ECH, an Epoxycyclohexenone Derivative That Specifically Inhibits Fas Ligand-Dependent Apoptosis in CTL-Mediated Cytotoxicity¹

Cytotoxic T lymphocytes play a critical role in protection against intracellular pathogens and tumor cells as well as autoimmunity and transplant rejection. To induce apoptosis of target cells, CTL mainly use two distinct pathways that are dependent on Fas ligand (FasL)³ and lytic granules that contain the pore-forming protein perforin and serine proteases termed granzymes (1, 2). CD8⁺ CTL eliminate target cells primarily via the perforin-dependent pathway, whereas the FasL-dependent killing pathway is dominantly used by CD4⁺ CTL (1). The perforin system is essential for the control of viral infection and tumor rejection, and the Fas/FasL system is important for lymphocyte homeostasis (1, 3, 4, 5). However, these two killing systems play different regulatory roles in various physiological and pathogenic situations.

CTL harbor lytic granules that store perforin and granzymes under the acidic environment (6). Upon TCR stimulation, CTL rapidly release lytic granules into the interface between CTL-target conjugates (1, 2). Perforin facilitates the translocation of granzymes A and B into the cytosol without pore formation on the plasma membrane (7). Granzymes A and B are major molecules that induce target cell death (1, 2). Granzyme B initiates caspase-dependent apoptosis that requires the release of proapoptotic mitochondrial factors (8, 9, 10). By contrast, granzyme A triggers a caspase-independent alternate cell death pathway characterized by ssDNA nicks (11, 12). However, in the granule-mediated killing pathway, granzyme B is critical for rapid induction of DNA fragmentation in target cells (13).

Upon TCR stimulation, FasL is newly synthesized and then transported to the cell surface of CTL (1). In the cytoplasmic region, Fas contains the death domain that is required for interaction with the adaptor protein Fas-associated death domain protein (FADD) (3, 4, 5). Membrane-bound FasL triggers Fas oligomerization, which allows the recruitment of FADD to the Fas death domain (3, 4, 5). FADD subsequently binds to procaspase-8 via the mutual interaction of their death effector domain, resulting in the formation of death-inducing signaling complex (DISC) (14). In the DISC, procaspase-8 immediately dimerizes and undergoes self-cleavage, generating the active heterotetramer composed of two large subunits and two small subunits (15, 16). Active caspase-8 cleaves downstream substrates such as procaspase-3 or Bid, essential for apoptosis execution (1, 2, 3, 4, 5).

Membrane-permeable small compounds are expected to be valuable tools to clarify the molecular basis of complex intracellular signaling pathways and to be potential therapeutic drugs as well. Hence, specific modulators of CTL-mediated cytotoxicity are highly useful. Previously, we reported that vacuolar type H⁺-ATPase inhibitor concanamycin A (CMA) is a specific inhibitor of the perforin-dependent killing pathway in target cell lysis mediated by CTL and NK cells (17, 18). However, the FasL-dependent killing pathway is totally insensitive to CMA (18). CMA perturbs acidification of lytic granules and raises the internal pH to around neutral (19). Neutralization of acidic pH induces inactivation of perforin in a Ca²⁺-dependent manner and subsequent proteolytic degradation of perforin by serine proteases (17, 20). To date, CMA has been frequently used to evaluate the contribution of the perforin-dependent killing pathway in cell-mediated cytotoxicity.

The early signal transduction of Fas-mediated apoptosis is a complex process regulated by various cellular proteins exerting proapoptotic and antiapoptotic functions. To search for specific modulators of Fas-mediated apoptosis, we have screened natural products, such as microbial metabolites, and identified several modulators that block or enhance the Fas death signals (21, 22, 23, 24). It was reported that an epoxycyclohexenone derivative, (2R,3R,4S)-2,3-epoxy-4-hydroxy-5-hydroxymethyl-6-(1E)-propenyl-cyclo-hex-5-en-1-one (ECH; see Fig. 1 A), prevents Fas-mediated apoptosis (25). However, the molecular target of ECH remained to be elucidated. Recently, we have shown that ECH inhibits apoptosis mediated by Fas and TNF receptor 1 through preventing activation of procaspase-8 in the DISC (24). By contrast, death receptor-independent apoptosis induced by chemical drugs and UV irradiation was totally insensitive to ECH (24). In the present report we have studied the inhibitory effects of ECH on two distinct CTL-mediated killing pathways. Our results demonstrate that ECH does not affect the perforin-dependent killing pathway, but selectively inhibits the FasL-dependent killing pathway mediated by Ag-specific CTL.

