Comparison of Apoptosis in Wild-Type and Fas-Resistant Cells: Chemotherapy-Induced Apoptosis Is Not Dependent on Fas/Fas Ligand Interactions

RECENT STUDIES from a number of laboratories indicate that widely used chemotherapeutic agents induce apoptosis in susceptible cells.1-3 For example, treatment with the topoisomerase II poison etoposide results in morphological and biochemical evidence of apoptosis in a variety of cell types.4-7 The events occurring between stabilization of topoisomerase II-DNA complexes and initiation of apoptosis are unclear, but subsequent events in all of these cell types include cleavage of DNA^4,6,7 and activation of cysteine-dependent aspartate-directed proteases8-11 called caspases.12 Although some mammalian cells have been shown to encode mRNA for at least 10 caspases, recent studies indicate that etoposide treatment results in selective activation of certain family members,11 including caspase-3, which cleaves PARP and DNA-protein kinase at DEVD-X bonds,13,14 and caspase-6, which recognizes the sequence VEID-N and cleaves the nuclear lamins.15,16 In contrast, other caspases, including caspases-1 and -2, do not appear to be activated during etoposide-induced apoptosis in human leukemia cells.11 

In a similar fashion, multiple caspases are activated during apoptosis triggered by Fas (CD95/APO-1) ligation.17-19 Fas-mediated apoptosis has been shown to be involved in the development of immune tolerance, regulation of immune responses, and killing of virally infected cells and tumor targets.18,20,21 Current evidence suggests that ligation of Fas leads to the binding of an adaptor protein FADD/MORT1, which in turn binds caspase-8 via its FADD death effector domain(s).17,19 In addition to caspase-8, caspase-10 has also been shown to contain FADD death effector domains, suggesting that the two proteases might associate with FADD.22 It is postulated that, upon binding to FADD, caspase-8 and possibly caspase-10 are activated by currently unidentified proteases and then become available for initiating a cascade of caspases.17,19 

The exact relationship between the pathways of chemotherapy-induced apoptosis and Fas/FasL-triggered apoptosis has been unclear. A number of studies using various inhibitors have raised the possibility that different apoptosis-inducing stimuli might activate pathways involving different proteases.23-26 In particular, the observation that Bcl-2 inhibits etoposide-induced apoptosis27,28 but not Fas-induced apoptosis16 is consistent with the claim that Fas- and chemotherapy-induced apoptosis proceed through distinct pathways.29 On the other hand, it has also been observed that Fas ligation and chemotherapeutic agents activate common downstream components of the apoptotic machinery. For example, treatment of Fas-expressing cells with anti-Fas or staurosporine results in proteolytic cleavage of procaspase-3, procaspase-6, and PARP.16,29 Likewise, sterol-regulatory element binding protein-2, a substrate for caspases-3 and -7, is cleaved after anti-Fas, staurosporine, or etoposide treatment.30 These observations suggest considerable overlap between the apoptotic pathways activated by chemotherapeutic agents and Fas ligation.

Friesen et al31 recently reported that treatment of CEM human T-cell leukemia cells with the anthracycline doxorubicin resulted in the induction of FasL expression followed by Fas/FasL-dependent apoptosis. Significantly, a Fas-resistant CEM subline was insensitive to apoptosis induced by doxorubicin. These results not only suggested that certain chemotherapeutic agents might initiate apoptosis by activating the Fas-dependent cell death pathway, but also led the investigators to propose that broad spectrum resistance to chemotherapeutic agents might be related to defects in the Fas pathway.31 The recent demonstration that FasL can be expressed in melanoma, hepatoma, and lung cancer cells32-34 raises the possibility that this molecule can be regulated in nonlymphoid cells and lends plausibility to the hypothesis that antineoplastic agents might induce cell death through the Fas/FasL pathway. Accordingly, the proposal that defects in the Fas pathway contribute to an important and previously unrecognized mechanism of chemotherapy resistance has wide-ranging implications for future studies of drug resistance in lymphoid and nonlymphoid malignancies.

