INTRODUCTION

Cytotoxic T lymphocytes (CTLs)¹ represent the body's primary defense against virus-infected cells (1). Recognition of a target cell bearing a foreign antigen through the T cell receptor (TCR) results in the induction of apoptotic death. During granule-mediated killing, exocytosis leads to the release of granule contents, including a family of serine proteases (the granzymes (2, 3)) and the pore-forming protein perforin (4), into the intercellular space. Polymerization of perforin in the target cell membrane facilitates granzyme entry into the target cell, where they induce apoptosis. In Fas-mediated cytotoxicity, TCR engagement results in up-regulation of Fas ligand (FasL) levels on the surface of the T cell. FasL then ligates the Fas receptor on the target cell thus setting in motion a series of events that results in death of the Fas-bearing cell (5).

A role for ICE (6, 7) in apoptosis was suggested by the discovery that ced-3, a gene isolated from the nematode Caenorhabditis elegans as being essential for programmed cell death during development, had homology to ICE (8). Multiple members of the ICE/Ced-3 family of cysteine proteases have since been isolated (reviewed in Ref. 9), including Ich1 (10), the human homologue of the murine Nedd-2 (11), TX/Ich-2/ICE_relII (12, 13, 14), ICE_relIII (14), Mch2 (15), Mch3 (16), ICE-LAP3 (17), and CPP32/Apopain/Yama (18, 19, 20). CPP32 was identified as the protease responsible for cleavage of the nuclear protein poly(ADP-ribose) polymerase (PARP) during the induction of apoptosis (19, 20). We have recently shown that CPP32 is involved in granule-mediated cytotoxicity since it is an intracellular substrate for granzyme B (21), an enzyme critical for the early induction of DNA fragmentation in target cells (22, 23). Other work has focused on the role of ICE/Ced-3 proteases in Fas-mediated killing (24, 25, 26) but the exact nature of the protease(s) involved has remained elusive.

Here we examine the role of ICE/Ced-3 proteases in Fas-mediated killing. Our data suggest the existence of an alternate pathway to DNA fragmentation which does not involve CPP32, since CPP32 is not cleaved in all cells following Fas ligation, and CPP32 inhibition does not suppress DNA fragmentation during Fas-mediated cytotoxicity.

MATERIALS AND METHODS

The cytolytic hybridoma PMM-1, derived from primary PEL cells, has been described previously (27, 28). Activation was achieved by stimulation for 3 h with PMA (10 ng/ml) and ionomycin (3 µg/ml). The L1210 variant transfected with mouse Fas cDNA (L1210-Fas) was kindly provided by Dr. Pierre Golstein, Centre d'Immunologie INSERM-CNRS de Marseille-Luminy, Marseille, France (29). EL4 and YAC are mouse lymphomas, Jurkat cells are human T cells. All cells were maintained by continuous culture in RHFM (Life Technologies, Inc.) as described previously (30).

⁵¹Chromium, [³H]thymidine, and [³⁵S]Met were obtained from DuPont NEN. Murine anti-human Fas antibody (IgM) was from Upstate Biotechnology Inc., hamster anti-mouse Fas (IgG) was from Pharmingen, and goat anti-hamster IgG was from Jackson ImmunoResearch Laboratories.

Target cells (10⁶) were incubated with appropriate numbers of effector cells at 37 °C for 4 h, and then all cells were lysed in 10 mM Hepes/KOH, pH 7.4, 2 mM EDTA, 0.1% (v/v) CHAPS, 5 mM dithiothreitol. Cellular debris was removed by centrifugation, and then the entire lysate was separated by 12% SDS-PAGE followed by electroblotting to polyvinylidene difluoride. The blot was probed using an antibody directed against the p17 subunit of CPP32. Detection was by probing with a horseradish peroxidase-conjugated secondary antibody, followed by ECL detection (Amersham).

Cytolytic activity, assessed by measuring ⁵¹Cr release from labeled target cells, and DNA fragmentation, assessed by measuring [³H]thymidine release, were as described previously (28). Briefly, labeled targets were incubated with PMM-1 at the indicated effector-to-target ratio in a total volume of 200 µl in 96-well V-bottom plates for 4 h at 37 °C in the presence or absence of the various inhibitors. CTL activity was calculated using the following formula: % lysis = 100 × (sample  spontaneous release)/(total  spontaneous release). Spontaneous release was determined by incubation of targets in the absence of CTL.

