We recently demonstrated that reperfusion rapidly induces the mitochondrial pathway of apoptosis in chick cardiomyocytes after 1 h of simulated ischemia. Here we tested whether ischemia-reperfusion (I/R)-induced apoptosis could be initiated by caspase-dependent cytochrome c release in this model of cardiomyocyte injury. Fluorometric assays of caspase activity showed little, if any, activation of caspases above baseline levels induced by 1 h of ischemia alone. However, these assays revealed rapid activation of caspase-2, yielding a 2.95 ± 0.52-fold increase (over ischemia only) within the 1st h of reperfusion, whereas activities of caspases-3, -8, and -9 increased only slightly from their baseline levels. The rapid and prominent activation of caspase-2 suggested that it could be an important initiator caspase in this model, and using specific caspase inhibitors given only at the point of reperfusion, we tested this hypothesis. The caspase-2 inhibitor benzyloxycarbonyl-Val-Asp(Ome)-Val-Ala-Asp(Ome)-CH2F was the only caspase inhibitor that significantly inhibited cytochrome c release from mitochondria. This inhibitor also completely blocked activation of caspases-3, -8, and -9. The caspase-3/7 inhibitor transiently and only partially blocked caspase-2 activity and was less effective in blocking the activities of caspases-8 and -9. The caspase-8 inhibitor failed to significantly block caspase-2 or -3, and the caspase-9 inhibitor blocked only caspase-9. Furthermore, the caspase-2 inhibitor protected against I/R-induced cell death, but the caspase-8 inhibitor failed to do so. These data suggest that active caspase-2 initiates cytochrome c release after reperfusion and that it is critical for the I/R-induced apoptosis in this model.
- caspase activation
- caspase inhibitor
ischemia-reperfusion (I/R) can induce apoptosis in cardiomyocytes, and several studies have implicated the “intrinsic” or mitochondrial pathway in the cell injury in models of myocardial infarction and cardiomyocyte death (see Ref. 4 for a recent review). Caspases (cysteinyl aspartate-specific proteases) play critical roles in the initiation and execution of apoptotic pathways, including those induced by I/R (10, 18, 21, 26, 38, 44). Several caspases have been described in vertebrates. Among these, important initiator functions have been documented for caspase-8 in death receptor-induced apoptosis and for caspase-9 in mitochondrial pathways (reviewed in Ref. 12). A large volume of literature also supports a role for caspase-3 as a central “executioner” or “downstream” caspase, and, for a number of years after its discovery in 1994 (22, 43), caspase-2 was thought to be primarily a downstream effector caspase as well (34). Across animal species, caspase-2 is one of the most conserved caspases (reviewed in Ref. 37) and, interestingly, shares structural and functional features of the “initiator” (caspase association recruitment domain) and the more downstream “effector” (substrate-specific) caspase subfamilies (27, 36). Recently, caspase-2 has been referred to as an “orphan” caspase (37), but interest in this unique caspase has increased after demonstrations of an apparent upstream role for caspase-2 in permeabilizing mitochondria and initiating the release of proapoptotic mitochondrial proteins (16, 23, 30, 32). Given that cytochrome c release is rapidly initiated after reperfusion of ischemic chick cardiomyocytes and is associated with caspase activation and apoptotic cell death (38), we hypothesized that caspase-2 activation may be involved in early reperfusion injury.
The aims of the present study were to examine the participation of certain caspases in the apoptosis induced in our isolated chick cardiomyocyte model of I/R injury and, in particular, the potential mediation of cytochrome c release from the mitochondria via caspase action in these cells. Here we report the activation of caspases-2, -3, -8, and -9 during reperfusion of ischemic chick cardiomyocytes. We also report the effect of caspase inhibition with selective inhibitors on the release of cytochrome c from mitochondria and on the sequential activation of these caspases after reperfusion. Results from these experiments showed that initiation of caspase-2-like (VDVADase) activity was early and prominent after reperfusion and that inhibition of this activity at the point of reperfusion blocks cytochrome c release and postmitochondrial activation of other caspases. To our knowledge, this is the first report of caspase activity, in particular caspase-2-like activity, in initiating cytochrome c release in I/R-induced apoptosis of cardiomyocytes.
