Caspase-10 triggers Bid cleavage and caspase cascade activation in FasL-induced apoptosis.

In contrast to caspase-8, controversy exists as to the ability of caspase-10 to mediate apoptosis in response to FasL. Herein, we have shown activation of caspase-10, -3, and -7 as well as B cell lymphoma-2-interacting domain (Bid) cleavage and cytochrome c release in caspase-8-deficient Jurkat (I9-2) cells treated with FasL. Apoptosis was clearly induced as illustrated by nuclear and DNA fragmentation. These events were inhibited by benzyloxycarbonyl-VAD-fluoromethyl ketone, a broad spectrum caspase inhibitor, indicating that caspases were functionally and actively involved. Benzyloxycarbonyl-AEVD-fluoromethyl ketone, a caspase-10 inhibitor, had a comparable effect. FasL-induced cell death was not completely abolished by caspase inhibitors in agreement with the existence of a cytotoxic caspase-independent pathway. In subpopulations of I9-2 cells displaying distinct caspase-10 expression levels, cell sensitivity to FasL correlated with caspase-10 expression. A robust caspase activation, Bid cleavage, and DNA fragmentation were observed in cells with high caspase-10 levels but not in those with low levels. In vitro, caspase-10, as well as caspase-8, could cleave Bid to generate active truncated Bid (p15). Altogether, our data strongly suggest that caspase-10 can serve as an initiator caspase in Fas signaling leading to Bid processing, caspase cascade activation, and apoptosis.

Cell death is an essential process in the regulation of cellular homeostasis. Dysfunction of the mechanisms involved can lead to human diseases such as cancers, autoimmune diseases (in the case of death defect), and neurodegenerative and immune deficiency diseases (in the case of death excess) (reviewed in Refs. 1 and 2). Two types of cell death have been clearly distinguished, apoptosis (programmed cell death) and necrosis (3). Apoptosis can be characterized by typical sets of changes including plasma membrane blebbing, cellular and chromatin condensation, nuclear fragmentation, and formation of apoptotic bodies. Phosphatidylserine externalization and DNA fragmentation are biochemical features of apoptosis. Most of these processes are mediated by caspases that cleave and inactivate proteins essential for cell survival (for review, see Ref. 4). In contrast to apoptosis, necrosis is an "accidental" cell death characterized by profound plasma membrane alterations and oncosis. Intermediate forms of programmed cell death, namely apoptosis-and necrosis-like cell death, have been recently described (for review, see Ref. 5).
Programmed cell death is essential for elimination of selfreactive lymphocytes and down-regulation of the immune response. Stimulation of Fas receptor (CD95) by FasL (CD95L or CD178) plays a crucial role in inducing lymphocyte programmed cell death. Fas cross-linking triggers activation of both caspase-dependent and -independent pathways (6,7). Caspase-independent cell death requires receptor-interacting protein (RIP) as an effector molecule, but cytotoxic signaling pathways activated by RIP remain largely unknown (6). Caspasedependent pathways have been extensively explored (for a recent review, see Ref. 8). Oligomerization of Fas by FasL leads to successive recruitment of FADD (9) and initiator caspase-8 (10) and -10 (11). Formation of this complex, termed DISC 1 (deathinducing signaling complex) (12), allows caspase-8 and -10 activation (10,11). Both caspases can directly activate effector caspases (13)(14)(15). Once activated, effector caspases specifically cleave and inactivate proteins leading to apoptosis. In addition, caspase-8 cleaves Bid, a pro-apoptotic member of the B cell lymphoma-2 superfamily (16,17). The terminal part of Bid containing a B cell lymphoma-2 homology (BH) 3 domain, namely truncated Bid, translocates to mitochondria and promotes cytochrome c release into the cytosol (16,17). In association with APAF-1 and pro-caspase-9, another initiator caspase, cytochrome c forms the apoptosome complex leading to the activation of caspase-9 that cleaves and activates effector caspases (18).
