Distinct Caspase Cascades Are Initiated in Receptor-mediated and Chemical-induced Apoptosis*

Release of cytochrome c is important in many forms of apoptosis. Recent studies of CD95 (Fas/APO-1)-induced apoptosis have implicated caspase-8 cleavage of Bid, a BH3 domain-containing proapoptotic member of the Bcl-2 family, in this release. We now demonstrate that both receptor-induced (CD95 and tumor necrosis factor) and chemical-induced apoptosis result in a similar time-dependent activation of caspases-3, -7, -8, and -9 in Jurkat T cells and human leukemic U937 cells. In receptor-medi-ated apoptosis, the caspase inhibitor, benzyloxycarbonyl-Val-Ala-Asp fluoromethyl ketone (Z-VAD.FMK), inhibits apoptosis prior to commitment to cell death by inhibiting the upstream activator caspase-8, cleavage of Bid, release of mitochondrial cytochrome c , processing of effector caspases, loss of mitochondrial membrane potential, and externalization of phosphatidylserine. However, Z-VAD.FMK inhibits chemical-induced apoptosis at a stage after commitment to cell death by inhibiting the initiator caspase-9 and the resultant postmitochondrial activation of effector caspases. Cleavage of Bid but not release of cytochrome c is blocked by Z-VAD.FMK demonstrating that in chemical-induced apoptosis cytochrome c release is caspase-independent and

Apoptosis is a major form of cell death characterized by a series of stereotypic morphological features. It occurs in two phases, an initial commitment phase followed by an execution phase involving the condensation and fragmentation of nuclear chromatin, dilation of the endoplasmic reticulum, and alterations to the cell membrane resulting in recognition and subsequent phagocytosis of the cell (1,2). Caspases, a family of cysteine proteases, play a critical role in the execution phase of apoptosis and are responsible for many of the biochemical and morphological changes associated with apoptosis (3,4). It has been proposed that "initiator" caspases with long prodomains, such as caspase-8 (MACH/FLICE/Mch5), either directly or indirectly activate "effector" caspases, such as caspases-3, -6, and -7 (3,5,6). These effector caspases then cleave intracellular substrates, such as poly(ADP-ribose) polymerase (PARP) 1 and lamins, during the execution phase of apoptosis. Caspase-8 is the most apical caspase in CD95 (Fas/APO-1)-induced apoptosis (7,8). Triggering of the CD95 receptor with its cognate ligand or agonistic antibody results in receptor trimerization and recruitment of CD95 receptor-associated protein with death domains (FADD/MORT1), which in turn binds to the death effector domains in the N-terminal region of caspase-8, resulting in its activation. As caspase-8 can activate all known caspases in vitro (6), it is a prime candidate for an initiator caspase in many forms of apoptosis in addition to CD95-induced apoptosis. Procaspase-9 has also been proposed as an initiator caspase; in the presence of dATP and cytochrome c, its long N-terminal domain interacts with Apaf-1 resulting in the activation of caspase-9 (9,10). Active caspase-9 can then activate the effector caspases-3, -6, and -7 (10,11). Thus there are at least two major mechanisms by which a caspase cascade resulting in the activation of effector caspases may be initiated as follows: one involving caspase-8 and the other involving caspase-9 as the most apical caspase.
Cytochrome c, which is usually present in the mitochondrial intermembrane space, is released into the cytosol following the induction of apoptosis by many different stimuli including CD95, tumor necrosis factor (TNF), and chemotherapeutic and DNA-damaging agents (12)(13)(14). Mitochondria have been proposed to act as an amplifier in CD95-induced apoptosis when activation of caspase-8 cleaves a cytosolic substrate leading to release of cytochrome c (15,16). Release of mitochondrial cytochrome c and activation of caspase-3 is blocked by anti-apoptotic members of the Bcl-2 family, such as Bcl-2 and Bcl-X L , and promoted by proapoptotic members, such as Bax and Bak (13,17,18). In chemical-or irradiation-induced apoptosis, cytochrome c release appears to be caspase-independent as it is not inhibited by the broad spectrum cell-permeable caspase inhibitor, Z-VAD.FMK (13,17,19,20). Mechanisms for the release of mitochondrial cytochrome c include opening of a mitochondrial permeability transition pore, the presence of a specific channel for cytochrome c in the outer membrane, or mitochondrial swelling and rupture of the outer membrane but without loss of mitochondrial membrane potential (14). None of these mechanisms appears generally applicable, as release of cytochrome c occurs in cells with normal mitochondrial membrane potential (13,17) and by a mechanism independent of rupture of the outer mitochondrial membrane (20). Two recent studies have highlighted another possible mechanism of mitochondrial cytochrome c release, involving Bid, a BH3 domain-containing proapoptotic Bcl-2 family member. Cleavage of Bid by caspase-8 results in translocation of the cleaved Bid to the mitochondria where it induces the release of cytochrome c, being 500-fold more potent than Bax (21,22). The BH3 domain of Bid is essential both for its proapoptotic activity and its ability to induce the release of cytochrome c (22,23).
