Induction of CPP32-like activity in PC12 cells by withdrawal of trophic support. Dissociation from apoptosis.

Inhibitors of interleukin-1β converting enzyme (ICE) and a related group of cysteine aspartases of the ICE/ced-3 family inhibit cell death in a variety of settings, including in PC12 cells and sympathetic neurons following withdrawal of trophic support. To assess the particular member(s) of the ICE/ced-3 family that are relevant to cell death and to position their activation within the apoptotic pathway, we have used specific substrates to measure ICE-like and CPP32-like enzymatic activity in naive and neuronally differentiated PC12 cells that had been deprived of trophic support (nerve growth factor and/or serum). Rapid induction of CPP32-like, but not ICE-like, activity was observed. c-Jun kinase activation and the action of bcl-2 and other survival agents, such as cell cycle blockers, a NO generator, N-acetylcysteine, aurintricarboxylic acid, and actinomycin D occurred at a point further upstream in the apoptotic pathway compared with the aspartase activation. In living cells, zVAD-FMK, a pseudosubstrate aspartase inhibitor, blocked the activity/activation of the aspartase at concentrations about one order of magnitude lower than those required to promote survival, raising the possibility that the CPP32-like aspartase is not the main death effector in this model.

Apoptosis is the predominant form of programmed cell death during development. It is characterized by cellular shrinkage, chromatin condensation, nuclear fragmentation, formation of apoptotic bodies, and the absence of an inflammatory response. Apoptotic neuronal cell death occurs normally during development and may also occur in a variety of neurological disorders. The development of strategies to arrest apoptotic death of neurons may, therefore, be used to ameliorate a number of neurological afflictions (1,2).
The temporal sequence of events leading from the apoptotic stimulus to nuclear fragmentation and death of cells, including neurons, is unknown, but a number of clues have emerged from the study of Caenorhabditis elegans, which exhibits a highly regulated pattern of cell death (3). Three genes in particular have emerged as important in this respect: ced-3 and ced-4 act as a pro-apoptotic genes (4), whereas ced-9 acts as an antiapoptotic gene (5). Great interest has been spurred by the discovery of mammalian homologues for these genes. Bcl-2 and a related group of proteins, some of which have an anti-apoptotic effect and some of which are pro-apoptotic, are the mammalian homologues of ced-9 (6,7). Interleukin-1␤ converting enzyme (ICE) 1 and a related group of cysteine proteases are the homologues of ced-3 (8,9). ICE and the other members of the family of ICE/ced-3-like proteases share the unusual ability to cleave proteins at an aspartate residue, hence the name aspartases. (9,10). Mammalian homologues of ced-4 have yet to be identified.
The general principles outlined above are relevant to apoptotic neuronal cell death. Bcl-2 prevents neuronal cell death in a variety of settings (7). CrmA inhibits cell death of chick dorsal root ganglia cells following withdrawal of NGF (29), whereas peptide pseudosubstrate inhibitors of ICE, which may have a broader action against other members of the ICE/ced-3 family, protect motor neurons in vivo and in vitro from programmed cell death (30). We have shown, using naive and neuronally differentiated PC12 cells and sympathetic neurons, that zVAD-FMK, a competitive irreversible inhibitor of ICE, and V-ICEinh, a newly designed peptide inhibitor that mimics the shared catalytic site of the cysteine aspartases of the ICE/ced-3 family, prevent cell death due to withdrawal of trophic support and to oxidative stress (31,32).
In this study, we have begun to investigate the specific aspartases that are activated in response to withdrawal of trophic support from neuronal cells. To achieve this, we used the strategy of measuring enzymatic activity using substrates of the most representative aspartases of each of the two subdivisions, i.e. ICE of the ICE-like and CPP32 of the ced3-like subdivision. To provide sufficiently homogeneous material for experimentation, we used cultured PC12 cells, which have proved to be useful for studying mechanisms of neuronal apoptotic death (33)(34)(35)(36)(37). In these cells, as is the case of their sympathetic neuron counterparts, the withdrawal of trophic support in the form of NGF and/or serum evokes stereotyped apoptotic death of the entire population. This approach has enabled us to identify a CPP32-like aspartase activity that is greatly increased prior to apoptotic death caused by trophic factor withdrawal and whose activation is inhibited by a number of agents that suppress death. These studies also permit us to draw conclusions about the temporal sequence of events leading to apoptosis and to position this aspartase activity relative to other events in the apoptotic pathway. Finally, we have been able to evaluate the potential role of this enzyme in the apoptotic cascade.