MATERIALS AND METHODS
The caspase-2 inhibitor benzyloxycarbonyl-Val-Asp(Ome)-Val-Ala-Asp(Ome)-CH2F (zVDVAD-fmk) and the caspase-8 inhibitor z-Ile-Glu(Ome)-Thr-Asp(Ome)-CH2F (zIETD-fmk) were purchased from Calbiochem (La Jolla, CA). The caspase-9 inhibitor z-Leu-Glu(Ome)-His-Asp(Ome)-CH2F (zLEHD-fmk) and the caspase-3/7 inhibitor Ac-Asp-Gln-Thr-Asp-H (Ac-DQTD-CHO) were obtained from Enzyme System Products (Livermore, CA) and Peptide International (Louisville, KY), respectively. Anticytochrome c antibody, used for Western blot analysis, was obtained from BD Biosciences Pharmingen (San Diego, CA). Buffer reagents were obtained from Sigma (St. Louis, MO). Propidium iodide (PI), for measuring cell viability, was obtained from Molecular Probes (Eugene, OR).
Cardiomyocyte preparation and cell culture.
Ventricular embryonic chick cardiomyocytes were prepared as previously described (38, 40). Briefly, 10-day-old chicken embryo hearts were resected, and ventricular tissue was minced into 0.5-mm fragments and enzymatically dispersed with 0.025% trypsin (Life Technologies, New York, NY). Isolated cells were collected by centrifugation and resuspended in complete medium 199 with supplements described previously (38, 40). Cells (0.7 × 106) were pipetted onto 25-mm glass coverslips, preplated for removal of fibroblasts, and placed in a humidified incubator (5% CO2). Experiments were performed with cardiac cell cultures at 3–5 days, by which time a synchronously contracting layer of cells could be visualized, and viability was >99%.
Glass coverslips with contracting cells were placed in a 1.2-ml Sykes-Moore perfusion chamber (Bellco Glass, Vineland, NJ) as described elsewhere (38, 40). Tubing to the chamber was made of stainless steel to minimize diffusion of ambient O2. Perfusate was pumped through the chamber (0.25 ml/min) by a roller pump via water-jacketed tubing that controlled for temperature (37°C). Normoxic perfusate used for baseline conditions and for reperfusion subsequent to ischemia consisted of oxygenated balanced salt solution, with 149 Torr Po2, 40 Torr Pco2, pH 7.4, and 4.0 meq/l K+, containing glucose (5.6 mmol/l). To simulate ischemia, balanced salt solution containing 2-deoxyglucose (20 mmol/l) was used without glucose and with 8.0 meq/l K+. Before use, the solution was equilibrated with 80% N2-20% CO2 to produce ∼5 Torr Po2, 144 Torr Pco2, and final pH 6.8. The typical experiment consisted of 1 h of equilibration (“baseline”), 1 h of simulated ischemia, and 3 h of reperfusion (38).
An Olympus IMT-2 inverted phase/epifluorescent microscope was used for cell imaging. Phase-contrast Hoffmann modulation optics and a charge-coupled device camera were used to monitor contractions and morphological membrane changes in the same field of cells (∼70 × 90 μm) over time. Fluorescent images were acquired from a cooled slow-scanning PC-controlled camera (Hamamatsu, Hamamatsu City, Japan), and changes in fluorescence intensity over time were quantified with MetaMorph software (Universal Imaging, Downington, PA).
Cell viability and cell contraction.
Cell viability was assessed with the fluorochrome PI (5 μmol/l). PI has been used previously to predict the transition from reversible to irreversible cell injury in cultured cardiomyocytes (3). It is excluded from viable cells and binds to chromatin after loss of cell membrane integrity, becoming highly fluorescent (excitation wavelength of 540 nm and a 590-nm band-pass emission filter). The dye was used to quantify cell death throughout each experiment, exhibiting minimal toxicity in control cells even after 10 h of exposure. All cells in the field were stained with PI at the end of the experiment by permeabilization with 300 μM digitonin, and percent cell death was calculated. Cell death was expressed as the PI fluorescence at any given time (≥2 measurements per hour) relative to the maximal value seen after digitonin exposure (100%). Cell contractions were assessed by observation of movement within the same field of cells as previously reported (39). A return of contraction after simulated I/R was indicated when contractions could be seen throughout the field of cells after the 3 h of reperfusion.