In this study we address the order in which caspases are activated in receptor-mediated and chemical-induced apoptosis. Our data support the hypothesis that caspase-8 and caspase-9 are the most apical caspases in receptor-mediated and chemical-induced apoptosis, respectively. In chemical-induced apoptosis, cytochrome c release is caspase-independent and is not mediated by cleavage of Bid in contrast to receptormediated apoptosis. We propose that caspases act solely as executioners of apoptosis in chemical-induced apoptosis, whereas in receptor-mediated apoptosis they also form an integral part of the cell death-inducing mechanism.

EXPERIMENTAL PROCEDURES
Materials-Media and serum were purchased from Life Technologies, Inc. (Paisley, UK). Z-VAD.FMK and Z-DEVD.AFC were from Enzyme Systems Inc. (Dublin, CA), and Z-IETD.CHO was kindly provided by Professor L. Rubin (Eisai London Research, London, UK). Anti-CD95 monoclonal antibody was obtained from Upstate Biotechnology Inc. (Lake Placid, NY). Annexin V/FITC kit was from Bender Medsystems (Vienna, Austria). DiOC 6 (3) was purchased from Molecular Probes (Eugene, OR). All other chemicals and human recombinant TNF-␣ were from Sigma (Poole, UK).
Cell Culture and Quantification of Apoptosis by Flow Cytometry-Jurkat T cells (clone E6-1) were obtained from ECACC and cultured in RPMI 1640 containing 10% fetal bovine serum and 1% Glutamax. Apoptosis in U937 cells was induced with etoposide (25 M) or TNF-␣ (10 ng/ml) and cycloheximide (0.9 M) (24). Some cells were treated for 1 h with Z-VAD.FMK prior to exposure to the apoptotic stimulus. Apoptosis was quantified by PS exposure or by loss of the mitochondrial membrane potential, assessed with DiOC 6 (3) (20,25).
Western Blot Analysis-Cell samples were prepared as described (26). Proteins were resolved on 12-15% SDS-polyacrylamide gels and blotted onto nitrocellulose (Hybond-C extra, Amersham, Bucks, UK). Caspases-3 and -7 were detected as described previously (26). A rabbit polyclonal antibody to caspase-8 was raised against the large subunit of caspase-8 (amino acids 210 -374). The antibody obtained was characterized by enzyme-linked immunosorbent assay and Western blot analysis, which verified that the antibody recognized intact procaspase-8 and the p43 and p18 subunits. An antibody to caspase-9 was also raised, which recognized both the inactive proform and the activated ϳ37and 35-kDa processed forms. Cytochrome c antibody was purchased from PharMingen (San Diego, CA). The Bid antibody was kindly provided by Dr. X. Wang (22). The PARP antibody was obtained from Dr. G. Poirier (Laval University, Quebec, Canada).
Preparation of Cytosol for Measurement of Cytochrome c-Cells were collected at the indicated times and washed once in ice-cold phosphatebuffered saline. Cell pellets were resuspended in cytosol extraction buffer, and cytosolic extracts were prepared by the method described previously (19).
Preparation of Cell Lysates for in Vitro Caspase Activation Studies and Immunodepletion of Caspases-Jurkat cell lysates were prepared as described previously (26) and activated by addition of dATP (2 mM), cytochrome c (0.25 mg/ml), and MgCl 2 (2 mM) (12). The proteolytic activity (cleavage of Z-DEVD.AFC) of the lysate was measured as described previously (26). Immunodepletion of capsase-8 or -9 from Jurkat cell lysate was performed as described previously (9).