L1210-Fas or YAC-1 cells were suspended in AIM-V serum-free media (Life Technologies, Inc.) and treated with 1 µg/ml hamster anti-mouse Fas IgG on ice for 30 min. Cells were then pelleted and resuspended in AIM-V containing 20 µg/ml goat anti-hamster IgG antibody, in order to cross-link ligated receptors. Incubation proceeded at 37 °C for the time points indicated. Cells were lysed in Nonidet P-40 lysis buffer (10 mM Tris, pH 7.4, 10 mM NaCl, 3 mM MgCl₂, 0.5% (v/v) Nonidet P-40), nuclei were pelleted and the supernatant was assayed for protein content using a BCA Protein Assay Kit from Pierce.

Cleavage of Murine Pro-IL-1

The cDNA encoding pro-IL-1 was a gift of Dr. Mike Tocci at Merck Research Laboratories, Rahway, NJ. RNA transcripts were produced by using T7 RNA polymerase (Life Technologies, Inc.) and a Promega In Vitro Transcription Kit. Template DNA was destroyed after the transcription reaction by treatment with E. coli DNaseI (Promega). In vitro translation was performed in a total volume of 50 µl using a rabbit reticulocyte lysate kit from Promega. Each reaction included 40 µCi of [³⁵S]Met. Resulting translation products were analyzed by SDS-PAGE followed by autoradiography. For cleavage assays, 50 µg of total protein from lysates of cells treated with anti-Fas antibody were incubated with 3 µl of [³⁵S]pro-IL-1 in a total volume of 30 µl at 37 °C overnight. Samples were analyzed by SDS-PAGE followed by autoradiography. In other assays, 45 ng of purified human recombinant ICE (generously provided by Nancy Thornberry, Merck Research Laboratories) was added to lysates prior to addition of [³⁵S]pro-IL-1. Analysis of ICE activity was then as outlined above.

Permeability of the inhibitors, as measured by inhibition of IL-1 release, was determined as outlined previously (7). Inhibition of CPP32 activity was determined as follows. Purified human CPP32 was a generous gift of Dr. Don Nicholson of Merck Frosst Canada Inc., Montréal, and was provided at a concentration of 6 units/µl. This stock solution was diluted 1:50 in 50 mM PIPES/KOH, pH 6.5, 2 mM EDTA, 0.1% (w/v) CHAPS, 5 mM DTT for the inhibitor study. The diluted CPP32 (1 µl) was combined with 1 µl of inhibitor stock (diluted in dimethyl sulfoxide) in a final volume of 20 µl and incubated at 37 °C for 15 min to allow binding of the inhibitor to CPP32. Following the preincubation, 5 µl of in vitro-translated, fast protein liquid chromatography-purified [³⁵S]PARP (21) was added to each reaction, and incubation proceeded at 37 °C for an additional 60 min. Reactions were separated by 12% SDS-PAGE, and cleavage was assessed by autoradiography.

RESULTS AND DISCUSSION

Previous work on CTL-mediated cytotoxicity has focused on the role of the apoptotic ICE/Ced-3 cysteine proteases (21, 24, 25, 31, 32). We have recently shown that the CTL-specific serine protease granzyme B, which is required for the early induction of target cell DNA fragmentation (22, 23), cleaves and activates the ICE/Ced-3 protease CPP32 during granule-mediated killing (21). Given these results using granule-mediated killing, we wished to further elucidate the role of ICE/Ced-3 proteases, particularly CPP32, during Fas-mediated killing.

Our first question was whether CPP32 is cleaved during Fas-induced apoptosis. EL4 and YAC-1 (mouse lymphoma) target cells were incubated with PMM-1 effectors for 4 or 24 h. These effectors kill exclusively through the Fas mechanism (28). Cell lysates were then assessed for CPP32 status by SDS-PAGE separation followed by Western blotting using an antibody directed against the large subunit of CPP32. The results, shown in Fig. 1A, indicate that when YAC-1 cells are incubated with effectors for 4 h, a band appears at 17 kDa. This band is not present in the time zero sample and represents cleavage of the CPP32 precursor (p32) to the active form (p17/p12). This cleavage product is not apparent in EL4 lysates, even after 24-h incubation with effector cells, and the cleavage in YAC-1 lysates is partial at best. Similar studies using a variant of L1210 transfected with murine Fas cDNA (L1210-Fas) as targets revealed that CPP32 is not cleaved during Fas killing in these cells (data not shown). Concurrent with this study, DNA fragmentation (assessed as [³H]thymidine release) was measured from YAC-1 and EL4 targets incubated with PMM-1 effectors and was found to be normal (Fig. 1B). Taken together, these results suggest that CPP32 may be cleaved in some cells during Fas-mediated killing, but is not essential for DNA fragmentation. These results are in direct contrast to our studies of granule-mediated killing, where we found complete cleavage of CPP32 in both EL4 and YAC-1 cells exposed to effectors for 4 h (21). In addition we have shown that inhibition of CPP32 activity suppresses target cell DNA fragmentation during granule-mediated killing suggesting a critical role for CPP32 in the induction of target cell death during this mode of cytotoxicity (34). It should be noted that DNA fragmentation during Fas-mediated cytotoxicity is consistently more pronounced in YAC-1 cells, where CPP32 is partially cleaved, than in EL4, in which no cleavage of CPP32 is apparent. Thus, cleavage of CPP32 may affect the efficiency of DNA fragmentation.