The H-2^d-specific CD8⁺ CTL clone OE4 (26) and the keyhole limpet hemocyanin (KLH)-specific H-2^d (I-E^d)-restricted CD4⁺ CTL clone BK-1 (27) were maintained in RPMI 1640 medium (Invitrogen, Carlsbad, CA) supplemented with 10% (v/v) heat-inactivated FCS (JRH Biosciences, Lenexa, KS), 50 μM 2-ME, 5% (v/v) rat spleen cell-conditioned medium (culture supernatant of rat spleen cells stimulated with 5 μg/ml CMA for 24 h), and penicillin-streptomycin-neomycin antibiotic mixture (Invitrogen). OE4 cells were stimulated with mitomycin C-treated spleen cells prepared from BALB/c mice every 2 wk. BK-1 cells were stimulated with 10 μg/ml KLH (Calbiochem, Darmstadt, Germany) in the presence of mitomycin C-treated BALB/c mouse spleen cells every 2 wk. BALB/c mouse-derived B lymphoma (A20, A20.HL, A20.FO) were maintained in RPMI 1640 medium containing 10% (v/v) FCS, 50 μM 2-ME, and penicillin-streptomycin-neomycin antibiotic mixture. A20.HL cells were the BALB/c B lymphoma transfected with L and H chain genes of anti-TNP IgM Ab (28). A20.FO cells were the Fas-negative variant subcloned from the Fas-positive parent cell line A20.2J (18). DBA/2 mouse-derived T lymphoma L5178Y and the Fas-expressing transfectant of L5178Y (L5178Y-Fas) (29) were maintained in RPMI 1640 medium containing 10% (v/v) FCS, 50 μM 2-ME, and penicillin-streptomycin-neomycin antibiotic mixture.

ECH was isolated from the culture broth of a producing fungal strain using a bioassay-guided purification procedure (24). CMA was purchased from Wako Pure Chemical Industries (Osaka, Japan). Recombinant human soluble FasL was a gift from Dr. J. Tschopp (Institute of Biochemistry, University of Lausanne, Epalinges, Switzerland) and was used as described previously (30).

Target cells were labeled with 37 kBq of [³H]TdR (ICN Biomedicals, Costa Mesa, CA) for 16 h and washed three times before use. The labeled cells (1 × 10⁵ cells/ml, 100 μl) were preincubated with the indicated concentrations of ECH for 1 or 2 h, then mixed with 100 μl of CTL in U-bottom, 96-well microtiter plates. The plates were centrifuged (300 × g, 3 min) and then incubated for 4 h. The drug concentration during coculture with CTL was diluted into half the initial drug concentration. At the end of the culture, 10 μl of 2% Triton X-100 was added to each well, and the cells were lysed by pipetting, followed by centrifugation (600 × g, 5 min). One hundred microliters of supernatants were harvested and measured for radioactivity. Specific ³H-labeled DNA release (percentage) was calculated using the following formula: (experimental cpm − spontaneous cpm)/(maximum cpm − spontaneous cpm) × 100.

Target cells were labeled with 1850 kBq of [⁵¹Cr]sodium chromate (Amersham Biosciences, Piscataway, NJ) in 100 μl of 50% (v/v) FCS for 1 h and washed three times with the medium. The labeled cells (1 × 10⁵ cells/ml, 100 μl) were preincubated with indicated concentrations of ECH for 1 h, then mixed with 100 μl of CTL in U-bottom, 96-well microtiter plates. The plates were centrifuged (300 × g, 3 min) and then incubated for 4 h. One hundred microliters of supernatants were harvested and measured for radioactivity. Specific ⁵¹Cr release (percentage) was calculated using the following formula: (experimental cpm − spontaneous cpm)/(maximum cpm − spontaneous cpm) × 100.

Analysis of conjugate formation was performed as described previously (31). Target cells (5 × 10⁵ cells/ml) were treated with or without ECH for 1 h, then stained with 62.5 μg/ml hydroethidine (Polysciences, Warrington, PA) for 30 min on ice. Effector cells (5 × 10⁵ cells/ml) were stained with 0.25 μM calcein-AM (Molecular Probes, Eugene, OR) for 30 min on ice. Both types of stained cells (5 × 10⁵ cells/ml, 250 μl) were washed twice with the medium, then transferred in a single tube. The cell mixtures were either left untreated or centrifuged (300 × g, 3 min), then incubated at 25°C for 30 min. The cells were resuspended carefully by pipetting and immediately analyzed by FACS.