RESULTS
Anti-CD95 Antibody and Etoposide Induce a Similar Timedependent Processing of Caspases-Both CD95 antibody and etoposide caused a time-dependent induction of apoptosis in Jurkat T cells, as assessed both by an increase in externalization of phosphatidylserine (PS) or by a decrease in mitochondrial membrane potential (⌬⌿ m ) ( Figs. 1 and 2). This was accompanied by a similar time-dependent processing of caspase-3, -7, -8, and -9 (Figs. 1 and 2). In Jurkat T cells, caspase-3 was present primarily as its intact 32-kDa proform (Fig. 1A, lane 1). Induction of both chemical-and receptormediated apoptosis resulted in loss of the proform of caspase-3 and appearance of three immunoreactive fragments of ϳ20 kDa (p20), ϳ19 kDa (p19), and ϳ17 kDa (p17), following initial cleavage at Asp-175 and then at Asp-9 and Asp-28 (27). Processing was first detected after 2-3 h treatment with either stimulus (Figs. 1A and 2A, lanes 2-7).
Caspase-7 was present in control Jurkat T cells primarily as a ϳ35-kDa protein (Fig. 1B, lane 1). Treatment with both anti-CD95 antibody and etoposide resulted in a time-dependent processing of caspase-7 accompanied by the formation of two major products. These were a ϳ32-kDa fragment, which probably represents the loss of the prodomain at DSVD 23 2A, and a ϳ19-kDa (p19) fragment, which corresponds to the catalytically active large subunit (Figs. 1B and 2B, lanes 2-7) formed following cleavage at IQAD 198 2S (27). Processing of caspase-7 was first observed 2 h after treatment with either stimulus.
In untreated Jurkat T cells, caspase-8 was present primarily as two isoforms of ϳ55 kDa (Fig. 1C, lane 1), possibly corresponding to caspase-8a and -8b (7, 28). Exposure to both anti-CD95 antibody and etoposide resulted in a time-dependent processing of caspase-8 initially to two fragments of ϳ43 and 41 kDa (p43 and p41, respectively), corresponding to cleavage of both caspase-8a and -8b between the large and small subunits. This was followed by the appearance of a p18 subunit resulting from removal of the death effector domains from the 43-and 41-kDa fragments (Figs. 1C and 2C, lanes 2-7) (6,28). An increase in the processing of caspase-8 was first observed 2 and 2.5 h after CD95 and etoposide treatment, respectively.
Thus, induction of apoptosis was accompanied by the activation of both the activator caspases -8 and -9 and the effector caspases-3 and-7, which all appeared to occur simultaneously making it extremely difficult to distinguish the order in which they were activated. In order to further address this problem, we used the broad-spectrum caspase inhibitor Z-VAD.FMK, which inhibits apoptosis in many but not all systems (3).

Z-VAD.FMK Inhibits Different Targets in Etoposide-and CD-95-induced
Apoptosis-Slight inhibition of CD95-induced caspase processing was observed at Z-VAD.FMK (0.1 M), with marked and almost total inhibition at 1.0 and 10 M, respectively (Fig. 1, A-D, lanes 8 -10). Higher concentrations of Z-VAD.FMK were required to inhibit etoposide-induced processing of these caspases (Fig. 2, A-D, lanes 8 -10). Z-VAD.FMK (10 and 25 M) largely but not completely inhibited the processing of caspases-3, -7, and -9 (Fig. 2, A-D, lanes 8 and 9). As caspase-9 processes caspases-3 and -7 (10,29), these data support the suggestion that one of the important targets of Z-VAD.FMK in etoposide-induced apoptosis may be either the processing or the activity of caspase-9. However Z-VAD.FMK (10 M) completely inhibited the processing of caspase-8, which suggested that caspase-8 was activated downstream of caspases-3 and -7. Most importantly these data strongly suggested that the order of caspase activation and the Z-VAD.FMK target(s) were different in etoposide-and CD95-induced apoptosis in Jurkat cells.