Caspase activity assay.
To assess caspase activities, caspase-2 (VDVADase), caspase-3 (DEVDase), caspase-8 (IETDase), and caspase-9 (LEHDase) activities were measured using standard fluorogenic substrates (12, 19, 20). Cells plated on coverslips were subjected to I/R at various time points and lysed in 250 μl of 1× lysis buffer. The lysates were collected and stored at −70°C until analyzed. The BD ApoAlert caspase profiling plate (Becton Dickinson, Palo Alto, CA), based on the cleavage of these fluorogenic substrates by activated caspases, was used according to the manufacturer's instructions. Briefly, 50 μl of 1× reaction buffer were added to each well used in the assay and incubated at 37°C for 5 min. Then 50 μl of cell lysate were added to each appropriate well and allowed to incubate at 37°C for 2 h. The plate was read in a CytoFluor II PerSeptive Biosystems (Farmington, MA) fluorometric plate reader with excitation at 360 nm and emission at 460 nm. Readouts were converted to Excel files, background fluorescence (from wells without cell lysates) was subtracted, and activities were calculated as fold increase above ischemia-only measurements. No statistically significant increases in caspase activities above time 0 for lysates from cells after ischemia only were observed (data not shown). To determine the specificity of the plate as well as the relative contribution of each caspase to the apoptotic death of chick cardiomyocytes, selective caspase inhibitors (for caspase-2, -3/7, -8, or -9 activity) were added at the point of reperfusion as previously described (38). Doses were determined in preliminary experiments, and the minimum effective (and nontoxic) dose was used for each inhibitor. Cell lysates were collected at various times for each experiment and condition as described above and run accordingly.
Western blot analyses for cytochrome c.
At designated times during the I/R protocol, cardiomyocytes were removed from the chamber and placed into ice-cold PBS. The mitochondrial free cytosolic proteins were prepared using a protocol previously described (38). Briefly, cells were harvested from coverslips using a cell scraper (Corning, Corning, NY), washed twice in ice-cold PBS by centrifugation at 200 g for 5 min, resuspended in 5 vol of mitochondrial isolation buffer (in mM: 220 mannitol, 50 KCl, 5 EGTA, 1 DTT, 2 MgCl2, 50 PIPES-KOH, pH 7.4, and 68 sucrose), and allowed to swell on ice for 20 min. Cells were homogenized with 40 strokes of a Teflon homogenizer and centrifuged at 14,000 g for 15 min at 4°C. The supernatants were transferred to a new microcentrifuge tube and subjected to additional centrifugation at 14,000 g for 15 min, and the cytosolic protein was collected and stored at −80°C until analysis. The protein concentration of the cytosolic extract was determined by the Bradford method (Bio-Rad Laboratories, Hercules, CA). Proteins were subjected to 15% SDS-PAGE and Western blot analysis using anticytochrome c monoclonal antibody (clone 7H8.2C12, PharMingen).
For each experiment, a field of ∼500 cells was observed. Treatment and control groups were matched in sets containing cells isolated and cultured on the same day to eliminate variability due to cell batch. Values are means ± SE, and two-tailed unpaired t-tests were performed as post hoc tests of significance, with P < 0.05 considered to be significant.
I/R-induced caspase activation and apoptosis.