Survival Assay-Naive and neuronally differentiated PC12 cells were mechanically dissociated from 100-mm dishes after five rinses with serum-free RPMI 1640 and were washed with the same medium three to four times by centrifugation and resuspension. Cells were replated in collagen-coated 35-mm dishes in the case of the naive cells and in collagen-coated 24-well dishes in the case of the neuronally differentiated cells. At the indicated time points, the numbers of viable cells were determined by quantifying the number of intact nuclei, as described previously (33). Counts were performed in triplicate and reported as means Ϯ S.E.
Preparation of Cell Lysates for Assay of Aspartase Activity-At the indicated time points following withdrawal of trophic support, naive and neuronally differentiated PC12 cells were harvested for the assays of aspartase activity. Cells were rinsed in cold phosphate-buffered saline and then collected in a buffer of 25 mM Hepes, pH 7.5, 5 mM EDTA, 1 mM EGTA, 5 mM MgCl 2 , 5 mM dithiothreitol, 10 g/ml each of pepstatin and leupeptin, and 1 mM phenylmethylsulfonyl fluoride. The cellular material was left for 20 min on ice and then was sonicated on ice. The lysate was centrifuged for 20 min at 160,000 ϫ g, and the supernatant was quick frozen with liquid nitrogen and stored at Ϫ80°C (9, 10). Protein concentrations of the lysates were quantified using the Bradford method (40).
Cleavage of Fluorogenic Substrate-Lysate (25 g of protein except where indicated) were incubated at 37°C in a buffer of 25 mM Hepes, pH 7.5, 10% sucrose, 0.1% CHAPS, and 10 mM dithiothreitol with the fluorogenic substrates DEVD-AFC, which mimics the cleavage site on PARP, and YVAD-AFC, which mimics the cleavage site on pIL-1␤. Typically, 15 M DEVD-AFC and 25 M YVAD-AFC (Enzyme Systems Products, Dublin, CA) were used. Cleavage of the substrates emitted a fluorescent signal that was quantified in a SLM 8000 fluorometer (excitation, 400 nm; emission, 505 nm).
Cleavage of Radioactively Labeled Substrate-Radioactively labeled substrates were generated by in vitro transcription/translation. An IL-1␤ construct encoding the 31-kDa pIL-1␤ protein was kindly provided by Dr. J. M. Hardwick (Johns Hopkins University) in a pSGS vector. Purified ICE, which was used as a positive control, was kindly provided by Dr. N. A. Thornberry (Merck Laboratories). A PARP construct encoding a 48-kDa truncated form of PARP (tPARP) comprising the first 338 N-terminal amino acids and including the cleavage site by CPP32 within the DNA-binding domain was constructed by Dr. John Collier in a PET15 vector and kindly provided to us by Dr. David Fisher (both of Harvard University). Following generation of mRNA using T7 RNA polymerase, translation was performed using reticulocyte lysate (Promega Corp.) in the presence of S 35 -labeled methionine (DuPont NEN). The radioactively labeled products were incubated for the indicated time periods at 37°C with the indicated amounts of lysates of PC12 cells in a buffer of 100 mM Hepes, pH 7.5, 10% sucrose, 0.1% CHAPS, and 10 mM dithiothreitol. Subsequently, the products of the reaction were resolved by electrophoresis on a SDS-polyacrylamide gel, 12 or 15%. The gel was dried and fluorography was performed (15,41).