The different cellular effects of Z-VAD.FMK on etoposideand CD95-induced apoptosis in Jurkat cells provided further support for this hypothesis. Both CD95 and etoposide induced apoptosis, as assessed either by an increase in externalization of PS (Fig. 3, B and D) or by a decrease in mitochondrial membrane potential (Fig. 3, G and I). Z-VAD.FMK (10 M) completely inhibited CD95-induced apoptosis assessed by both these criteria (Fig. 3, C and H). However, Z-VAD.FMK (25 M) did not inhibit etoposide-induced loss of mitochondrial membrane potential (Fig. 3J) but did inhibit PS exposure (Fig. 3E). Z-VAD.FMK (25 M) also did not inhibit etoposide-induced decrease in cell size as measured by forward light scatter (Fig.  3, E and J). Taken together, these data suggest that in CD95induced apoptosis in Jurkat cells the intracellular target of Z-VAD.FMK is most likely the activator caspase-8 acting upstream of mitochondria, whereas in etoposide-induced apoptosis the target(s) is the activation/processing of caspase-9, which is activated downstream of perturbation of mitochondria. In order to substantiate this hypothesis, we examined the effects of Z-VAD.FMK on the release of mitochondrial cytochrome c.
Z-VAD.FMK Inhibits Receptor-mediated but Not Chemicalinduced Cytochrome c Release-In agreement with other studies, CD95 induced a time-dependent increase of cytochrome c in the cytosol, most probably due to an increased release of mitochondrial cytochrome c (Fig. 4A). Z-VAD.FMK (10 M) inhibited this increase (Fig. 4A), further demonstrating that it inhibited a caspase upstream of the mitochondrial changes. Etoposide also induced a time-dependent increase in cytosolic cytochrome c; however, Z-VAD.FMK (25 M) did not inhibit this increase (Fig. 4B) further supporting the hypothesis that the target of Z-VAD.FMK in etoposide-induced apoptosis is downstream of mitochondria.
In order to determine if such differences between receptormediated and chemically induced apoptosis also occurred in other cells, we examined a human lymphoid tumor cell line, U937, which is sensitive to tumor necrosis factor (TNF-␣). Both TNF/cycloheximide and etoposide caused an induction of apoptosis in U937 cells, assessed by PS externalization, which was accompanied by an increase in cytosolic cytochrome c, processing of caspase-3, and cleavage of PARP, a commonly used measure of caspase-3-like enzymic activity (Fig. 5, A-C). Z-VAD.FMK completely inhibited TNF/cycloheximide-induced apoptosis assessed by all these criteria. In contrast, in etoposide-induced apoptosis Z-VAD.FMK completely inhibited PS externalization and the cleavage of PARP but only partially inhibited processing of caspase-3 and did not inhibit the increase in cytosolic cytochrome c. Thus Z-VAD.FMK was more effective at blocking the activity rather than the processing of caspase-3. Taken together, the data from both Jurkat and U937 cells support the hypothesis that Z-VAD.FMK inhibits a target upstream of mitochondria in receptor (CD95 or TNF)-mediated apoptosis, whereas in etoposide-induced apoptosis the Z-VAD.FMK target is downstream of mitochondria.
Activation of Jurkat Lysate Results in Initial Processing of Caspase-9 -In order to gain further insight into the order of caspase activation in chemical-mediated apoptosis, the processing of various caspases was studied in Jurkat cell lysates, a well established model for understanding postmitochondrial caspase cascades (12). Activation of lysates, which resulted in an increased caspase-3-like DEVDase activity, was accompanied by a time-dependent processing of caspases (Fig. 6). Processing of the effector caspases-3 and -7 was first observed after 5 min, and these caspases were almost completely processed after 60 min (Fig. 6, lanes 1-7). Caspase-9 was very rapidly and extensively processed to both its p37 and p35 fragments, initial processing being observed by 1 min and virtually complete processing noted at 30 min (Fig. 6, lanes 1-7). In marked contrast, processing of caspase-8 was first observed at 30 min, when a small amount was processed only to its ϳ41and 43-kDa fragments (Fig. 6, lanes 1-7). Co-incubation of lysates with two different caspase inhibitors, Z-VAD.FMK or benzyloxycarbonyl-Ile-Glu-Thr-Asp-aldehyde (IETD.CHO), resulted in a marked inhibition of the processing of all the caspases with caspase-9 being somewhat less sensitive to inhibition than the other caspases (Fig. 6, lanes 8 and 9). Thus caspase inhibitors may also block the processing of caspases activated by a caspase cascade downstream of mitochondria.