In the present study, flow cytometry and a novel affinity-labeling procedure that detects active caspases were used to compare apoptotic pathways activated by anti-Fas and etoposide in human Jurkat T-leukemia cells. Multiple independently derived Fas-resistant Jurkat cell lines were studied to examine the possibility that etoposide, doxorubicin, topotecan, cisplatin, staurosporine, methotrexate, and γ-irradiation trigger cell death through the Fas/FasL pathway. Results of these experiments indicate that (1) the downstream effectors of the programmed cell death process induced by etoposide and anti-Fas are indistinguishable, and (2) the triggering of apoptosis by a broad array of chemotherapeutic agents occurs in an Fas/FasL-independent fashion.

Materials.Purchased reagents included etoposide, ICR 191, ethyl mercurithiosalicylate (EMS), doxorubicin, cisplatin, methotrexate, staurosporine, RNase A, 7-amino-4-trifluoromethylcoumarin (AFC), and propidium iodide from Sigma (St Louis, MO); saponin from CalBiochem (La Jolla, CA); DEVD-AFC and YVAD-AFC from Enzyme Systems Products (Dublin, CA); fluorescein-conjugated anti-Fas monoclonal antibody (MoAb) UB2 and ZB4 IgG₁ blocking anti-Fas MoAb from Immunotech (Westbrook, ME); rabbit anticaspase-1 from Oncogene Research (Cambridge, MA); and monoclonal anti-caspase-2 and anti-caspase-3 from Transduction Labs (Lexington, KY). Topotecan was generously provided by SmithKline Beecham Pharmaceuticals (King of Prussia, PA). The monoclonal anti-Fas antibody 7C11 (IgM) and polyclonal anti-caspase-7 were kind gifts from Michael J. Robertson (Dana-Farber Cancer Institute, Boston, MA) and Vishva Dixit (University of Michigan Medical School, Ann Arbor), respectively. All other reagents were obtained or synthesized as previous described.4,11 

Tissue culture.Jurkat human T-cell leukemia cells obtained from the American Type Culture Collection (Rockville, MD) were cultured in RPMI 1640 medium containing 10% bovine serum, 100 μg/mL gentamicin, and 2 mmol/L glutamine. The Fas-resistant Jurkat clones JM3A5, JM14A5, JM18A1, and JM19A6 were generated by incubating wild-type Jurkat cells with 5 μg/mL ICR 191, a frameshift mutagen, for 5 hours on one to five separate occasions. The Fas-resistant clone JM5A5 was generated by incubating wild-type Jurkat cells with 400 μg/mL EMS for 24 hours. Each of the Fas-resistant cell lines was independently derived from a separate wild-type Jurkat cell pool. Anti-Fas (7C11; 0.5 μg/mL) was added to the mutagenized cells 1 week after the last mutagenesis, followed by cloning by limiting dilution in the presence of anti-Fas (7C11; 0.5 μg/mL). The surface expression of Fas was measured using a fluorescein-conjugated anti-Fas MoAb, UB2, followed by flow cytometry as previously described.35 For induction of apoptosis, cells were treated with the indicated concentrations of etoposide, doxorubicin, topotecan, cisplatin, methotrexate, or staurosporine (added from 1000X stocks in dimethyl sulfoxide [DMSO]) as indicated in the individual figure legends.

Enzyme assays.After treatment, cytosol was prepared at 4°C and assayed for activity as previously described.11 In brief, cells were sedimented at 200g for 10 minutes, washed twice in serum-free RPMI 1640, and incubated for 20 minutes in buffer A (25 mmol/L HEPES [pH 7.5 at 4°C], 5 mmol/L MgCl₂ , 1 mmol/L EGTA supplemented immediately before use with 1 mmol/L α-phenylmethylsulfonyl fluoride [PMSF], 10 μg/mL pepstatin A, and 10 μg/mL leupeptin) and lysed with 20 to 30 strokes in a tight-fitting Dounce homogenizer. After removal of nuclei by sedimentation at 16,000 g for 3 minutes, the supernatant was supplemented with 0.5 mmol/L EDTA and sedimented at 280,000g_max for 60 minutes in a Beckman TL-100 ultracentrifuge (Palo Alto, CA). After addition of dithiothreitol (DTT) to a final concentration of 2 mmol/L, the supernatant (cytosol) was frozen in 50-μL aliquots at −70°C. Experiments showed that DEVD-AFC cleavage activity was stable for at least 3 months. All experiments described in the present study were performed within 2 weeks of extract preparation.