Parallel c-Jun Kinase and DEVD-AFC Cleavage Activity-PC12 cells were deprived of trophic support, as described above, and replated on 100-mm dishes. At the indicated time points, the cells were mechanically detached, and one-half were processed for the DEVD-AFC cleavage activity assay, as described above, while the other one-half were processed for the c-Jun kinase activity assay, as described previously. Briefly, c-Jun kinase activity was affinity purified from cell extracts with recombinant glutathione S-transferase-c-Jun. c-Jun kinase activity recovered from each sample was determined by an in vitro solid phase kinase radioassay. Glutathione S-transferase-c-Jun was resolved by PAGE (8.5% acrylamide), and incorporation of 32 PO 4 was quantified by autoradiography and densitometry (32,42,43).

Withdrawal of Trophic Support Does Not Activate ICE-like
Protease Activity-The fluorogenic substrate YVAD-AFC, which mimics the cleavage site at which ICE cleaves pIL-1␤, was used to detect ICE-like protease activity in lysates of PC12 cells that were plated in serum-free medium for 1, 2, 3, 6, 12, or 24 h with or without NGF treatment. Under these conditions, approximately 75% of the untreated PC12 cells undergo apoptotic death within 24 h, whereas Ͼ90% of the NGF-treated cells survive (44). There was no difference in substrate cleavage between the two sets of lysates. The low level of activity observed was not inhibited by high concentrations of the competitive irreversible ICE inhibitors AcYVAD-CMK or zVAD-FMK (Ͼ10 ϫ substrate concentration), implying that this was a nonspecific activity (data not shown). These results were confirmed with the use of radioactively labeled pIL-1␤ as a substrate. Despite demonstrable cleavage by purified ICE, there was no cleavage of pIL-1␤ by lysates of serum-deprived PC12 cells undergoing death (Fig. 1). In addition, there was no induction of YVAD-AFC cleavage in lysates of neuronally differentiated PC12 cells that had been deprived of NGF for 5, 24, or 36 h (data not shown). These findings indicate that PC12 cells do not show detectable activation of ICE-like protease activity before or during apoptotic death triggered by loss of trophic support.
Withdrawal of Trophic Support Results in Activation of a CPP32-like Protease Activity-By contrast with our findings for ICE activity, lysates of serum-deprived PC12 cells consistently brought about cleavage of the fluorogenic substrate DEVD-AFC ( Fig. 2A). This substrate mimics the PARP cleavage site by CPP32. The activity was inhibited in vitro by very low concentrations of the reversible pseudosubstrate inhibitor DEVD-CHO (apparent K i , 0.12 nM). However, even with satu-rating concentrations of the inhibitor (1 M with a substrate concentration of 15 M), there was a very small amount of residual activity, which was calculated as 0.007 nmol of AFC released/g protein/min. This activity appears to represent background nonspecific substrate cleavage activity and was subsequently subtracted from all measurements. With a concentration of 15 M, DEVD-AFC cleavage by up to 100 g of lysate protein was linear for at least 2 h. The apparent K m for the activity was calculated to be 16.7 M.
Significant DEVD-AFC cleavage activity was observed by 2 h following serum deprivation. By 5 h, the activity reached a peak and thereafter plateaued until 10 h. By 17 h, activity began to decrease. Lysates of serum-deprived PC12 cells that had been treated with NGF showed an approximately 7-fold lower activity ( Fig. 2A). Parallel assessment of surviving cells revealed that the DEVD-AFC cleavage activity preceded the onset of death and roughly correlated with the rate of death (Fig. 2B).
These results with DEVD-AFC were confirmed by measuring the cleavage of radioactively labeled tPARP (Fig. 3). The time course for the appearance of this activity following serum deprivation was similar to the one observed for the fluorogenic substrate. The 48-kDa substrate yielded products of approximate sizes of 28 and 20 kDa. This cleavage activity was inhibited by 50 nM DEVD-CHO (Fig. 4). The size of the products argues for cleavage of PARP at residue 216, as has been observed for CPP32 (15,16).