In our next series of experiments, we used a battery of peptide inhibitors, summarized in Table I, which were synthesized for their ability to inhibit ICE/Ced-3 proteases. Briefly, inhibitor 1 contains the prototypic sequence of an ICE/Ced-3 inhibitor, VAD, with an amino-terminal phenylalanine which lacks an amino group. Inhibitor 2 is the Ac-YVAD-CHO ICE inhibitor often used to study ICE/Ced-3 proteases (K_i,ICE = 0.76 nM (7)). Inhibitor 3 was originally identified as a CPP32 inhibitor (19) with the structure Ac-DEVD-CHO (K_i,CPP32 = 0.2 nM (19)). Inhibitor 4 is the reduced aldehyde of inhibitor 2 and acts as a negative control. Although extensive kinetic studies have not been performed on CPP32 in the presence of these inhibitors, in an in vitro assay looking at the ability of CPP32 to cleave PARP, inhibitors 1 and 2 inhibit CPP32 in the range 100 nM to 1 µM, far above the K_i of inhibitor 3, suggesting that these peptides are poorly inhibitory toward CPP32 (data not shown).

Table I. Summary of structures and properties of peptide inhibitors Inhibitor Structure Inhibition of ICEa Inhibition of CPP32a Features Reference 1 Ph(Ch2)2C(=O))-Val-Ala-AspC(=O)CH2O-Ph(F)5 1,100,000 M1s1 NDb Competitive, irreversible 33 2 Ac-Tyr-Val-Ala-Asp-CHO 0.76 nM >10,000 nM Competitive reversible 7 3 Ac-Asp-Glu-Val-Asp-CHO ND 0.2 nM Competitive reversible 19 4 Ac-Tyr-Val-Ala-Asp-CH2OH >10,000 nM ND Competitive reversible N. A. Thornberry, unpublished data a  Inhibition constants: k (second order rate constant) for inhibitor 1; Ki for inhibitors 2-4. b  ND, not determined.

Studies using these inhibitors addressed their effect on ⁵¹Cr or [³H]thymidine release (as a measure of membrane integrity or DNA fragmentation, respectively) from labeled target cells, using PMM-1 cells (which kill through the Fas pathway) as effectors. The results using L1210-Fas as target cells are shown in Fig. 2, A and B. Similar results were obtained using EL4 cells as targets (data not shown). In these experiments, only inhibitor 1 had a significant effect on target cell ⁵¹Cr and [³H]thymidine release induced by ligation of the Fas receptor. Intriguingly, inhibitor 3, the CPP32 inhibitor, had no effect on either chromium or thymidine release, suggesting no role for CPP32 in cell death following ligation of Fas. This result is consistent with our finding that CPP32 is not cleaved in these cells during Fas-induced apoptosis (Fig. 1A).

For comparison, we studied the effect of these inhibitors on Fas-induced death of YAC-1 cells, in which CPP32 is cleaved following Fas ligation (Fig. 1A). Again, labeled target cells were incubated with PMM-1 effectors in the presence or absence of the inhibitors, and then ⁵¹Cr or [³H]thymidine release was measured. The data, presented in Fig. 2, C and D, clearly show that inhibitor 1 has a significant effect on both ⁵¹Cr and [³H]thymidine release from these cells following Fas ligation. Additionally, inhibitor 3 (the CPP32 inhibitor) also suppressed [³H]thymidine release from YAC-1 cells. This is in contrast to the results using inhibitor 3 with L1210-Fas cells (Fig. 2, A and B). When taken together with the results from Figs. 1 and 2, A and B, these studies lead to the conclusion that CPP32 is not essential for Fas-induced cell death since it is not always cleaved during Fas-mediated apoptosis. In cells which exhibit no CPP32 cleavage, the CPP32 inhibitor (inhibitor 3) affects neither ⁵¹Cr nor [³H]thymidine release. However, there is a correlation between CPP32 cleavage and inhibition of DNA fragmentation by inhibitor 3, suggesting that CPP32 may be involved in Fas-induced cell death in some cells.