To elucidate further the role of caspases-8 and -9 in the postmitochondrial caspase cascade, they were immunodepleted from lysates, and the subsequent ability of the lysate to be activated was assessed. In the presence of dATP/cytochrome c, lysates from control Jurkat cells exhibited a marked DEVDase activity ( Table I). Immunodepletion of caspase-8 caused only a very slight decrease in DEVDase activity demonstrating that caspase-8 contributed little to this activity (Table I). Western blot analysis demonstrated that immunodepletion of caspase-8 resulted only in loss of this caspase but not of other caspases (data not shown). In contrast, immunodepletion of caspase-9 resulted in complete inhibition of DEVDase activity (Table I) without loss of caspases-3, -7, and -8 (data not shown) demonstrating the key role of caspase-9 in the postmitochondrial processing of caspases. Taken together these results lend strong support to the hypothesis that caspase-9 is the first caspase activated in a caspase cascade following perturbation of mitochondria and release of cytochrome c. Z-VAD.FMK Inhibits Cleavage of Bid in Receptor-and Chemical-mediated Apoptosis-Cleavage of Bid is important for the release of mitochondrial cytochrome c in CD95-induced apoptosis (21,22). We wished to investigate whether this mechanism of cytochrome c release is important in other receptormediated cell deaths or in chemical-mediated apoptosis. Bid was present as an ϳ26-kDa protein in control Jurkat cells and was cleaved initially to two major fragments of ϳ15 and 14 kDa (Fig. 7A), most probably following cleavage at LQTD 60 2G and IEAD 75 2S (21,22). In CD95-mediated apoptosis, cleavage of Bid was first observed at 30 min (Fig. 7A). Z-VAD.FMK inhibited cleavage of Bid in a concentration-dependent manner (Fig.  7A) commensurate with its ability to inhibit caspase processing and release of cytochrome c (Figs. 1 and 4A). Etoposide also induced a time-dependent cleavage of Bid in Jurkat cells with a small amount of fragmentation first being observed at 2 h (Fig. 7B). Z-VAD.FMK inhibited this cleavage in a concentrationdependent manner (Fig. 7B), but it did not inhibit etoposide-induced release of cytochrome c (Fig. 4B). Thus, in etoposide-induced cell death in Jurkat cells, cleavage of Bid was not required for the release of mitochondrial cytochrome c.
In order to extend these studies, we examined cleavage of Bid in other cell types. In U937 cells, induction of apoptosis with either TNF/cycloheximide or etoposide resulted in cleavage of Bid (Fig. 5D, lanes 2 and 4). Z-VAD.FMK prevented cleavage of Bid induced by both apoptotic stimuli (Fig. 5D, lanes 3 and 5); however, it only inhibited cytochrome c release in TNF/cycloheximide-but not in etoposide-induced apoptosis (Fig. 5A).
Recently we have shown that in human monocytic THP.1 cells, induction of apoptosis by either etoposide or N-tosyl-L-phenylalanyl chloromethyl ketone (TPCK), a chymotrypsin-like protease inhibitor, is accompanied by the release of mitochondrial cytochrome c, which is not blocked by Z-VAD.FMK (20). Both these chemical stimuli also induced a time-dependent loss of intact Bid accompanied by the formation of fragments of ϳ15 and 14 kDa, both of which were prevented by Z-VAD. FMK   FIG. 6. Caspase-9 is processed before caspases-3, -7, and -8 in activated lysates. Lysates from Jurkat cells were prepared and activated in the presence of dATP (2 mM) and cytochrome c (0.25 mg/ ml). DEVDase activity was measured fluorimetrically as described under "Experimental Procedures." Processing of caspase-3 (A), caspase-7 (B), caspase-8 (C), and caspase-9 (D) was measured from 0 -60 min by Western blot analysis as described in the legend to Fig. 1.

TABLE I
Immunodepletion of caspase-9 but not caspase-8 inhibits cytochrome c/dATP-dependent activation Jurkat lysate was prepared and immunodepleted with either preimmune serum, caspase-8, or caspase-9 antibody, and the proteolytic activity (DEVDase) of the control lysate and the immunodepleted lysates was then measured as described under "Experimental Procedures."