OE4 cells (1 × 10⁶ cells) were washed with PBS and lysed with 1% Triton X-100, 50 mM Tris-HCl (pH 7.4), and protease inhibitor mixture (Complete; Roche, Mannheim, Germany) on ice for 15 min. After centrifugation (10,000 × g, 5 min), supernatants were collected. Postnuclear lysates (30 μg/lane) were separated by 10% SDS-PAGE and analyzed by Western blotting using ECL detection reagents (Amersham Biosciences). Anti-mouse perforin Ab P1-8 (32) was provided by Dr. H. Yagita (Juntendo University School of Medicine, Tokyo, Japan).

OE4 cells (1 × 10⁶ cells) were washed with PBS and lysed with 0.5% Triton X-100, 10 mM HEPES-NaOH (pH 7.4), 150 mM NaCl, 1 mM CaCl₂, and 1 mM MgCl₂ on ice for 15 min. After centrifugation (10,000 × g, 5 min), supernatants were collected. Postnuclear lysates (5 μg) were incubated with 200 μl of the reaction mixture (200 μM 5,5′-dithio-bis-(2-nitrobenzoic acid) plus either 200 μM CBZ-Gly-Arg-thiobenzylester for the granzyme A substrate or Boc-Ala-Ala-Asp-thiobenzylester for the granzyme B substrate (Enzyme Systems Products, Dublin, CA) in 10 mM HEPES-NaOH (pH 7.4), 150 mM NaCl, 1 mM CaCl₂, and 1 mM MgCl₂) at room temperature. Absorbance at 415 nm was measured.

OE4 cells (1 × 10⁶ cells/ml, 100 μl) were transferred into 96-well microtiter plates coated with anti-mouse CD3 Ab 145-2C11 (10 μg/ml). After centrifugation (300 × g, 3 min), the cells were incubated for 4 h. Culture supernatants were removed and measured for the activity of granzyme A as described above.

B lymphoma A20 cells are sensitive to Fas-mediated apoptosis and exhibited DNA fragmentation characteristic of apoptosis within 4 h upon exposure to cross-linked FasL, whereas Fas-negative A20.FO cells were totally resistant to cross-linked FasL (Fig. 1,B). ECH inhibited FasL-induced DNA fragmentation in a dose-dependent manner, and complete inhibition was observed when A20 cells were pretreated with 50–100 μM ECH for 1 h (Fig. 1,C). ECH was only diluted into half of the initial concentration during exposure to FasL, because DNA fragmentation was partially reversed when ECH was removed from A20 cells during the coculture with FasL (Fig. 1,D). Under these experimental conditions, ECH was not cytotoxic to the cell, as no DNA fragmentation was induced by ECH alone (Fig. 1,D). Twenty-four-hour incubation of A20 cells with ECH resulted in a marked reduction of live cellnumber without an induction of DNA fragmentation (Fig. 1, E and F). Less than 20% of the total cells were stained with trypan blue when A20 cells were treated with 50 μM ECH for 24 h (data not shown). In addition to A20 cells, the T lymphoma L5178Y cells were used for the second cell line to study the biological activity of ECH. Although L5178Y cells were insensitive to cross-linked FasL, Fas-transfected L5178Y cells (L5178Y-Fas) were highly susceptible to cross-linked FasL (Fig. 1,G). DNA fragmentation induced by cross-linked FasL was completely inhibited when L5178Y-Fas cells were pretreated with 100 μM ECH for 2 h, and ECH alone did not induce DNA fragmentation (Fig. 1 H).

KLH-specific H-2^d-restricted CD4⁺ CTL clone BK-1 cells are perforin-negative (33), and the killing pathway of this clone isexclusively dependent on FasL (29). BK-1 cells induced DNA fragmentation in KLH-pulsed A20.HL cells, but not A20.HL cells without Ag (Fig. 2,A). ECH markedly inhibited DNA fragmentation induced by BK-1 cells when KLH-pulsed A20.HL cells were pretreated with 50–100 μM for 1 h (Fig. 2).