Similar results were achieved with neuronally differentiated PC12 cells that were deprived of NGF in serum-free medium (Fig. 5). These also showed induction of DEVD-AFC cleavage activity that preceded cell death. However, the activity peaked later (after 24 h), was of a lower magnitude, and was more protracted in time course than for serum-deprived naive cells. This is consistent with the more gradual evolution of cell death of neuronally differentiated cells after NGF deprivation. Under these conditions, approximately 65% of cells undergo apoptotic death within 48 h (33). Lysates of cells treated with NGF showed a much lower activity (Fig. 5). Cleavage of tPARP was again observed at the same site and was inhibited by 50 nM DEVD-CHO (Fig. 4).
The ICE Inhibitors zVAD-FMK and AcYVAD-CMK Inhibit the DEVD-AFC Cleavage Activity in Vitro-Our previous work has shown that the irreversible competitive pseudosubstrate inhibitors of ICE zVAD-FMK and AcYVAD-CMK, which act by mimicking the cleavage site on pIL-1␤, offer protection from cell death in serum-deprived PC12 cells. zVAD-FMK is much more effective in this model, offering partial protection at a concentration of 10 M and almost complete protection at 50 M, whereas AcYVAD-CMK is only partially effective at a concentration of 250 M (31). We tested these agents for their FIG. 1. Cleavage of pIL-1␤ by ICE but not by PC12 lysates. 0.1 unit of ICE or lysates of PC12 cells (containing 10 g of protein), which had been deprived of serum for 6 or 12 h, were incubated with radioactively labeled 31-kDa pIL-1␤ for 1 h at 37°C. The samples were then resolved by SDS-PAGE on a 15% polyacrylamide gel, and fluorography was performed. The expected products of 17 and 14 kDa, denoted by arrows, were generated by ICE but not by the PC12 lysates.

FIG. 2.
A, DEVD-AFC cleavage activity in lysates of naive PC12 cells treated with no additives or NGF following various times of serum deprivation. At the indicated time points following replating in serumfree medium, cells that had been treated with no additives (Ⅺ) or 100 ng/ml NGF (ࡗ) were harvested, and DEVD-AFC cleavage activity was measured in the corresponding lysates, as described under "Materials and Methods." Data are the means; bars, S.E. (n ϭ 3). B, parallel assessment of survival of naive PC12 cells following serum deprivation. Survival is reported relative to the number of initially plated cells. Data are the means; bars, S.E. (n ϭ 3). ability to affect the cleavage of DEVD-AFC by lysates of serumdeprived PC12 cells. The activity was inhibited in a dose-dependent fashion by both agents, but at doses at least 100 times higher than the ones used for DEVD-CHO, which, as noted previously, had a K i of 0.12 nM. Between the two agents, zVAD-FMK was clearly more potent. Over a 2-h period, inhibition of approximately 50% of the activity required 100 nM zVAD-FMK or 2 M AcYVAD-CMK (Fig. 6). zVAD-FMK was also about 25-50 times more potent as an inhibitor of purified ICE in in vitro assays using as substrates YVAD-AFC or pIL-1␤ (data not shown). These data would suggest that the superior efficacy of zVAD-FMK as a survival agent is due to its superior potency as a general inhibitor of ICE/ced-3 aspartases. Indeed, it is thought that fluoromethylketones, despite a more gradual onset of action, are more potent inhibitors of cysteine proteases than chloromethylketones (45).