Jurkat T lymphoma cells contain caspase-8 and -10, and caspase-8 has been shown to be essential for apoptosis induction in response to Fas stimulation (21). The mitochondrial pathway involving cytochrome c release was reported to be critical for caspase cascade activation in Jurkat cells, classified * This work was supported by INSERM, Paul Sabatier University, and Grant 3417 from the Association pour la Recherche sur le Cancer. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Herein, we have shown that FasL can induce apoptosis of caspase-8-deficient Jurkat cells, which was associated to endogenous caspase-10, -3, and -7 activation, as well as Bid cleavage and cytochrome c release. Cell sensitivity to FasL correlated with caspase-10 expression in caspase-8-deficient cells. A caspase-10 inhibitor prevented Bid processing, caspase cascade activation, and apoptosis. In addition, evidence is provided for the first time that Bid is a direct substrate of caspase-10, strongly indicating that, like caspase-8, caspase-10 can lead to activation of the mitochondrial pathway.
Flow Cytometry Analyses-CD95 cell surface expression was determined after incubation of cells for 30 min at 4°C with or without anti-CD95-PE or an irrelevant antibody coupled to PE. To allow study of phosphatidylserine externalization, cells were labeled with Annexin V-fluorescein isothiocyanate (250 ng/ml) and propidium iodide (12.5 g/ml) (Immunotech) for 10 min at 4°C. To allow study of hypodiploidy, cells were washed in phosphate-buffered saline and permeabilized in ethanol (70%) for 10 min at Ϫ20°C. Cells were next incubated for 30 min at 37°C with RNase (1 mg/ml) and propidium iodide (0.1 mg/ml). Percentage of hypodiploid cells carrying DNA content below cells in G 0 /G 1 was quantified by flow cytometry. Analyses were performed on a FACScan (BD Biosciences) cytometer.
Cell-free System Assay-Mitochondria-free cytosolic protein extracts (250 g) from A3 or I9 -2 cells were incubated in the presence of one unit (as defined by the manufacturer) of recombinant caspase-8 or -10 for different times up to 90 min at 37°C in a final volume of 200 l of buffer containing 50 mM HEPES, pH 7.5, 150 mM NaCl. At the indicated times, reaction was stopped by freezing on dry ice. 20 g of protein were subjected to SDS-PAGE and Western blotting analysis.
Alternatively, 1 g of His-tagged mouse recombinant Bid (26) was incubated in the presence of one unit of recombinant caspase at 37°C in a final volume of 60 l of buffer containing 50 mM HEPES, pH 7.5, 150 mM NaCl. At the indicated times, 100 ng of protein were immediately frozen on dry ice and analyzed by Western blotting with anti-His antibody (clone sc8036, Santa Cruz). Alternatively, extracts were separated on 15% SDS-PAGE and stained with Coomassie blue. Of note, mouse and human Bid proteins share 63.5% identity and the LQTD2G caspase cleavage motif (16).

FasL Can Trigger Apoptosis in Caspase-8-deficient Jurkat
Cells-Caspase-8 has previously been reported to be essential in anti-Fas-induced Jurkat apoptosis (21). However, two different groups have recently shown that caspase-8-independent cell death occurs in response to high FasL concentration (6,7). Both studies reached the conclusion that this caspase-8-independent cell death is a form of necrosis rather than apoptosis. Herein, we have re-examined caspase-8-independent cell death in Jurkat cells. First, we compared the sensitivity to FasL of parental (clone A3), caspase-8-deficient (clone I9 -2), and FADD-deficient (clone I2-1) Jurkat cell lines (21,28). Caspase-8 and FADD expression was analyzed by Western blotting to confirm respective protein deficiency in I9 -2 and I2-1 clones (Fig. 1A). Flow cytometry analysis indicated that the various cell lines exhibited similar Fas expression on the cell surface (Fig. 1B). When incubated at a low FasL concentration (15 ng/ml), apoptotic features were seen only in A3 cells in agreement with previous reports (21,28). When a higher FasL concentration (500 ng/ml) was used, caspase-8-deficient cells displayed an apoptotic phenotype with pronounced cellular condensation and nuclear fragmentation (Fig. 1C). Flow cytometry analysis indicated that 54.4 Ϯ 4.4% and 11.5 Ϯ 3.6% (mean Ϯ S.D. of three independent experiments) of caspase-8deficient cells were scored Annexin-V-positive when incubated in the presence or absence of 500 ng/ml FasL, respectively. FADD-deficient cells completely resisted FasL-induced apoptosis at all doses tested (up to 500 ng/ml) (Fig. 1C).