To assay for caspases, aliquots containing 50 μg of cytosolic protein (estimated by the bicinchoninic acid method36 ) in 50 μL buffer A were diluted with 225 μL of freshly prepared buffer B (25 mmol/L HEPES [pH 7.5], 0.1% [wt/vol] CHAPS, 10 mmol/L DTT, 100 U/mL aprotinin, 1 mmol/L PMSF) containing 100 μmol/L substrate and incubated for 2 hours (DEVD-AFC, VEID-AFC) or 4 hours (YVAD-AFC) at 37°C. Reactions were terminated by addition of 1.225 mL ice-cold buffer B. Fluorescence was measured in a Sequoia-Turner fluorometer (Mountain View, CA) using an excitation wavelength of 360 nm and emission wavelength of 475 nm. Reagent blanks containing 50 μL of buffer A and 225 μL of buffer B were incubated at 37°C for 2 hours, then diluted with 1.225 mL ice-cold buffer B. Standards containing 0-1500 pmol of AFC were used to determine the amount of fluorochrome released.

Affinity labeling.The affinity label Z-EK(bio)D-amok was synthesized as previously described.11 Aliquots containing 50 μg of HL-60 cytosolic protein or 75 μg of Jurkat cytosolic protein were incubated for 1 hour at 20-22°C with 1 μmol/L Z-EK(bio)D-aomk (added from a 25 μmol/L stock in DMSO). At the completion of the incubation, extracts were diluted with 1/2 volume of 3X concentrated SDS sample buffer, heated to 95°C for 3 minutes, subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 16% (wt/vol) acrylamide gels, transferred to nitrocellulose, probed with peroxidase-labeled streptavidin, and visualized by enhanced chemiluminescence. Control experiments revealed that P20 subunits of all active recombinant caspases tested, including caspases-1, -2, -3, -4, and -6, reacted with Z-EK(bio)D-amok.11 Additional experiments indicated that as little as 1 ng of purified caspase-1 could be detected.

Immunoblotting.After incubation with apoptosis-inducing stimuli for the indicated length of time, cells were sedimented at 200 g for 10 minutes at 4°C, washed once with ice-cold RPMI containing 10 mmol/L HEPES (pH 7.4 at 4°C), and solubilized in lysis buffer consisting of 6 mol/L guanidine hydrochloride, 250 mmol/L Tris-HCl (pH 8.5 at 21°C), 10 mmol/L EDTA, 150 mmol/L β-mercaptoethanol, and 1 mmol/L freshly added PMSF. After sonication, treatment with iodoacetamide to block free sulfhydryl groups, dialysis into 0.1% (wt/vol) SDS, and lyophilization as described,37 samples containing 50 μg protein in SDS sample buffer consisting of 4 mol/L deionized urea, 2% (wt/vol) SDS, 62.5 mmol/L Tris-HCl (pH 6.8 at 21°C), and 1 mmol/L EDTA were heated to 65°C for 20 minutes; subjected to SDS-PAGE on gels with 5% to 15% (wt/vol) acrylamide gradients; and transferred to nitrocellulose. Blots were probed as previously described4,11 using an MoAb that recognizes PARP, chicken polyclonal antiserum raised against rat lamin B₁ , and antibodies to caspases -2, -3, and -7 described above.

Flow cytometry.Propidium iodide staining was performed as previously described.38 Briefly, after the indicated treatments, cells were washed with phosphate-buffered saline (PBS); resuspended in 20 mmol/L HEPES (pH 7.4) containing 120 mmol/L NaCl, 60 μg/mL saponin, 50 μg/mL RNase A, and 20 μg/mL propidium iodide; incubated at 37°C for 1 hour; and analyzed (20,000 events/sample) on a Becton Dickinson FACScan (Mountain View, CA). Hypodiploid cells were quantitated in each sample.