Inhibitors of Cell Death Block Activation of DEVD-AFC/ tPARP Cleavage-We next assessed whether a variety of agents that inhibit or delay death of naive PC12 cells following serum deprivation would affect activation of DEVD-AFC/ tPARP cleaving activity. These agents include ATA (100 M), an inhibitor of a variety of enzymes and nucleases that blocks apoptotic death in many systems, including serum/NGF-deprived PC12 cell cultures (34); 8-(4-chlorophenylthio)-cAMP (100 M), a permeant cAMP analog (46); N-acetylcysteine (NAC; 60 mM), a reducing agent that blocks death by a transcription-dependent mechanism (47,48); the NO generator, SNAP, which may exert its survival effect through cGMP activation (200 M) (49); and the G 1 -S phase blocker deferoxamine (3 mM) (50). PC12 cells overexpressing the apoptosis-inhibitory protein bcl-2 were also tested, as was a line of control cells transfected with the same vector, but lacking bcl-2 (39). In each case, treatment was for 5 h during serum deprivation. In the case of deferoxamine, cells were also pretreated overnight prior to the onset of serum deprivation. Assays were performed using both the fluorogenic and the radioactively labeled substrates and relative survival was assessed in each case at 24 h in the same set of cells. Each survival agent suppressed the activation of DEVD-AFC (Table I) and tPARP cleavage activity (Fig. 3). As shown in Table I, there was a good quantitative correlation between the capacities of the various molecules to promote survival and to block activation of cleavage. In addition, lysates of bcl-2-overexpressing cells in particular showed almost complete absence of DEVD-AFC and tPARP cleavage, whereas, in contrast, lysates from control-transfected cells showed the expected levels of cleavage activity (Figs. 3 and 7). The pharmacological agents also inhibited activation of DEVD-AFC cleavage activity in neuronally differentiated PC12 cells, assessed 24 h after NGF withdrawal (8-(4-chlorophenylthio)-cAMP by 100%, ATA by 80%, NAC by 66%, SNAP by 89%, and deferoxamine by 94%, in a mean of two experiments).
Actinomycin D, an inhibitor of transcription, and flavopiridol, a cyclin-dependent kinase inhibitor, block death of NGFdeprived neuronally differentiated PC12 cells but not of naive PC12 cells (36,43). Lysates of NGF-differentiated cells that had been deprived of NGF for 24 h, but treated with 10 M actinomycin D or 3 M flavopiridol, showed little cleavage of DEVD-AFC or tPARP ( Fig. 8 and Table I). Parallel assessment at day 3 following NGF/serum withdrawal confirmed that both agents were effective in promoting survival. Naive PC12 cells were also treated with actinomycin D and flavopiridol and, as anticipated, there was no promotion of survival. There was also no blockade of the stimulation of cleavage activity (data not shown). This indicates that the inhibition of the cleavage acti-

FIG. 3. tPARP cleavage by lysates of naive PC12 cells following various times of serum deprivation and inhibition of activation by treatment with a variety of agents. PC12 cells
were serum-deprived in the presence or absence of a variety of agents that promote survival. The agents used were: 100 ng/ml NGF, 100 M ATA, 100 M 8-(4chlorophenylthio)-cAMP, 60 mM NAC, 200 M SNAP, and 3 mM deferoxamine. At the indicated time points, the cells were harvested, and the corresponding lysates (containing 4 g of protein) were incubated for 2 h at 37°C with tPARP. The samples were then resolved by SDS-PAGE on a 12% polyacrylamide gel, and fluorography was performed. The 48-kDa substrate yielded products of approximate lengths of 28 and 20 kDa, denoted by arrows. PC12neo.1 and bcl-2.1 cell lysates were processed in the same fashion.

FIG. 4. Inhibition of tPARP cleavage by DEVD-CHO.
Lysates of naive PC12 cells (4 g of protein), harvested at 5 h following serum deprivation, and of neuronally differentiated PC12 cells (16 g of protein), harvested at 24 h following NGF/serum deprivation, were incubated with radioactively labeled tPARP for 2 h at 37°C, in the presence or absence of 50 nM DEVD-CHO. The samples were subsequently resolved by SDS-PAGE on a 12% polyacrylamide gel, and fluorography was performed. Cleavage products are denoted by arrows. vation in the neuronally differentiated cells was due to an upstream action of these agents on the survival/death mechanism, rather than to a direct effect on aspartase activation.
Taken together, these findings in naive and neuronally differentiated PC12 cells demonstrate that a variety of agents that inhibit or delay death also block the activation of DEVD-AFC/tPARP cleavage activity. The blockade of the protease activation appears to correlate with the promotion of survival.