FasL Induces Caspase-dependent and Caspase-independent Cell Death in Caspase-8-deficient Jurkat Cells-DNA fragmentation in caspase-8-deficient cells was next assessed by measuring hypodiploidy. Although 15 ng/ml FasL did not induce hypodiploidy in these cells, 14.4 Ϯ 1.6% (mean Ϯ S.D. of three independent experiments) of them displayed DNA fragmentation in response to 500 ng/ml FasL ( Fig. 2A). This process was completely abolished by the broad-spectrum caspase inhibitor Z-VAD-fmk, indicating involvement of caspases in DNA fragmentation. In accordance with a previous study (6), cell death was not fully impaired by Z-VAD-fmk in caspase-8-deficient cells even though caspase activation was completely inhibited Caspase-10 in Fas Signaling ( Fig. 3B and data not shown). As a matter of fact, 45.9 Ϯ 3.4% (mean Ϯ S.D. of three independent experiments) of cells were still Annexin-V-positive in the presence of Z-VAD-fmk and FasL (Fig. 2B). The cell size was reduced as evaluated by flow cytometry (Fig. 2C) and microscopic examination (Fig. 2D). It is of interest that Z-VAD-fmk abrogated nuclear fragmentation, but not nuclear condensation and membrane alterations, as illustrated by propidium iodide uptake (Fig. 2D). Thus, Z-VADfmk weakly affected FasL-induced cytotoxicity, but the latter was shifted into cell death without nuclear fragmentation.
FasL Promotes Bid Cleavage, Cytochrome c Release, and Caspase Activation in Caspase-8-deficient Jurkat Cells-Caspase activation in I9 -2 cells was next analyzed by Western blotting (Fig. 3). Cleavage of caspase-3, -7, and PARP was observed in I9 -2 cells incubated for 16 h with FasL at concentrations higher than 60 ng/ml (Fig. 3A). Because caspase-10 can be recruited to and activated at the DISC in Jurkat cells (15,23), we examined caspase-10 expression using a specific anti-caspase-10 antibody (15,23). The level of pro-caspase-10 clearly decreased in I9 -2 cells incubated with FasL, in a dosedependent manner. Interestingly, Bid content also decreased in I9 -2 cells treated with 500 ng/ml FasL. The disappearance of caspase-10 and Bid in I9 -2 cells treated with FasL was completely abolished in the presence of Z-VAD-fmk, strongly suggesting a caspase-dependent proteolytic processing of both proteins (Fig. 3B). Next, we examined the effect of Z-AEVD-fmk, a caspase-10 inhibitor (29). Z-AEVD-fmk (10 M) inhibited the processing of caspase-10, Bid, and PARP (Fig. 4A) as well as caspase-3 activation (data not shown) and hypodiploidy at 16 h (Fig. 4B). A time course experiment of I9 -2 cells incubated with 500 ng/ml indicated that caspase-10 and -7 and PARP were cleaved at as early as 4 h of treatment with FasL (Fig. 5). As a consequence, apoptosis as evaluated by the quantification of hypodiploid cells increased to reach a maximum at 8 h. Cytochrome c efflux was also observed but only at late time points (8 and 16 h). This phenomenon might indicate that early caspase activation did not require cytochrome c release. Alternatively, our assay might not be sensitive enough to detect low amounts of cytochrome c released at early time points. Together, our data strongly support the notion that FasL can promote caspase activation, Bid processing, and cytochrome c release even in the absence of caspase-8 and that caspase-10 is likely involved in these events.