Z-VAD.FMK Inhibits
Receptor-mediated Apoptosis prior to Involvement of Mitochondria-CD95-mediated apoptosis involves the initial formation of a death-inducing signaling complex (DISC), which is formed following the recruitment of FADD (Fas-associating protein with a death domain) (30). FADD binds through its death effector domain (DED) to one of the two N-terminal DEDs of caspase-8, resulting in its activation (7,8). Active caspase-8 may induce apoptosis either directly following direct activation of other caspases or indirectly following cleavage of cytosolic factors, such as Bid, leading to involvement of mitochondria and release of cytochrome c, which together with Apaf-1 results in activation of caspase-9 and the effector caspases (6,16,21,22). Interestingly no major differences in CD95-induced apoptosis were observed in T cells from caspase-9 Ϫ/Ϫ and Apaf-1 Ϫ/Ϫ knock-out compared with wild type mice (31)(32)(33). However, CD95-induced apoptosis was markedly reduced in Apaf-1 Ϫ/Ϫ compared with wild type embryonic fibroblasts (34) suggesting that the same apoptotic stimulus may utilize different caspases in different cell types. Although mitochondria may be initially bypassed following direct activation of the effector caspases-3 and -7 by caspase-8, these effector caspases may then subsequently disrupt mitochondrial membrane function (Fig. 8). This will lead to a release of intermembrane proteins, such as cytochrome c, resulting in a further activation of caspases and the establishment of a self-amplification loop (35). Thus receptor-mediated apoptosis may initially either directly involve or bypass the mitochondria dependent on the relative concentrations of caspase-8, the effector caspases and cytosolic factors, such as Bid, in a specific cell type (16) (Fig. 8).
Anti-apoptotic members of the Bcl-2 family block mitochondrial cytochrome c release (13,17) and also block CD95-and TNF-induced apoptosis in some (16,29,36) but not all circumstances (37,38). Thus in a variety of cell types, apoptotic signaling in response to CD95 or TNF stimulation is regulated at least in part by a Bcl-2 and/or Bcl-x L -inhibitable step. In this regard, CD95 has recently been shown to activate two distinct cell death pathways, one involving FADD, which is Bcl-2-insensitive, and one involving Daxx, which is Bcl-2 sensitive (39).
Furthermore, Bcl-2 protects mice from the lethal hepatic apoptosis induced by anti-CD95 antibody (40). Taken together these studies demonstrate a physiological role for the involvement of mitochondria in many instances of receptor-mediated apoptosis. In the present study, apoptosis induced by two death receptors, CD95 and TNF, was inhibited by low concentrations of Z-VAD.FMK at a very early stage prior to processing of caspases-3, -7, -8, and-9, cleavage of Bid, release of mitochondrial cytochrome c, loss of mitochondrial membrane potential, and externalization of PS. These data strongly support the hypothesis that the primary target of Z-VAD.FMK in receptormediated apoptosis is inhibition of the activation of procaspase-8 prior to any perturbation of mitochondria resulting in the inhibition of all downstream biochemical effects (Fig. 8).
Z-VAD.FMK Inhibits Chemical-induced Apoptosis Downstream of Mitochondria-Following release from mitochondria, cytochrome c binds to Apaf-1 resulting in the activation of caspase-9, which in turn activates the effector caspases responsible for the cleavage of many of the substrates associated with the characteristic biochemical and morphological changes of apoptosis (9). In agreement with data from chemical-and radiation-induced apoptosis in other cells (13,17,19,20), Z-VAD.FMK did not inhibit release of mitochondrial cytochrome c in etoposide-induced apoptosis in both Jurkat T and U937 cells. Neither did it completely inhibit processing of caspases-3, -7, and -9 (Figs. 2, 4, and 5). Taken together with the demonstration of the critical role of caspase-9 in cytochrome c-catalyzed caspase activation in lysates ( Fig. 6 and Table I), these data support the hypothesis that the Z-VAD.FMK target in chemical-induced apoptosis is the processing or the activity of caspase-9. Recent studies on caspase-9 knock-out mice demonstrate that caspase-9 is a critical upstream activator of a caspase cascade in vivo and in some situations is essential for the processing of caspase-3 (31,32).
Cleavage of Bid was observed in chemical-as well as receptor-mediated apoptosis. Bid is most likely cleaved by low concentrations of caspase-8 but can be cleaved by higher concentrations of caspase-3 (21). Thus in chemical-induced apoptosis, inhibition of Bid cleavage by Z-VAD.FMK may be due to inhibition of processing/activity of either caspase-3 or -8. In addition, effector caspases may activate caspase-8 at a later stage of the apoptotic process, resulting in cleavage of Bid and release of mitochondrial cytochrome c, thereby further amplifying the apoptotic program (Fig. 8). Most importantly, in contrast to receptor-mediated apoptosis, Z-VAD.FMK blocked cleavage of Bid but not release of cytochrome c (Figs. 4, 5 and 7) demon- strating that cytochrome c release is caspase-independent and is not mediated by cleavage of Bid in three different models of chemical-induced apoptosis. Whether the mechanism of mitochondrial cytochrome c release involves another proapoptotic BH3 domain-containing member of the Bcl-2 family, such as Bik, Hrk, Blk, or Bim, is not known. Interestingly EGL-1, a Caenorhabditis elegans BH3-containing proapoptotic protein, has recently been isolated, and this functions upstream of CED-9, the C. elegans homolog of Bcl-2, to induce apoptosis (41).