Reverse transcription-polymerase chain reaction (RT-PCR).RT-PCR for procaspase transcripts was performed as described previously11 except that primers for procaspase-7 (Accession no. U37449) were (forward primer) TCAAGTGCTTCCGAAGCCTGG and (reverse primer) TGACCCATTGCTTCTCAGCTAGAATGTAC. Control PCR reactions included (1) HL-60 cDNA template and human β-actin primers (forward-GTGGGGCGCCCCAGGCACCA, reverse-CTCCTTATGTCACGCACGATTTC) to validate the cDNA, and (2) 0.05 μg HL-60 poly(A⁺) RNA template to confirm that the products observed were derived from cDNA rather than contaminating genomic DNA. After amplification, 10% of each PCR reaction was applied to a 3.5% composite agarose gel (2.5% NuSieve, 1% Seakem-GTG; FMC Bioproducts, Rockland, ME) containing 0.5 μg/mL ethidium bromide in 1x TBE buffer. After electrophoresis for 160 volt-hours, reaction products were visualized on a UV transilluminator (256 nm) and photographed.

Comparison of anti-Fas– and etoposide–induced apoptosis in wild-type Jurkat T-cell leukemia cells.To investigate the possible relationship between chemotherapy- and Fas-induced apoptosis, we initially compared the effects of the topoisomerase II–directed agent etoposide and the anti-Fas MoAb 7C11 on wild-type Jurkat cells. Treatment with 100 ng/mL anti-Fas resulted in the appearance of cells with subdiploid DNA content within 3 hours (Fig 1A). Similarly, incubation with 68 μmol/L etoposide induced DNA fragmentation within 6 hours (Fig 1A). Examination of anti-Fas– and etoposide–treated cells by electron microscopy confirmed that these cells displayed the characteristic morphological features of apoptosis, including cytoplasmic blebbing, chromatin condensation, and nuclear fragmentation (data not shown).

RT-PCR showed that untreated Jurkat cells constitutively expressed multiple different caspases (Fig 1B). The identity of the observed PCR products was confirmed by sequencing. This sequence analysis also revealed that the multiple products observed from primers for caspases-8 and -10 were derived from alternatively spliced transcripts as described in Boldin et al.39 In further experiments, Northern blotting showed that full-length message was detectable for procaspases-2, -3, -6, -7, -8, and -10. In contrast, levels of mRNA for procaspases-1, -4, and -5 were below the limit detected by Northern blotting (data not shown).

Immunoblotting showed that PARP and lamin B₁ , two caspase substrates,4,8,40 were cleaved upon treatment with anti-Fas or etoposide (Fig 1C). Anti-Fas treatment led to the cleavage of PARP within 1.5 hours and lamin B₁ within 2 hours, whereas etoposide treatment resulted in cleavage of PARP and lamin B₁ within 3 hours. In addition, levels of procaspases-3 and -7 decreased within 2 hours after the addition of anti-Fas and 4 hours after the addition of etoposide (Fig 1C). In contrast, no cleavage of procaspase-2 was seen 3 hours after anti-Fas or 5 hours after etoposide was added to the wild-type Jurkat cells (Fig 1C). These observations suggested that etoposide and anti-Fas treatments resulted in activation of similar apoptotic proteases.