Activation of c-Jun Kinase Activity Temporally Precedes Activation of DEVD-AFC Cleavage-Recent findings indicate that activation of c-Jun kinase accompanies, and appears to be necessary for, death of trophic factor-deprived PC12 cells (32,36,43). To determine the temporal sequence in which c-Jun kinase and DEVD-AFC cleavage activities are activated, each was monitored in the same set of cells following trophic factor deprivation. In two independent experiments, one of which is shown in Fig. 9, activation of c-Jun kinase activity clearly preceded that of DEVD-AFC cleavage.
Quantitative Dissociation between the Ability of zVAD-FMK to Inhibit Activity/Activation of DEVD-AFC/tPARP Cleavage and to Promote Survival-As mentioned above, zVAD-FMK is the most potent and effective pseudosubstrate ICE inhibitor that we have tested in terms of promoting survival of serumdeprived PC12 cells. We, therefore, investigated whether zVAD-FMK affects the activity/activation of DEVD-AFC/ tPARP cleavage in intact, serum-deprived PC12 cells. Cultures were deprived of serum for 5 or 18 h in the presence of various concentrations of zVAD-FMK and washed free of the drug, and the cell lysates were assessed for their capacity to cleave DEVD-AFC or tPARP (Figs. 10 and 11). At both time points, exposure of the cells to as little as 1 M peptide resulted in substantial diminution of activity in the lysates (Ͼ50%); at 5 M, activity was less than 15% of that in control lysates. To assess the possibility that residual unbound zVAD-FMK present in cell extracts might interfere with the assay by inhibiting  substrate cleavage, we mixed lysates of control, serum-deprived cells with lysates of serum-deprived cells treated with 10 M zVAD-FMK. No inhibition of DEVD-AFC cleavage activity was observed in the mixed extracts (data not shown). zVAD-FMK is an irreversible inhibitor and would be expected to remain bound to target protease(s) in the cell extracts. Thus, in living cells deprived of trophic support, zVAD-FMK inhibits either the activation or activity of one or more DEVD-AFC/ tPARP-cleaving proteases.
We next compared the potency with which zVAD-FMK blocks DEVD-AFC cleavage activity/activation with its potency for inhibition of cell death in the same set of cultures. The data in Figs. 10 and 11 show that the potencies for the two actions do not match. Approximately 50% of the activity/activation was blocked by 1 M zVAD-FMK, whereas 10 -25 M was needed to inhibit death by 50%. DISCUSSION This study comprised three major aims. The first was to determine whether cell death brought on by withdrawal of trophic support in PC12 cell cultures is associated with activa-tion of an aspartase. The second was to position this activation relative to other elements that regulate cell death. The third was to test the causal role of the aspartase activation in the cell death process.
Trophic Factor Deprivation Activates a CPP32-like Aspartase-Our results demonstrate that there is no detectable induction of YVAD-AFC/pIL-1␤ cleavage activity in PC12 cells following withdrawal of trophic support in PC12 cell cultures. This is consistent with our prior observation that such treatment does not increase secretion of IL-1␤ in the same model (31). In contrast, an aspartase cleaving DEVD-AFC and tPARP is activated in naive and neuronally differentiated PC12 cells prior to their death following trophic factor deprivation. Cleavage activity was of lower magnitude and more protracted in neuronally differentiated compared to naive cultures, consistent with the more gradual evolution of cell death in the former. We have not purified the responsible aspartase, but its kinetic Neuronally differentiated PC12 cells were NGF/serum-deprived for 24 h in the presence of no additives or 100 ng/ml NGF, 3 mM flavopiridol or 10 M actinomycin D. Lysates of these cells (10 g of protein) were then incubated with tPARP at 37°C for 2 h. The samples were then resolved by SDS-PAGE on a 12% polyacrylamide gel, and fluorography was performed. The arrow denotes the 28-kDa cleavage product.