Recombinant Caspase-10 Can Cleave Bid-Active caspase-8 can promote effector caspase activation (14) as well as Bid

Caspase-10 in Fas Signaling
processing leading to cytochrome c release from the mitochondria into the cytosol (16,17). Little is known about caspase-10 substrate specificity. Notably, to our knowledge, it has never been demonstrated that Bid could be a substrate of caspase-10. Thus, we compared the activity of human recombinant caspase-8 and -10 produced in Escherichia coli. A mitochondria-free cytosolic protein extract from parental A3 cells was used as a source of substrate and incubated for different times in the presence or absence of recombinant caspases. Substrate cleavage was next evaluated by Western blotting (Fig. 8). Recombinant caspase-8 and -10 had no or little effect on endogenous caspase-8 in agreement with a previous study (15). In contrast to recombinant caspase-8, recombinant caspase-10 triggered an obvious endogenous caspase-10 processing. Both caspase-8 and -10 activated caspase-3 and -7. Interestingly, Bid concentration clearly decreased after 60 min of incubation in both cases indicating that caspase-10, like caspase-8, was capable of cleaving Bid. Caspase-3 and -7 cleavage, as well as Bid proteolysis, were also effective when a cytosolic protein extract from I9 -2 cells was incubated in the presence of either recombinant caspase, indicating that endogenous caspase-8 did not interfere in our system (data not shown). In addition, both recombinant caspases were equally active toward recombinant Bid protein, therefore suggesting that Bid could serve as a substrate of caspase-10 ( Fig. 9). Bid proteolysis was characterized by the generation of an active form of Bid (p15) (Fig. 9B), which is known to trigger cytochrome c release (16,17). Thus, our in vitro study demonstrates that caspase-10 can serve as an initiator caspase in Fas signaling, leading to Bid processing and cytochrome c release. DISCUSSION The present study demonstrates the existence of a caspasedependent signaling pathway in response to FasL in caspase-8-deficient Jurkat cells. Thus, in contrast to a previous report showing that caspase-8 deficiency leads to a complete block of caspase activation in Fas signaling (21), we have established that FasL can activate the caspase cascade in a caspase-8independent manner. Caspase-8-independent pathway leads to (i) Bid processing (Figs. 3, 4, and 7), (ii) effector caspase activation (Figs. 3 and 5), (iii) DNA fragmentation (Figs. 2, 4, and 7), and (iv) nuclear fragmentation (Figs. 1 and 2). All these events could be inhibited by the broad-spectrum caspase inhibitor Z-VAD-fmk, arguing that caspases are functionally and actively involved in this form of cell death (Figs. 2, 3, 6, and 7). Z-AEVD-fmk, a caspase-10 inhibitor, had a comparable effect, strongly supporting the notion that caspase-10 acts upstream of FasL-induced caspase-8-independent apoptosis. However, this signaling pathway seems to play a minor function in Jurkat cells and can be activated only in the presence of high FasL concentrations (from 60 ng/ml), possibly because of a low expression level of caspase-10 in I9 -2 cells (15,23). Hence, our study confirms the important role of caspase-8 and highlights the existence of an alternative pathway involving caspase-10 in FasL-induced caspase activation.
A similar observation was made in caspase-8-deficient Jurkat cells treated by tumor necrosis factor-related apoptosisinducing ligand (TRAIL) (15). Endogenous caspase-10 was recruited to and activated at the TRAIL receptor DISC level in Jurkat cells upon stimulation with TRAIL even in the absence of caspase-8. TRAIL-induced apoptosis was delayed but still occurred in caspase-8-deficient cells. Enforced expression of caspase-10a, -10b, or -10d sensitized TRAIL-induced apoptosis, indicating that caspase-10 could function as an initiator caspase. Endogenous caspase-10, similarly to caspase-8, can be recruited to the Fas DISC level in Jurkat cells (15) and in PBL (20). Accordingly, a clear caspase-10 processing was detected in caspase-8-deficient cells in response to FasL (Figs. 3-5 and 7). Opposite findings were reported by Rieux-Laucat and co-workers (30) showing no caspase-10 activation despite PARP cleavage in FasL-treated I9 -2 cells. These authors have discussed the possibility that a caspase-10-independent pathway might be responsible for PARP processing. Herein, use of a highly specific anti-caspase-10 antibody, as demonstrated by two different groups (15,23), allowed us to show an obvious caspase-10 processing in FasL-treated I9 -2 cells (Figs. 3-5  and 7).
