(ADP-ribose), the enzymatic product of PARP-1. The poly(ADP-
ribose) immunoreactivity induced by either MNNG or CI was confined to neurons and attenu-
ated in preconditioned cultures. The effect of preconditioning was negated by the addition of
M) butnot by CHX (10
g/ml) addedimmediately after PC. Cells were fixed40
min after onset of CI or MNNG incubation. Similar results were observed in four independent
experiments. Scale bar, 25
7970 • J. Neurosci., September 3, 2003 • 23(22):7967–7973 Garnier et al. •Ischemic Preconditioning by Caspase Cleavage of Poly(ADP-Ribose) Polymerase-1
MNNG, and showed an attenuated resistance to a subsequent isch-
The extent of caspase-3 activation induced by ischemic pre-
conditioning was small relative to that induced by staurosporine.
The relatively modest activation of caspase-3 is consistent with
the observation that preconditioning, unlike staurosporine, did
not lead to neuronal death. Although caspase-3 activation is an
effector mechanism of cell death during apoptosis, caspases can
also be activated in processes that do not lead to cell death. For
example, caspase-3 cleavage of the transcription factor GATA-1
regulates differentiation of erythroid cells (De Maria et al., 1999).
Reversible, sublethal activation of caspase-3 has been demon-
strated previously in neurons (Francois et al., 2001), and
caspase-3 cleavage of glutamate receptors has been proposed as a
mechanism for synaptic plasticity (Glazner et al., 2000). The
present results suggest that PARP-1 cleavage is an additional
mechanism by which caspase-3 may act outside of its classical
role as an effector of cell death.
A comparison of Figure 1C and Figure 2 shows that the degree
of neuroprotection achieved by ischemic preconditioning was
substantial, approaching that achieved with the PARP inhibitor
DPQ and with PARP-1 gene deletion. If the neuroprotective ef-
fect of preconditioning effect is caused primarily by PARP-1
cleavage, a substantial fraction of neuronal PARP-1 must be
cleaved after preconditioning. It is difficult to quantify the frac-
tion of neuronal PARP-1 cleaved after preconditioning because
of the presence of astrocytes in the cocultures. However, the stud-
ies of PARP-1 function showed that preconditioning markedly
decreased poly(ADP-ribose) formation in both MNNG- and CI-
treated neurons (Fig. 4B), suggesting that substantial PARP-1
inactivation does occur after ischemic preconditioning.
Modest ischemic insults can trigger apoptosis (Bonfoco et al.,
1995), and it is likely that the sublethal insults that induce pre-
conditioning activate the caspase cascade by similar mechanisms.
Several studies support NMDA receptor activation and nitric ox-
ide production as important upstream components of neuronal
preconditioning (Grabb and Choi, 1999; Gonzalez-Zulueta et al.,
2000; Nandagopal et al., 2001). Although not investigated here,
these upstream events may be coupled to caspase-3 activation in
several ways. One well established sequence involves mitochon-
drial depolarization and cytochrome c release, which triggers the
intrinsic pathway of caspase activation (Lemasters et al., 1999;
Kroemer and Reed, 2000). Evidence also suggests that the Fas
pathway of caspase activation can be induced by transient ATP
depletion (Feldenberg et al., 1999) and that NMDA receptor ac-
tivation and nitric oxide production can trigger caspase activa-
tion through the endoplasmic reticulum stress response (Rao et
It is generally accepted that delayed-onset ischemic tolerance
requires de novo protein synthesis (Kirino, 2002). Consistent with
this, we observed that ischemic tolerance was substantially re-
duced in cultures treated with the protein synthesis inhibitor
CHX. Because caspase-3 is abundant in cells in an inactive form,
it is unlikely that de novo synthesis of caspase-3 is required for
PARP-1 cleavage. Accordingly, we observed no effect of CHX
treatment on the preconditioning-induced caspase-3 activation
or PARP-1 inhibition. This suggests the possibility that precon-
ditioning may induce synthesis of proteins that influence the
PARP-1 cell death pathway at steps downstream of PARP-1 acti-
vation. Alternatively, CHX may have cytoprotective effects
through mechanisms independent of protein synthesis inhibi-