To exclude the possibility that ECH pretreatment of target cells decreases CTL-target interaction, ECH-pretreated A20.HL cells were mixed with BK-1 cells, and resultant conjugate formation was analyzed by FACS (Fig. 3,A). A brief centrifugation facilitated the formation of CTL-target conjugates. ECH did not influence conjugate formation at 50 μM. It should be noted that this concentration resulted in a profound inhibition of DNA fragmentation induced by BK-1 cells (Fig. 2 A). As a slight reduction of CTL-target conjugates was observed when A20.HL cells were pretreated with 100 μM ECH, ECH might affect the interaction between CTL and target cells at higher concentrations.

The H-2^d-specific CD8⁺ CTL clone OE4 cells kill target cells via both the perforin-dependent pathway and the FasL-dependent pathway (18). As observed with BK-1 cells, ECH did not substantially prevent conjugate formation between OE4 and A20 cells up to 50 μM (Fig. 3,B). Perforin and granzymes stored in lytic granules are released upon TCR activation and induce apoptosis in target cells. The cellular levels of perforin and granzymes were thus analyzed in OE4 cells exposed to ECH. In contrast to CMA that induced a marked reduction of perforin, ECH failed to decrease the cellular amount of perforin in OE4 cells (Fig. 4,A). The enzyme activities of granzymes A and B in ECH-treated OE4 cells were also unimpaired (Fig. 4,B). Stimulation of OE4 cells with plate-coated anti-CD3 Ab induced exocytosis of lytic granules into the culture medium, and ECH only marginally reduced granule exocytosis even at 50 μM (Fig. 4 C). In our experimental system, effector CTL were exposed to ECH at half the initial concentrations for pretreatment of target cells. To avoid any direct effects on lytic granules and the granule exocytosis pathway during the killing assay, ECH was used at concentrations <100 μM for pretreatment of target cells.

To determine whether ECH inhibits perforin-dependent DNA fragmentation, two different Fas-negative lymphomas were used as target cells against OE4 cells. As shown in Fig. 1, A20.FO and L5178Y cells were completely resistant to soluble FasL. Therefore, DNA fragmentation of these target cells induced by OE4 cells is solely dependent on the perforin/granzyme system. ECH did not affect DNA fragmentation of A20.FO cells (Fig. 5, A and B) and L5178Y cells (Fig. 5,C) at concentrations up to 50 μM. Only slight reduction of DNA fragmentation was observed when A20.FO cells (Fig. 5,A) and L5178Y cells (Fig. 5, C and D) were pretreated with 100 μM ECH.

Fas-positive target cells are killed by OE4 cells via the perforin-dependent pathway and the FasL-dependent pathway. To block the perforin-dependent killing pathway, OE4 cells were pretreated with 100 nM CMA for 2 h. ECH alone did not affect DNA fragmentation of Fas-positive A20 cells induced by OE4 cells (Fig. 6, A and B). However, FasL-dependent DNA fragmentation induced by CMA-treated OE4 cells was prevented by ECH in a dose-dependent manner, and ECH completely inhibited DNA fragmentation when A20 cells were pretreated with 50 μM (Fig. 6, A and B). These results demonstrate that ECH inhibits FasL-dependent DNA fragmentation mediated by CD8⁺ CTL, but not perforin/granzyme-dependent DNA fragmentation. Likewise, in L5178Y-Fas cells, ECH dose-dependently inhibited DNA fragmentation induced by CMA-treated OE4 cells, and complete inhibition was observed when L5178Y-Fas cells were pretreated with 100 μM ECH (Fig. 6, C and D). The CMA-insensitive killing of OE4 cells corresponded to 40 and 70% in A20 cells (Fig. 6,B) and L5178Y-Fas cells (Fig. 6 D), respectively, suggesting that the FasL-dependent killing pathway plays a more dominant role in the induction of apoptosis in L5178Y-Fas cells than A20 cells. This difference might explain the observation that ECH alone exerts a stronger inhibitory activity toward DNA fragmentation of L5178Y-Fas cells.

To generalize the selective inhibitory effects of ECH on FasL-dependent DNA fragmentation, alloantigen-specific bulk CTL were induced by in vitro MLC for 4 days and used as effector CTL. In agreement with the observations using OE4 cells, ECH barely influenced DNA fragmentation of A20.FO cells induced by MLC cells (Fig. 7, A and B). In Fas-positive A20 cells, FasL-dependent DNA fragmentation induced by CMA-treated MLC cells was almost completely inhibited by ECH at 25 μM (Fig. 7,C). By contrast, ECH alone did not significantly influence DNA fragmentation induced by MLC cells under these conditions (Fig. 7, C and D).