As we previously reported (38), chick cardiomyocytes exposed to 1 h of simulated ischemia followed by 3 h of reperfusion exhibited significant reperfusion injury. We have confirmed these results in this study with the finding that cells exposed to prolonged ischemia for 4 h exhibit significantly less cell death (16.3 ± 1.1%, n = 6) than cells exposed to 1 h of ischemia followed by 3 h of reperfusion (53.5 ± 3.5%, n = 5, P < 0.001; Fig. 1A). With the use of fluorogenic substrate assays for activities of caspases-2, -3, -8, and -9, no significant elevation above preischemic levels of activity of any of these caspases could be measured after 1 h of ischemia only (data not shown). Within the 1st h of reperfusion, only caspase-2 activity was consistently and significantly increased, reaching nearly a threefold increase over caspase-2 activity measured after 1 h of ischemia only (Fig. 1B). Activity of caspases-3, -8, and -9 also tended to be elevated by the end of the 1st h of reperfusion, but these apparent increases were not statistically significant (P = 0.5 for caspase-3, P = 0.5 for caspase-8, and P = 0.1029 for caspase-9). However, by the end of the 3 h of reperfusion, approximately fivefold or more increases in activity over levels measured after ischemia only could be measured for all four caspases (Fig. 1C).
Effect of caspase-2 inhibition on caspase activity and cell death.
The rapid and prominent appearance of caspase-2-like activity suggested that this caspase could be an important initiator caspase in the cardiomyocyte response to reperfusion as has been described in some cells after other apoptotic stimuli (16, 23, 30, 32, 37). To examine this possibility, we used the selective caspase-2 inhibitor zVDVAD-fmk given only at reperfusion to block caspase-2 activity and examined the effects of this inhibitor on cell death and on measured caspase activities. Blockade of caspase-2 activity significantly inhibited the reperfusion-induced cell death to levels seen for ischemia only (Figs. 1A and 2A). Furthermore, inhibition of caspase-2 resulted in return of spontaneous contractions of cardiomyocytes after reperfusion. This blockade of cell death by zVDVAD-fmk was associated with complete inhibition of the activity of caspase-2 and the activities of the three other caspases (-3, -8, and -9) (Fig. 2B).
Effect of inhibition of other caspases on caspase activity and cell death.
The activation of caspase-8 has been implicated in the death of cardiomyocytes in some model systems (9, 26, 33), and during the later periods of reperfusion, prominent activation of caspase-8 was observed in our cell system (Fig. 1C). Therefore, we tested the effects of the selective caspase-8 inhibitor zIETD-fmk on the reperfusion-induced activation of caspases and cell death. Addition of this inhibitor at reperfusion had no significant effect on the progressive cell death throughout the 3 h of reperfusion after ischemia (Fig. 3A). In addition, although the inhibitor effectively blocked the activation of its target caspase (caspase-8), it failed to significantly block the activation of caspases-2, -3, and -9, and each of these caspases reached their previously demonstrated (Fig. 1C) maximum (∼5-fold) activities by the end of the 3-h reperfusion period (Fig. 3B).
We previously reported that inhibition of caspase-3/7 and -9 could block, at least partially, the acute apoptotic death of chick cardiomyocytes (38). However, caspase-9 inhibition blocked only caspase-9 throughout the reperfusion period, and caspase-3/7 inhibition transiently blocked activation of the caspases, including caspase-3 itself (Fig. 4). By the end of the 3-h reperfusion period, activation of the caspases approached the levels seen in the absence of inhibitors (Fig. 1C).
Effect of inhibition of caspase activity on cytochrome c release.
Our previous studies of the induction of apoptosis in chick cardiomyocytes demonstrated a rapid release of cytochrome c from the mitochondria that peaked at 30 min after the beginning of reperfusion and measurable caspase-9 activation by 1–2 h (38). If caspase-2 is upstream of other caspases tested, including caspase-9, it is possible that its activation is necessary for cytochrome c release and postmitochondrial initiation of the caspase cascade in these cells. Therefore, we tested the effects of the inhibitors used in the experiments described above on cytochrome c release from the cardiomyocyte mitochondria. The inhibitors were tested for their abilities to block cytochrome c release into the cytosol after 30 min of reperfusion, which we previously had shown to elicit peak measurable levels of cytochrome c release into the cytoplasm (38). As described above, the inhibitors were given only at the point of reperfusion. Only the caspase-2 inhibitor zVDVAD-fmk was capable of significantly blocking cytochrome c release (Fig. 5).