Figure4. Preconditioningproduces caspase-mediatedcleavageof PARP-1.A, Westernblots
with antibody tothe 89 kDa PARP-1 fragment were prepared fromcultures 24 hr after sham or
15 min PC. The increase produced by preconditioning was attenuated in cultures treated with
the caspase inhibitor DEVD-CHO (25
M). Values above the bands denote the increase ⫾ SE
relative to the paired control (**p ⬍ 0.01; n ⫽ 3). B, Immunostaining for the 89 kDa PARP-1
cleavage product showed it to accumulate in neurons after preconditioning. The accumulation
was almost totally blocked by 25
M DEVD-CHO, as observed in the Western blots. Similar
results were observed in three independent experiments. Scale bar, 100
Figure5. Timecourse ofcaspase-3 activationand PARP-1cleavageafter preconditioning.A,
Westernblots withantibodyto thecleaved(activated) 17kDacaspase-3 fragmentorthe 89kDa
(inactivated) PARP-1 fragment were prepared from cultures 0 –24 hr after a sham or 15 min
ischemicpreconditioning. Valuesabovethe bandsdenotethe foldincrease⫾ SE,relative tothe
0 hr time point (*p ⬍ 0.05; **p ⬍ 0.01; n ⫽ 3– 4). B, Caspase-3 enzymatic activity was
measured in cultures lysed at sequential time points after sham washes only (0 min PC) or 15
min of ischemic preconditioning (15 min PC). PC produced a significant increase in caspase-3
within 3 hr of PC. The increase was blocked in cultures treated with the caspase inhibitor
DEVD-CHO after PC, but not in cultures treated with CHX. Staurosporine (150 n
M) was used as a
positive control. Fluorescence values are expressed as increase over control, normalized to the
increase induced by staurosporine. *p ⬍ 0.05; **p ⬍ 0.01; representative of three indepen-
dent experiments each with n ⫽ 3– 4.
Garnier et al. •Ischemic Preconditioning by Caspase Cleavage of Poly(ADP-Ribose) Polymerase-1 J. Neurosci., September 3, 2003 • 23(22):7967–7973 • 7971
tion, as has been suggested by previous studies (Ratan et al.,
During apoptosis, PARP-1 cleavage by caspase-3 and other
enzymes prevents the PARP-1 activation and resultant ATP de-
pletion that would otherwise be triggered during DNA fragmen-
tation (Oliver et al., 1998; Boulares et al., 1999; Le et al., 2002).
Support for caspase-3 as a mediator of PARP-1 cleavage after
ischemic preconditioning was provided in the present study by
biochemical measurements showing increased caspase-3 activity
in cell lysates, by Western blots showing the active caspase-3 frag-
ment in preconditioned cultures, and by the block of PARP-1
cleavage produced by 25
M DEVD-CHO, a concentration at
which DEVD-CHO is a relatively specific caspase-3 inhibitor
(Margolin et al., 1997). These findings are in agreement with a
study published during the preparation of this manuscript that
reports widespread caspase-3 activation in a rat brain model of
ischemic preconditioning (McLaughlin et al., 2003). It is possi-
ble, however, that other proteolytic enzymes may also be in-
volved. PARP-1 can also be cleaved by caspase-7, caspase-8, and
calpains (Germain et al., 1999; Wang, 2000; Le et al., 2002), and
because DEVD-CHO is not entirely specific for caspase-3, the
results obtained with DEVD-CHO do not exclude the possibility
that other caspases or calpains may contribute to PARP-1 inacti-
vation in ischemic preconditioning. Conversely, the possibility
that DEVD-CHO does not entirely block PARP-1 cleavage means
that the observed effect of DEVD-CHO on ischemic precondi-
tioning may underestimate the true contribution of PARP-1
cleavage to this process.
Aizenman E, Sinor JD, Brimecombe JC, Herin GA (2000) Alterations of
N-methyl-D-aspartate receptor properties after chemical ischemia.
J Pharmacol Exp Ther 295:572–577.
Bonfoco E, Krainc D, Ankarcrona M, Nicotera P, Lipton SA (1995) Apopto-
sis and necrosis: two distinct events induced, respectively, by mild and
intense insults with N-methyl-D-aspartate or nitric oxide/superoxide in
cortical cell cultures. Proc Natl Acad Sci USA 92:7162–7166.
Boulares AH, Yakovlev AG, Ivanova V, Stoica BA, Wang G, Iyer S, Smulson M
(1999) Role of poly(ADP-ribose) polymerase (PARP) cleavage in apo-
ptosis. Caspase 3-resistant PARP mutant increases rates of apoptosis in
transfected cells. J Biol Chem 274:22932–22940.