The function of caspase-10 in Fas signaling is still controversial. Overexpression of caspase-10 has been shown to bypass caspase-8 deficiency and restore sensitivity to FasL (15,20). In contrast, using a similar strategy Walczak and co-workers (23) did not succeed in sensitizing caspase-8-deficient cells to FasL. Different expression levels of caspase-10 might explain these conflicting observations indicating that high caspase-10 concentration might be required to overcome caspase-8 deficiency.
In addition, caspase-10-enforced expression in I9 -2 cells might perturb cell response to FasL non-specifically due to ectopic expression. Thus, previous studies based on caspase-10 overexpression in I9 -2 cells do not allow drawing any conclusion on the physiological function of caspase-10 in Fas signaling. Herein, we isolated various I9 -2 subcultures that display different caspase-10 expression levels. The sensitivity of these I9 -2 subpopulations correlated with their content of caspase-10. As a matter of fact, PARP and Bid cleavage as well as DNA fragmentation occurred proportionally to caspase-10 expression. Consequently, our data indicate that caspase-10 can initiate caspase cascade activation in Fas signaling. In humans, gene mutations affecting the catalytic enzyme activity of either caspase-8 (31) or caspase-10 (32) are responsible for autoimmune lymphoproliferative syndrome. In vitro, PBL carrying either mutation resisted Fas ligation, implying that both caspase-8 and -10 are involved in FasL-induced apoptosis (31,32).
To date, it is still unclear whether or not caspase-8 and -10 share similar substrates. For instance, both caspases are capable of activating effector caspases in vitro (13)(14)(15). However, whereas caspase-3 was equally activated in FasL-treated I9 -2 cells overexpressing caspase-8 or caspase-10, receptor-interacting protein processing occurred only in caspase-8 expressing cells (20). We therefore carried out in vitro experiments to compare the substrate specificity of human recombinant caspase-8 and -10. In a cell-free system, caspase-8 and -10 cleaved effector caspases as well as Bid (Fig. 8). In addition, caspase-10, like caspase-8, directly processed recombinant Bid leading to active truncated Bid (p15) generation (Fig. 9). Thus, both caspases might be capable of cleaving Bid in vivo and trigger the activation of the mitochondrial pathway in a similar fashion. Accordingly, Bid processing (Figs. 3, 4, and 7) and cytochrome c release (Fig. 5) occurred in FasL-treated caspase-8-deficient cells.
Recent studies have suggested the involvement of a caspaseindependent pathway in T lymphocyte cell death induction. In vivo, transgenic mice overexpressing cytokine response modifier A, a Poxvirus-encoding caspase inhibitor, did not develop lymphoproliferative syndrome (33). In vitro studies demonstrated that FasL can trigger cytotoxic caspase-independent pathways in Jurkat cells and in activated PBL (6,7,30,34). Accordingly, Z-VAD-fmk did not completely abolish I9 -2 cell death in response to 500 ng/ml FasL (Fig. 2). Caspase-independent cell death was characterized by phosphatidylserine externalization and absence of nuclei and DNA fragmentation. Membrane alterations were present, in agreement with previous reports showing that caspase-independent pathways could trigger a necrosis-like cell death (6,7). However, most of the dead cells displayed strong nuclear condensation and decreased cell size. Importantly, we noticed similar alterations in activated PBL that were co-incubated with FasL and Z-VADfmk (data not shown). These observations might indicate that, according to recent definitions (5), FasL-induced caspase-independent pathway triggers apoptosis-like rather than necrosislike cell death.
Collectively, the present study indicates that at least two caspase-8-independent pathways operate in T lymphocyte to mediate Fas-induced cell death: (i) a caspase-dependent pathway involving caspase-10 activation, Bid processing, cytochrome c release, and classical apoptosis, and (ii) a caspaseindependent pathway leading to apoptosis-like cell death.