FIG. 9. Parallel assessment of DEVD-AFC cleavage and c-Jun kinase activation. At the indicated time points following serum deprivation, PC12 cells were harvested for the c-Jun kinase (Ⅺ) and the DEVD-AFC cleavage (ࡗ) assay. c-Jun kinase activity was measured as reported previously (32,42,43). The two activities were normalized to each other on a scale of arbitrary units so that they would be comparable. Similar results were achieved in another experiment under the same conditions. properties (K m for DEVD-AFC and K i for DEVD-CHO) resemble closely those reported for CPP32 (15,28). Moreover, DEVD-CHO inhibited tPARP cleavage at nanomolar concentrations, and it was at least 100 times more potent than zVAD-FMK or AcYVAD-CMK in its ability to inhibit the DEVD-AFC cleavage activity in vitro, consistent with the relative specificity of inhibition by DEVD-CHO that has been observed for purified CPP32 (15). In addition, the cleavage products of approximately 28 and 20 kDa would be consistent with a cleavage at residue 216 of PARP, as has been reported for CPP32 (15,16). Recently, it has been reported that Mch2-␣, CMH1, Nedd2, TX, and ICE are also able to cleave PARP (17,18,51). The exact kinetic properties of these reactions have not been reported, but it appears that the affinity of these aspartases for PARP may be considerably lower than in the case of CPP32. Taken together, our findings indicate that trophic factor withdrawal from PC12 cells activates an aspartase activity that is distinct from ICE and that resembles or is equivalent to CPP32.
Ordering Cell Death Pathways-The detection of an aspartase activity that increases under conditions that lead to neuronal cell death has permitted us to determine its place relative to other steps in the apoptotic pathway. A variety of agents that inhibit death of naive or neuronally differentiated PC12 cells following withdrawal of trophic support suppressed activation of DEVD-AFC/tPARP cleavage. These results indicate that all of these agents, with their varying modes of action, act at some point upstream of the aspartase activation.
The suppression of DEVD-AFC/tPARP cleavage activation in serum-deprived PC12 cells overexpressing bcl-2 merits special attention. In mammalian systems, it has been shown that cell death caused by overexpression of ICE-like proteases can be protected by bcl-2 overexpression (29). One interpretation of such findings has been that bcl-2 acts downstream of the aspartases. An alternative explanation that would be consistent with our findings and with studies in C. elegans (6) is that bcl-2 acts upstream to interfere with activation of both endogenous and overexpressed members of the aspartase family.
The inhibition of aspartase activation by ATA is also of particular interest. ATA is an endonuclease inhibitor but also stimulates several proximal signaling pathways (phosphatidyl-inositol-3-kinase, phospholipase C, and ERK phosphorylation) in PC12 cells (52). Our results placing ATA upstream of aspartase activation would be more consistent with an action early in the apoptotic pathway rather than with a late action, such as direct inhibition of DNA fragmentation.
We and others have hypothesized that neuronal cell death caused by withdrawal of trophic support is due to an aberrant attempt at cell cycle re-entry. In support of this, agents that block PC12 cell proliferation, including permeant cAMP derivatives, the G 1 -S inhibitor deferoxamine, and the cyclin-dependent kinase inhibitor flavopiridol, all promote survival of trophic factor-deprived PC12 cells and sympathetic neurons (43,46,50). Our findings that these agents also suppress stimulation of DEVD-AFC/tPARP cleavage activity suggest that the hypothesized attempt at cell cycle entry/re-entry occurs upstream of the aspartase activation.
It has been shown in a number of neuronal systems, in vitro and in vivo, that programmed cell death may be dependent on active transcription (1,53). Our data, demonstrating that actinomycin D blocks activation of the DEVD-AFC/tPARP-cleaving aspartase in neuronally differentiated PC12 cells, would suggest that amongst the newly transcribed or rapidly turning over genes there should be those that regulate the expression or activation of the aspartase-cleaving DEVD-AFC/tPARP and possibly of other members of the ICE/ced-3 family. By contrast, in naive PC12 cells, actinomycin D does not promote survival and does not inhibit the activation of DEVD-AFC cleavage, implying that the aspartases in this setting can be activated through signaling of molecules that are already expressed in the cell. These molecules may represent part of the cell cycle machinery that would be expected to be active in dividing cells.