Burkle A, Chen G, Kupper JH, Grube K, Zeller WJ (1993) Increased
poly(ADP-ribosyl)ation in intact cells by cisplatin treatment. Carcino-
Chen J, Simon R (1997) Ischemic tolerance in the brain. Neurology
Chen J, Graham SH, Zhu RL, Simon RP (1996) Stress proteins and tolerance
to focal cerebral ischemia. J Cereb Blood Flow Metab 16:566–577.
Chen M, Won DJ, Krajewski S, Gottlieb RA (2002) Calpain and mitochon-
dria in ischemia/reperfusion injury. J Biol Chem 277:29181–29186.
Cohen GM (1997) Caspases: the executioners of apoptosis. Biochem J
Cosi C, Suzuki H, Milani D, Facci L, Menegazzi M, Vantini G, Kanai Y, Skaper
SD (1994) Poly(ADP-ribose) polymerase: early involvement in
glutamate-induced neurotoxicity in cultured cerebellar granule cells.
J Neurosci Res 39:38– 46.
Currie RW, Ellison JA, White RF, Feuerstein GZ, Wang X, Barone FC (2000)
Benign focal ischemic preconditioning induces neuronal Hsp70 and pro-
longed astrogliosis with expression of Hsp27. Brain Res 863:169–181.
D’Amours D, Germain M, Orth K, Dixit VM, Poirier GG (1998) Proteolysis
of poly(ADP-ribose) polymerase by caspase 3: kinetics of cleavage of
mono(ADP-ribosyl)ated and DNA-bound substrates. Radiat Res
D’Amours D, Desnoyers S, D’Silva I, Poirier GG (1999) Poly(ADP-
ribosyl)ation reactions in the regulation of nuclear functions. Biochem J
De Maria R, Zeuner A, Eramo A, Domenichelli C, Bonci D, Grignani F,
Srinivasula SM, Alnemri ES, Testa U, Peschle C (1999) Negative regula-
tion of erythropoiesis by caspase-mediated cleavage of GATA-1. Nature
Dirnagl U, Simon RP, Hallenbeck JM (2003) Ischemic tolerance and endog-
enous neuroprotection. Trends Neurosci 26:248–254.
Eliasson MJ, Sampei K, Mandir AS, Hurn PD, Traystman RJ, Bao J, Pieper A,
Wang ZQ, Dawson TM, Snyder SH, Dawson VL (1997) Poly(ADP-
ribose) polymerase gene disruption renders mice resistant to cerebral
ischemia. Nat Med 3:1089–1095.
Endres M, Wang ZQ, Namura S, Waeber C, Moskowitz MA (1997) Isch-
emic brain injury is mediated by the activation of poly(ADP-
ribose)polymerase. J Cereb Blood Flow Metab 17:1143–1151.
Feldenberg LR, Thevananther S, del Rio M, de Leon M, Devarajan P (1999)
Partial ATP depletion induces Fas- and caspase-mediated apoptosis in
MDCK cells. Am J Physiol 276:F837–846.
Francois F, Godinho MJ, Dragunow M, Grimes ML (2001) A population of
PC12 cells that is initiating apoptosis can be rescued by nerve growth
factor. Mol Cell Neurosci 18:347–362.
Germain M, Affar EB, D’Amours D, Dixit VM, Salvesen GS, Poirier GG
(1999) Cleavage of automodified poly(ADP-ribose) polymerase during
apoptosis. Evidence for involvement of caspase-7. J Biol Chem
Glazner GW, Chan SL, Lu C, Mattson MP (2000) Caspase-mediated degra-
dation of AMPA receptor subunits: a mechanism for preventing excito-
toxic necrosis and ensuring apoptosis. J Neurosci 20:3641–3649.
Gonzalez-Zulueta M, Feldman AB, Klesse LJ, Kalb RG, Dillman JF, Parada LF,
Dawson TM, Dawson VL (2000) Requirement for nitric oxide activa-
tion of p21(ras)/extracellular regulated kinase in neuronal ischemic pre-
conditioning. Proc Natl Acad Sci USA 97:436– 441.
Grabb MC, Choi DW (1999) Ischemic tolerance in murine cortical cell cul-
ture: critical role for NMDA receptors. J Neurosci 19:1657–1662.