Caspases Are an Integral Part of the Cell Death-inducing Mechanism in Receptor-mediated Apoptosis whereas in Chemical-induced Apoptosis They Act Solely as Executioners-Previous work has demonstrated an increased clonogenic potential of cells treated with CD95 in the presence of caspase inhibitors, such as Z-VAD.FMK (42,43). Taken together with the present and other studies (7,8), it suggests that the activator caspase-8 is an integral component of the cell death-inducing mechanism and that in receptor-mediated apoptosis activation of caspase-8 represents a commitment point to cell death. At a later stage in the apoptotic program other effector caspases are activated and are responsible for the characteristic biochemical and morphological features of the apoptotic phenotype. We cannot exclude a similar role for other activator caspases, such as caspase-10, which like caspase-8 possesses a long N-terminal prodomain with two death effector domains (27).
In chemical-induced apoptosis, Z-VAD.FMK inhibited the activation of caspases and PS externalization but not the loss of mitochondrial cytochrome c (Figs. 2, 4, and 5) or the loss of clonogenic potential (43). Thus in chemical-induced apoptosis, Z-VAD.FMK does not inhibit the commitment to cell death but rather inhibits all the biochemical changes associated with caspase activation occurring following perturbation of mitochondria and loss of cytochrome c. We propose that Z-VAD.FMK inhibits the processing and/or the activity of caspase-9 and thus blocks the subsequent activation of effector caspases, resulting in the block of a self-amplification loop (Fig.  8). Thus in chemical-induced apoptosis, caspase activation oc-curs after commitment to cell death and is primarily responsible for those stereotypic biochemical and morphological changes commonly associated with the apoptotic phenotype. These results have important clinical implications for the use of caspase inhibitors as potential therapeutic agents. Cells, which are committed to die but with caspases inhibited, may die more slowly and with a different morphology from cells dying with characteristic apoptotic morphology. This hypothesis is compatible with recent reports describing a caspaseindependent cell death, such as Bax-or Bak-induced cell death (44,45).
In summary, we have demonstrated that in receptor-mediated apoptosis, Z-VAD.FMK inhibited apoptosis prior to commitment to cell death by inhibiting the upstream activator caspase-8. This resulted in inhibition of cleavage of Bid, loss of mitochondrial cytochrome c, processing of caspases, loss of mitochondrial membrane potential, and externalization of PS. In chemical-induced apoptosis, Z-VAD.FMK inhibited apoptosis downstream of cell death commitment by inhibiting the initiator caspase-9 and the resultant postmitochondrial activation of effector caspases, cleavage of Bid, and PS externalization but not loss of cytochrome c or mitochondrial membrane potential. Thus in chemical-induced apoptosis, a mechanism other than cleavage of Bid is responsible for the release of cytochrome c. FIG. 8. Scheme for receptor-and chemical-induced apoptosis. Triggering of death receptors, such as CD95 (Fas/ Apo-1) or TNF, results in recruitment of adapter molecules, such as FADD and TRADD. The DED of FADD binds to the DED of caspase-8 leading to the oligomerization and activation of caspase-8. Caspase-8 may directly activate the effector caspases, Ϫ3 and Ϫ7, responsible for many of the biochemical and morphological changes associated with the apoptotic phenotype. Alternatively, caspase-8 can act indirectly by cleaving cytosolic factors, such as Bid, leading to the release of cytochrome c and activation of caspase-9 with subsequent activation of the effector caspases. Active caspases-3 or -7 may then further activate procaspase-9 or alter mitochondrial membrane permeability setting up a self-amplification loop (----) or they may cleave DNA fragmentation factor (DFF) or inhibitor of caspase-activated DNase (ICAD) leading to the internucleosomal cleavage of DNA. In chemical-mediated apoptosis, undefined signals lead to perturbation of mitochondria (Mit) and loss of cytochrome c and then activation of caspases.