We wondered whether the results using inhibitor 1 could be explained by inhibition of ICE itself, since other investigators have previously found that an ``ICE'' activity is involved in Fas-mediated cytotoxicity (24, 25), a result which we have confirmed using lysates from anti-Fas-treated cells in a fluorogenic assay (data not shown). To evaluate ICE activity, cell lysates from YAC-1 cells treated with anti-Fas antibody were tested for their ability to cleave pro-IL-1 to produce the 17-kDa IL-1. To date, ICE is the only known ICE/Ced-3 protease capable of activating pro-IL-1. [³⁵S]Pro-IL-1, which is translated as a doublet due to the use of an alternate initiation codon, was incubated overnight with lysates from YAC-1 cells treated with anti-Fas antibody for up to 8 h. The results, shown in Fig. 3A, revealed that treatment of YAC-1 cells with anti-Fas did not result in the appearance of any pro-IL-1-cleaving activity. Similar results were obtained using lysates from both L1210-Fas and human Jurkat cells treated with anti-Fas (data not shown). The lack of ICE activity was not due to the presence of a cellular inhibitor of ICE since supplementing the lysates with purified recombinant human ICE resulted in cleavage of pro-IL-1 (Fig. 3B), detected by appearance of the 17.5-kDa mature IL-1. Therefore, the increased ICE activity seen during Fas-mediated apoptosis is not due to activation of ICE itself, and the inhibitory action of inhibitor 1 cannot be explained by inhibition of ICE.

One possible explanation for the lack of effect of some of these inhibitors during Fas-mediated killing would be that the inhibitor fails to enter the target cell (7). Using a whole cell assay for IL-1 release, we have found that although the IC₅₀ values for these inhibitors are relatively high compared to the K_i value for ICE (IC_50,1 = 3 µM; IC_50,2 = 2.5 µM) they are still below the concentration used in these studies (40 µM). Our evidence also supports the idea that inhibitor 3 is able to enter the target cells. In Fas-mediated killing, we have shown that 40 µM inhibitor 3 can suppress DNA fragmentation from YAC-1 cells, which correlates with cleavage of CPP32 in these cells. In addition, we have previously found that 40 µM inhibitor 3 can suppress granule-mediated DNA fragmentation in YAC-1, EL4, and L1210-Fas cells (34). Therefore, it seems likely that the failure of these inhibitors to affect Fas-mediated killing cannot be explained by poor membrane permeability.

These results suggest the involvement of an ICE-like protease during Fas-mediated killing. This protease is suppressed by inhibitor 1 but is not ICE itself since no ICE activity is apparent in these cells following Fas ligation, and the prototypical ICE inhibitor Ac-YVAD-CHO (inhibitor 2) has no effect on either ⁵¹Cr or [³H]thymidine release from labeled targets during Fas-mediated cytotoxicity. The inhibition profile seen with inhibitor 1 suggests that the protease in question goes on to induce both membrane blebbing and DNA fragmentation within the target cell. While CPP32 may be cleaved and activated during this process, its activity is not essential since CPP32 is not cleaved in all cells. However, in cases where CPP32 is cleaved, this activity is involved in DNA fragmentation. Additionally, it seems that a second mechanism exists for the induction of target cell DNA fragmentation which does not proceed through CPP32 since fragmentation proceeds in the absence of CPP32 activation. This is the first demonstration of apoptosis in the absence of CPP32 activation and suggests that a cell may possess multiple mechanisms which culminate in apoptotic cell death. Since Mch3 has similar kinetics for inhibition by Ac-DEVD-CHO (inhibitor 3, K_i,Mch3 = 1.8 nM (16)), these results also suggest that Mch3 is not essential for Fas-mediated apoptosis. The presence of an alternate pathway for induction of DNA fragmentation, which does not involve CPP32, is intriguing and has not been previously demonstrated. Based on our results, we would predict that an ICE/Ced-3 protease exists which is suppressed by inhibitor 1 but not by inhibitor 3 and is involved in the induction of ⁵¹Cr and [³H]thymidine release during Fas-induced apoptosis. Further work to determine the nature of this pathway and the proteases involved will be essential to the understanding of apoptosis and CTL-mediated cytotoxicity and may aid in the design of novel immunosuppressants.