Ha HC, Snyder SH (1999) Poly(ADP-ribose) polymerase is a mediator of
necrotic cell death by ATP depletion. Proc Natl Acad Sci USA
Heller B, Wang ZQ, Wagner EF, Radons J, Burkle A, Fehsel K, Burkart V, Kolb
H (1995) Inactivation of the poly(ADP-ribose) polymerase gene affects
oxygen radical and nitric oxide toxicity in islet cells. J Biol Chem
Hoshida S, Yamashita N, Otsu K, Hori M (2002) The importance of man-
ganese superoxide dismutase in delayed preconditioning. Involvement of
reactive oxygen species and cytokines. Cardiovasc Res 55:495–505.
Ishida T, Yarimizu K, Gute DC, Korthuis RJ (1997) Mechanisms of ischemic
preconditioning. Shock 8:86–94.
Figure6. Caspaseinhibition reducestheresistance toCIconferred byischemicprecondition-
ing. Cultures were treated sham washes only (0 min PC), with 15 min of chemical ischemia (15
min PC), or with 15 min of chemical ischemia in the presence of 25
M DEVD-CHO. After 24 hr
the cultures were exposed to CI for durations of 0, 30, or 45 min. DEVD-CHO was added 2 hr
before onsetof PC andmaintained until onsetof CI. Neuronal deathwas assessed 24hr after CI.
**p ⬍ 0.01; representative of three independent experiments each with n ⫽ 3– 4.
7972 • J. Neurosci., September 3, 2003 • 23(22):7967–7973 Garnier et al. •Ischemic Preconditioning by Caspase Cleavage of Poly(ADP-Ribose) Polymerase-1
Kirino T (2002) Ischemic tolerance. J Cereb Blood Flow Metab
Kroemer G, Reed JC (2000) Mitochondrial control of cell death. Nat Med
Lazebnik YA, Kaufmann SH, Desnoyers S, Poirier GG, Earnshaw WC (1994)
Cleavage of poly(ADP-ribose) polymerase by a proteinase with properties
like ICE. Nature 371:346–347.
Le DA, Wu Y, Huang Z, Matsushita K, Plesnila N, Augustinack JC, Hyman
BT, Yuan J, Kuida K, Flavell RA, Moskowitz MA (2002) Caspase activa-
tion and neuroprotection in caspase-3-deficient mice after in vivo cere-
bral ischemia and in vitro oxygen glucose deprivation. Proc Natl Acad Sci
Leist M, Raab B, Maurer S, Rosick U, Brigelius-Flohe R (1996) Conven-
tional cell culture media do not adequately supply cells with antioxidants
and thus facilitate peroxide-induced genotoxicity. Free Radic Biol Med
Lemasters JJ, Qian T, Bradham CA, Brenner DA, Cascio WE, Trost LC, Nish-
imura Y, Nieminen AL, Herman B (1999) Mitochondrial dysfunction in
the pathogenesis of necrotic and apoptotic cell death. J Bioenerg Bi-
Lipton P (1999) Ischemic cell death in brain neurons. Physiol Rev
Lobner D, Choi DW (1996) Preincubation with protein synthesis inhibitors
protects cortical neurons against oxygen-glucose deprivation-induced
death. Neuroscience 72:335–341.
Margolin N, Raybuck SA, Wilson KP, Chen W, Fox T, Gu Y, Livingston DJ
(1997) Substrate and inhibitor specificity of interleukin-1 beta-
converting enzyme and related caspases. J Biol Chem 272:7223–7228.
McLaughlin B, Hartnett KA, Erhardt JA, Legos JJ, White RF, Barone FC,
Aizenman E (2003) Caspase 3 activation is essential for neuroprotection
in preconditioning. Proc Natl Acad Sci USA 100:715–720.
Nandagopal K, Dawson TM, Dawson VL (2001) Critical role for nitric oxide
signaling in cardiac and neuronal ischemic preconditioning and toler-
ance. J Pharmacol Exp Ther 297:474– 478.
Oliver FJ, de la Rubia G, Rolli V, Ruiz-Ruiz MC, de Murcia G, Murcia JM
(1998) Importance of poly(ADP-ribose) polymerase and its cleavage in
apoptosis. Lesson from an uncleavable mutant. J Biol Chem
Pieper AA, Verma A, Zhang J, Snyder SH (1999) Poly (ADP-ribose) poly-
merase, nitric oxide and cell death. Trends Pharmacol Sci 20:171–181.
Rajdev S, Reynolds IJ (1994) Glutamate-induced intracellular calcium
changes and neurotoxicity in cortical neurons in vitro: effect of chemical
ischemia. Neuroscience 62:667–679.