The ⁵¹Cr release assay has been widely used for the measurement of target cell lysis in CTL-mediated cytotoxicity. As observed with the DNA fragmentation assay (Fig. 1,C), ECH inhibited cytolysis of A20 cells induced by soluble FasL in a dose-dependent manner (Fig. 8,A). Consistent with this observation, ECH strongly prevented FasL-dependent target cell lysis mediated by BK-1 cells (Fig. 8,B) as well as OE4 cells pretreated with CMA (Fig. 8,C). By contrast, perforin-dependent target cell lysis mediated by OE4 cells was only marginally affected by ECH (Fig. 8 D).

Recently, we have shown that ECH inhibits Fas-mediated apoptosis by blocking activation of procaspase-8 in the DISC (24). In this work we have investigated whether ECH inhibits the perforin-dependent killing pathway and the FasL-dependent killing pathway in CTL-mediated cytotoxicity. In the short term killing assay based on two different murine Fas-positive/negative target cells vs the CD4⁺ and CD8⁺ CTL clones as well as alloantigen-specific bulk MLC cells, ECH profoundly blocked the FasL-dependent DNA fragmentation and cytolysis of target cells, but barely prevented the perforin/granzyme-dependent DNA fragmentation and cytolysis of target cells. Moreover, ECH did not influence the cellular levels of perforin and granzymes A/B and only marginally reduced the granule exocytosis pathway in response to CD3 stimulation. Thus, our present results demonstrate that ECH is a highly selective inhibitor to block the FasL-dependent killing pathway in CTL-mediated cytotoxicity.

Death receptor-independent apoptosis induced by chemical compounds (i.e., staurosporine, MG-132, and ceramide) and UV irradiation was insensitive to ECH, whereas ECH markedly inhibited apoptosis induced by anti-Fas Ab, FasL, or TNF (24). These results suggest that ECH selectively blocks death receptor-mediated apoptosis that requires activation of procaspase-8. In the granule-dependent killing pathway, granzymes A and B have major roles in inducing target cell death upon translocation into the cytosol (1, 2). Granzyme A induces a caspase-independent cell death characterized by ssDNA nicks, which are mediated by granzyme A-activated DNase, NM23-H1 (34), whereas granzyme B initiates procaspase-3 processing, and the release of proapoptotic mitochondrial factors that facilitate the full activation of procaspase-3 (8, 9, 10). In the short term killing of target cells, granzyme B is critically involved in a rapid induction of DNA fragmentation (13). Consistent with this idea, the synthetic caspase inhibitor benzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethyl ketone completely prevented DNA fragmentation of A20 cells induced by the CD8⁺ CTL clone in the short term assay (data not shown), confirming that granzyme B is a main factor that induces DNA fragmentation in the perforin-dependent killing pathway. ECH failed to prevent the perforin/granzyme B-dependent DNA fragmentation. These findings provide additional evidence that ECH specifically inhibits death receptor-mediated apoptosis, but does not affect death receptor-independent apoptosis.

The major molecular target of ECH in Fas-mediated apoptosis is procaspase-8, and ECH affected neither active caspase-8 nor activation of procaspase-3 and procaspase-9 at the cellular level (24). ECH has a molecular structure of α,β-unsaturated ketone and epoxide that might be reactive to thiol residues of proteins. In agreement with this hypothesis, glutathione or cysteine neutralized ECH binding to procaspase-8 (24). Recently, we have reported that the mycotoxin penicillic acid (PCA) inhibits Fas-mediated apoptosis by blocking activation of procaspase-8 (23). Although ECH and PCA are structurally unrelated, PCA exhibited the inhibitory activity similarly to ECH, in that PCA preferentially inhibited activation of procaspase-8, but did not affect active caspase-8 in living cells (23). Moreover, activation of procaspase-3 and procaspase-9 in the cell was only weakly inhibited by PCA (23). PCA contains α,β-unsaturated lactone able to bind to sulfhydryl groups. Analysis by mass spectrometry revealed that PCA binds to the cysteine residue of the active center in procaspase-8 (23). Although we have not yet determined the binding sites of ECH on procaspase-8, it seems likely that ECH binds to the active center cysteine and inactivates its intrinsic proteolytic activity.