Caspase-2 activation can occur early in the apoptotic process in many cells. The early activation of caspase-2 and the blockade of cytochrome c release by inhibition of caspase-2-like activity support recent evidence that activation of caspase-2 can precede the release of cytochrome c from the mitochondria (16, 30). Our data extend this evidence by suggesting that reoxygenation of ischemic tissue can also induce caspase-2-like activity as an initiator of cytochrome c release and reperfusion injury. Guo et al. (16) showed that physiological amounts of recombinant caspase-2 protein could directly induce the mitochondrial release of cytochrome c and lead to activation of the apoptotic caspase cascade. A similar proximal role for caspase-2 in etoposide-induced cell death of Jurkat T lymphocytes has been reported; its inhibition with the inhibitor zVDVAD-fmk or procaspase-2 antisense transfection blocked the release of cytochrome c (32). Although the exact mechanism of caspase-2 activation is not clear, a number of stimuli can lead to its early activation in the apoptotic cascade (reviewed in Ref. 37). Oxidative stress-dependent apoptosis was associated with caspase-2 cleavage and cytochrome c release in putrescine-mediated cell death in murine myeloma cells, but it was not clear in this study whether caspase-2 activation took place before mitochondrial release of cytochrome c (15).
Broad and selective inhibition of caspases can be instructive in ordering the events of the caspase cascade (34). Our data suggest that caspase-2 is activated before caspases-3, -9, and -8 (Fig. 1) and, furthermore, that caspase-2-like activity is necessary for the subsequent activation of these caspases via cytochrome c release and for I/R-induced apoptotic death of the cardiomyocytes (Fig. 2). In recent studies of I/R injury in whole rabbit models, the general caspase inhibitor Ac-YVAD-chloromethylketone decreased activity of caspases and reduced the number of apoptotic cardiomyocytes and infarct size in the myocardium of treated rabbits (18, 44). Both studies showed a decrease in caspase-3 activity, and one also demonstrated a blockade of caspase-2 activity (18). However, temporal ordering of caspase activation, particularly with respect to cytochrome c release, was not examined and may not have been possible in this whole animal model. Blockade of early caspase-2 activation (by zVDVAD-fmk) has been shown to be more protective than caspase-3 inhibition in a number of cellular models (6, 7, 46), including one examining staurosporine-induced apoptosis of chick cardiomyocytes (31). The release of cytochrome c from mitochondria on reperfusion of ischemic cells, however, has not been previously associated with caspase-2-like activity. Our observations also suggest that, in cardiomyocytes, caspase-8 activation may be, at least in large part, a downstream or amplification event (Fig. 1, B and C). Downstream activation of caspase-8 has been reported in a number of models (11, 42, 45), and progressive caspase-8 activation was recently observed in an isolated Langendorff-perfused rat heart model (33). Roles for caspases-3 and -8 as downstream amplifiers and executioners in apoptotic pathways have been suggested (42). Our data are consistent with a downstream role for caspase-8 in I/R-induced apoptosis in chick cardiomyocytes. This may involve an executioner role for caspase-8 similar to that of caspase-3 (14), or it may act as a mitochondrial amplifier (41). In our model, selective inhibition of caspase-9-, -8-, or -3-like activities did not block cytochrome c release from the mitochondria in cardiomyocytes (Fig. 5), suggesting that these caspases may have primarily postmitochondrial and/or amplification roles in these cells.
Our data showing transient inhibition of caspase-2 activation by the caspase-3 inhibitor (Fig. 4B) suggest that inhibition of caspase-3-like activity is detected by the VDVAD substrate or that some activation of caspase-2 may be due to caspase-3/7 mediation. However, blockade of caspase-3 activity (in contrast to caspase-2 inhibition) fails to block cytochrome c release (Fig. 5). It is possible that, in the intact cardiomyocyte, caspase-3 plays an amplification role downstream of mitochondria and contributes to further caspase activation (potentially including additional caspase-2 cleavage). This pathway does not appear to be essential for this initiating mitochondrial event in our model. In a Langendorff-perfused rat heart model, caspase-3 inhibition (DEVD-CHO) prevented ischemia-induced DNA fragmentation but only partially blocked cytochrome c release from mitochondria, likely through blockade of amplification (5). In addition, studies of granzyme B-induced apoptosis of Jurkat T cells showed that caspase-3 may help permeabilize the mitochondria and, therefore, may represent an amplification loop (25). However, this “ancillary” pathway does not appear to be essential for apoptosis of these T cells (25), and the lack of any effect of caspase-3-like inhibition on cytochrome c release in our cardiomyocytes (Fig. 5) suggests, similarly, that this pathway may not be critical in I/R-induced apoptosis of these cells. We have summarized a potential sequence of events in this model of I/R-induced apoptosis in cardiomyocytes in Fig. 6.