Our current findings, when viewed in concert with previous work showing differential effects of various survival agents on c-Jun kinase activation in the same system (32), permit further ordering of the cell death pathway. We showed previously that NGF, bcl-2, ATA, NAC, cAMP, and the NO generator DETA⅐NO suppress c-Jun kinase activation, whereas deferoxamine, actinomycin D, and flavopiridol had no effect, despite demonstrable effects on survival. This would place the former agents upstream of c-Jun kinase activation and the latter agents either downstream of c-Jun kinase activation or on an independent pathway. To further investigate the relationship between c-Jun kinase and aspartase activation, we performed parallel experiments that show that c-Jun kinase activation precedes DEVD-AFC cleavage activation. This is in agreement with our previous finding that zVAD-FMK does not inhibit c-Jun kinase activation (32). If, therefore, c-Jun kinase activation and the aspartase cleaving DEVD-AFC/tPARP are part of the same pathway, this would place NO generators, cAMP, NAC, ATA, and bcl-2 upstream of both and would position c-Jun kinase activation upstream of DEVD-AFC/tPARP cleavage activation. In this model, flavopiridol, deferoxamine, and actinomycin D should exert their effect at a point following c-Jun kinase but preceding DEVD-AFC/tPARP cleavage activation (Fig. 12A). Alternatively, it is possible that DEVD-AFC/ tPARP cleavage and c-Jun kinase activation occur on separate, parallel pathways of the apoptotic response (Fig. 12B). It is tempting to suggest that in neuronally differentiated PC12 cells, c-Jun kinase activation may be directly involved in activating c-jun and in promoting early gene induction that ultimately leads to the activation of the aspartases.
Dissociation between Blockade of DEVD-AFC/tPARP Cleavage Activity and Promotion of Survival-We observed that for intact cells, there was a discrepancy between the concentrations of zVAD-FMK that block the activity/activation of DEVD-AFC/tPARP cleavage and that promote survival. The latter required approximately 10-fold higher concentrations. This difference cannot be simply a requirement for stringent inhibition of cleavage activity; other treatments that inhibited DEVD-AFC/tPARP cleavage activation by 80 -90% promoted significant rescue of cells, whereas a concentration of zVAD-FMK (5 M) that inhibited activity/activation to a similar extent provided no rescue above controls. Our data would, therefore, support the idea that although regulation of DEVD-AFC/ tPARP cleaving activity correlates with cell death, it is not the major determinant of apoptosis in our paradigm. Thus, some other aspartase(s) with less affinity for zVAD-FMK may be the death effector(s). In support of this proposition, mice with a null mutation for PARP do not appear to show any differences in programmed cell death (54). Moreover, recent findings from our laboratories implicate an aspartase other than CPP32 in death of PC12 cells and sympathetic neurons caused by withdrawal of trophic support. 2 It is possible that in other systems, the aspartase(s) responsible for cleaving DEVD-AFC/tPARP has a more direct role as death effector(s). However, it should be noted that much of the evidence for this is based on studies with inhibitors whose selectivity within the aspartase family of proteases is incompletely established.
The relative positions within the apoptotic pathway for the CPP32-like aspartase and the postulated "death" aspartase are not entirely clear. When applied to intact cells, zVAD-FMK, a competitive irreversible inhibitor of ICE-like proteases, may act either by directly blocking the CPP32-like aspartase itself or by inhibiting another member of the ICE/ced-3 family that is, in turn, responsible for activating DEVD-AFC/tPARP-cleavage. Thus, one possibility is that the CPP32-like aspartase and the "death" aspartase(s) are activated in parallel, possibly by means of the same upstream mechanism. Alternatively, if the two are on the same pathway, the "death" aspartase(s) should be upstream. The converse could not be true, because in that case activation of the "death" aspartase(s), and therefore apoptosis, would be inhibited by the low doses of zVAD-FMK sufficient to suppress DEVD-AFC/tPARP cleavage activity/ activation.