DNA fragmentation was partially reversed when ECH was removed from A20 cells during the 4-h coculture with FasL. In our earlier paper (24), we showed that ECH affects neither cell surface Fas expression nor Fas-FasL interaction. Thus, it seems likely that a portion of procaspase-8 becomes functional by dissociation from ECH or is newly synthesized during the 4-h incubation, although we cannot rule out the possibility that ECH affects FasL expression on CTL.

In the short term culture, ECH alone did not induce apoptosis or necrosis. However, in the long term culture, ECH markedly reduced live cells without an induction of DNA fragmentation, whereas a portion of ECH-treated cells became positive for trypan blue staining. Thus, these results indicate that ECH inhibits proliferation, but can also induce necrotic cell death in the long term culture. Molecular mechanisms of ECH on inhibiting proliferation or inducing necrosis remain to be elucidated.

ECH inhibited Fas-mediated apoptosis in several human cell lines (24) (data not shown) and two murine lymphoma cell lines tested in this paper at the minimum inhibitory concentrations of ECH ranging from 10–100 μM. ECH inhibited Fas-mediated apoptosis at relatively lower concentrations in human cell lines (24), and 100 μM ECH was required for the complete inhibition of the Fas-mediated apoptosis in L5178Y-Fas cells. The culture medium for murine lymphoma cells contains 50 μM 2-ME, which might neutralize ECH due to the presence of thiol residues. However, the inhibitory doses of ECH unaltered in culture medium without 2-ME (data not shown). In addition, the FasL susceptibility is unlikely to involve the inhibitory doses of ECH, because L5178Y-Fas cells and A20 cells were equivalently susceptible to FasL, but Fas-mediated apoptosis in A20 cells was more strongly inhibited by ECH. A possible explanation might be that intracellular competitors such as glutathione and/or thiol-containing protein(s) antagonize ECH and determine the strength of its inhibitory activity on procaspase-8.

We previously showed that CMA is a specific inhibitor of the perforin-dependent killing pathway (18). CMA induces inactivation and subsequent proteolytic degradation of perforin in lytic granules upon neutralization of acidic pH (17, 19, 20). By contrast, the FasL-dependent killing pathway mediated by Ag-specific CTL was suppressed by an inhibitor of glycoprotein transport, brefeldin A (18), as FasL is newly synthesized upon TCR activation and transported to the cell surface. Ca²⁺ dependency was frequently used to distinguish between the perforin-based cytotoxicity and the FasL-based cytotoxicity. However, Ca²⁺ is not only required for the function of perforin, but also for TCR-mediated early signaling leading to FasL expression (35, 36, 37). Therefore, the Ca²⁺-independent cytotoxicity previously observed is largely due to the pre-existing FasL and is believed to be insensitive to brefeldin A. As ECH blocks activation of procaspase-8 in target cells, it is thought that ECH is a more general nonpeptide inhibitor of the FasL-mediated killing pathway exerted by CTL, NK cells, or other types of cells, even though they constitutively express cell surface FasL.

The Fas/FasL system plays an essential role in lymphocyte homeostasis, and mutations in this system lead to the accumulation of abnormal T cells in the peripheral lymphoid organs and the development of severe autoimmune diseases in humans and mice (3, 4, 5). In addition to death-inducing functions, our and other groups reported that Fas provides costimulatory signals for human T cells, and caspase inhibitors block T cell proliferation (38, 39, 40), suggesting that caspase activation is required for T cell proliferation. The FasL-dependent killing of target cells by CTL is involved in the pathogenesis of experimental autoimmune encephalomyelitis and fulminant hepatitis (41, 42, 43). In neuronal diseases such as Huntington diseases, procaspase-8 is activated independently of death receptors (44, 45). Taken together, caspase-8 inhibitors such as ECH might be therapeutic candidates to treat autoimmune diseases and neuronal diseases.

Mice deficient in perforin, granzymes, Fas, or FasL have been used to study the biological significance of the perforin/granzyme system and the Fas/FasL system in various in vivo models and in vitro killing assays. These knockout mice, however, are not always applicable to all experimental settings. In this report we have shown that ECH is a specific inhibitor of the FasL-dependent killing pathway, but does not affect the perforin/granzyme-dependent killing pathway. Thus, ECH is a highly useful tool to evaluate the FasL-dependent killing pathway in cell-mediated cytotoxicity and might be applicable for all CTL-target combinations.

We thank Dr. N. Shinohara for CTL clones and their target cells, and Drs. J. Tschopp and H. Yagita for reagents. We also thank Dr. R. C. Budd for critical reading of the manuscript.