Rao RV, Peel A, Logvinova A, del Rio G, Hermel E, Yokota T, Goldsmith PC,
Ellerby LM, Ellerby HM, Bredesen DE (2002) Coupling endoplasmic
reticulum stress to the cell death program: role of the ER chaperone
GRP78. FEBS Lett 514:122–128.
Ratan RR, Murphy TH, Baraban JM (1994) Macromolecular synthesis in-
hibitors prevent oxidative stress-induced apoptosis in embryonic cortical
neurons by shunting cysteine from protein synthesis to glutathione.
J Neurosci 14:4385–4392.
Shamloo M, Wieloch T (1999) Changes in protein tyrosine phosphoryla-
tion in the rat brain after cerebral ischemia in a model of ischemic toler-
ance. J Cereb Blood Flow Metab 19:173–183.
Shamloo M, Rytter A, Wieloch T (1999) Activation of the extracellular
signal-regulated protein kinase cascade in the hippocampal CA1 region in
a rat model of global cerebral ischemic preconditioning. Neuroscience
Shimazaki K, Ishida A, Kawai N (1994) Increase in bcl-2 oncoprotein and
the tolerance to ischemia-induced neuronal death in the gerbil hippocam-
pus. Neurosci Res 20:95–99.
Shimizu S, Nagayama T, Jin KL, Zhu L, Loeffert JE, Watkins SC, Graham SH,
Simon RP (2001) bcl-2 Antisense treatment prevents induction of tol-
erance to focal ischemia in the rat brain. J Cereb Blood Flow Metab
Smith PK, Krohn RI, Hermanson GT, Mallia AK, Gartner FH, Provenzano
MD, Fujimoto EK, Goeke NM, Olson BJ, Klenk DC (1985) Measure-
ment of protein using bicinchoninic acid. Anal Biochem 150:76– 85.
Sordet O, Rebe C, Plenchette S, Zermati Y, Hermine O, Vainchenker W,
Garrido C, Solary E, Dubrez-Daloz L (2002) Specific involvement of
caspases in the differentiation of monocytes into macrophages. Blood
Swanson RA, Benington JH (1996) Astrocyte glucose metabolism under
normal and pathological conditions in vitro. Dev Neurosci 18:515–521.
Szabo C, Dawson VL (1998) Role of poly(ADP-ribose) synthetase in inflam-
mation and ischaemia-reperfusion. Trends Pharmacol Sci 19:287–298.
Takahashi K, Pieper AA, Croul SE, Zhang J, Snyder SH, Greenberg JH (1999)
Post-treatment with an inhibitor of poly(ADP-ribose) polymerase atten-
uates cerebral damage in focal ischemia. Brain Res 829:46–54.
Tokime T, Nozaki K, Sugino T, Kikuchi H, Hashimoto N, Ueda K (1998)
Enhanced poly(ADP-ribosyl)ation after focal ischemia in rat brain.
J Cereb Blood Flow Metab 18:991–997.
Virag L, Szabo C (2002) The therapeutic potential of poly(ADP-ribose)
polymerase inhibitors. Pharmacol Rev 54:375–429.
Wang KK (2000) Calpain and caspase: can you tell the difference? Trends
Wang ZQ, Auer B, Stingl L, Berghammer H, Haidacher D, Schweiger M,
Wagner EF (1995) Mice lacking ADPRT and poly(ADP-ribosyl)ation
develop normally but are susceptible to skin disease. Genes Dev
Ying W, Han SK, Miller JW, Swanson RA (1999) Acidosis potentiates oxi-
dative neuronal death by multiple mechanisms. J Neurochem
Ying W, Sevigny MB, Chen Y, Swanson RA (2001) Poly(ADP-ribose) gly-
cohydrolase mediates oxidative and excitotoxic neuronal death. Proc Natl
Acad Sci USA 98:12227–12232.
Ying W, Chen Y, Alano CC, Swanson RA (2002) Tricarboxylic acid cycle
substrates prevent PARP-mediated death of neurons and astrocytes.
J Cereb Blood Flow Metab 22:774–779.
Yu SW, Wang H, Poitras MF, Coombs C, Bowers WJ, Federoff HJ, Poirier
GG, Dawson TM, Dawson VL (2002) Mediation of poly(ADP-ribose)
polymerase-1-dependent cell death by apoptosis-inducing factor. Science
Garnier et al. •Ischemic Preconditioning by Caspase Cleavage of Poly(ADP-Ribose) Polymerase-1 J. Neurosci., September 3, 2003
• 23(22):7967–7973 • 7973