It is important to note some potential limitations to the present study. As we previously discussed (38), our model of simulated ischemia may not adequately reproduce all the characteristics of “clinical” ischemia. Although there is no accepted experimental standard that replicates clinically relevant ischemic exposure for cells, we believe that our model recreates a number of important components, including hypoxia, hypercarbia, hyperkalemia, and substrate deprivation (38–40). We also previously discussed the potential limitations that this embryonic avian system may not model the events of apoptosis during I/R in a mammalian system. However, this cellular system is one of the few that exhibits several important qualities, including reperfusion-induced oxidant generation, synchronous contractile function, and preconditioning protection (38–40). Furthermore, we believe that the well-documented conservation of apoptotic molecules and mechanisms (1, 7, 43) make this model a useful one in dissecting some of these mechanisms.
Caution must be exercised during interpretation of temporal events associated with caspase activation using inhibitors that may overlap in function. A caspase-3 DEVD-based inhibitor has been shown to partially inhibit caspase-2 in some systems (24), probably because of one of its preferred peptide substrates, DEHD (13, 36). Therefore, we utilized the DQTD form, which also has been shown to effectively inhibit caspase-3 activity on specific endogenous substrates in intact cardiomyocytes (38). Although there are no reports of direct inhibition of caspase-2 by the caspase-3 inhibitor zDQTD-CHO, it is possible that this inhibitor can inhibit at least a portion or “subset” of the caspase-2 activity. In any case, this potential and transient blockade of caspase-2 by the DQTD caspase-3 inhibitor as measured by the activity assays is insufficient to block cytochrome c release from our cardiomyocytes (Fig. 5). These observations may be due to exquisite intracellular sensitivity of mitochondrial cytochrome c release to low levels of active (or a subset of) caspase-2, i.e., levels below the detection limit of our activity assays. Additionally or alternatively, because cytochrome c release can occur within minutes (before we can effectively measure caspase activity), the observed DQTD-fmk-mediated blockade of caspase-2 activity at 1 h may not reflect persistent (and sufficient) intracellular activity at these very early time points.
Together, these data support the hypothesis of caspase-initiated cytochrome c release and apoptosis of cardiomyocytes after simulated I/R. Although caspase-2 activity appears critical in I/R-induced apoptosis in these cardiomyocytes, we cannot discount entirely the potential involvement of caspase-3 in contributing to or amplifying caspase-2 activation in the intact cell. Furthermore, the molecular mechanism(s) of caspase-2 activation by I/R or other stimuli is unclear (reviewed in Ref. 37). Oligomerization or dimerization may be a general mechanism for the activation of initiator caspases such as caspases-2 and -9 (2, 8), but the trigger for this oligomerization in I/R-induced caspase-2 activation is not known. Although the subcellular localization of caspase-2 may provide clues about the mechanism of activation, its precise organellar location (e.g., mitochondria vs. nucleus, Golgi, or other) is a matter of controversy (28, 35) and may vary according to cell type (37). Finally, regardless of specificity of caspase inhibitors, it is difficult to establish exclusive roles for individual caspases. Indeed, cooperation between caspases, at least through amplification, is likely, and any therapeutic approach will need to consider this.
This study was supported by National Heart, Lung, and Blood Institute Grants R01-HL-65558 (K. J. Hamann), R01-HL-66026 (K. J. Hamann), and R01-HL-68951 (T. L. Vanden Hoek).
↵* Y. Qin and T. L. Vanden Hoek contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 by the American Physiological Society