Enzyme assays using the fluorogenic substrates DEVD-AFC, VEID-AFC, and YVAD-AFC likewise suggested that Fas- and etoposide-stimulated apoptotic pathways were similar. The loss of procaspase-3 from immunoblots (Fig 1C) was reflected in a >100-fold increase in cytosolic DEVD-AFC cleavage activity and a >20-fold increase in cytosolic VEID-AFC cleavage activity (Fig 1D) upon treatment with anti-Fas or etoposide. With both stimuli, kinetics of appearance and magnitude of maximal activity were similar when activity was assayed in whole cell extracts rather than cytosol (data not shown). No YVAD-AFC cleavage activity was detected throughout the entire period of incubation with either anti-Fas or etoposide (Fig 1D) even though the same assay readily detected ICE activity in cytosol from THP.1 monocytic leukemia cells.11 

Affinity labeling with Z-EK(bio)D-amok, which selectively modifies the large subunits of active caspases,11 provided further support for the similarity of the Fas- and etoposide-induced apoptotic pathways. Examination of Z-EK(bio)D-aomk-labeled cytosolic (Fig 1E) or whole cell extracts (data not shown) by unidimensional SDS-PAGE followed by blotting with peroxidase-labeled biotin showed that five bands were specifically labeled beginning 1 hour after anti-Fas treatment. The same five bands were detected 3 hours after etoposide was added to the cells (Fig 1E). Further experiments involving recombinant caspases and two-dimensional electrophoresis showed that the major labeled polypeptides migrating at the positions of IRP₁ and IRP₃ correspond to two different molecular-weight forms of active caspase-6, whereas the major polypeptides migrating at the positions of IRP₂ and IRP₄ correspond to two different molecular-weight forms of caspase-3.11 The identities of the additional minor bands (not labeled in Fig 1E) remain to be determined.

Aside from the kinetics, etoposide-induced apoptosis in Jurkat cells was indistinguishable from Fas-induced apoptosis by all criteria examined.

Anti-Fas blocking antibody does not inhibit chemotherapy-induced apoptosis.To examine the potential role of Fas/FasL interactions in initiating etoposide-induced apoptosis, Jurkat cells were preincubated with ZB4, a monoclonal anti-Fas antibody that blocks Fas-mediated killing.41 Results of these experiments are shown in Fig 2. Control experiments confirmed that ZB4 inhibited apoptosis induced by ligation of the T-cell receptor with anti-CD3, a treatment that initiates cell death by upregulating FasL expression.38,42 ZB4 also inhibited apoptosis induced by the anti-Fas antibody 7C11. In contrast, ZB4 had no effect on apoptosis induced by various concentrations of etoposide. Likewise, induction of apoptosis by the anthracycline doxorubicin was unaffected by ZB4. These results suggested that etoposide and doxorubicin do not require Fas/FasL interactions to produce apoptotic cell death.

Chemotherapy-induced apoptosis in Fas-resistant JM3A5 cells.To further explore the relationship between the Fas/FasL pathway and chemotherapy-induced apoptosis, multiple Fas-resistant Jurkat cell lines were generated by mutagenesis followed by selection in medium containing 7C11 monoclonal anti-Fas antibody. Results in one cell line (designated JM3A5) are presented in detail in Fig 3. In contrast to parental Jurkat cells, JM3A5 cells displayed no evidence of DNA fragmentation after treatment with 100 ng/mL anti-Fas antibody (Fig 3A). Likewise, there was no cleavage of PARP, lamin B₁ , procaspase-3, or procaspase-7 after anti-Fas treatment of JM3A5 cells (Fig 3B). Consistent with these results, anti-Fas treatment did not result in activation of caspases capable of cleaving the fluorogenic substrate DEVD-AFC (Fig 3C). Electron microscopy (not shown) revealed that anti-Fas-treated JM3A5 cells were indistinguishable from untreated cells, confirming that JM3A5 cells did not undergo apoptosis after Fas ligation.

Treatment of the same JM3A5 cell line with etoposide produced all of the hallmark changes of apoptosis (Fig 3). In particular, DNA fragmentation was first evident 3 hours after the addition of etoposide and increased thereafter as assessed by flow cytometry (Fig 3A). Likewise, apoptotic proteases in the JM3A5 cells were also activated by etoposide treatment as indicated by cleavage of PARP, lamin B₁ , procaspase-3, and procaspase-7 (Fig 3B) and the appearance of DEVD-AFC cleavage activity in cytosol (Fig 3C). Electron microscopy confirmed that etoposide treatment was accompanied by chromatin condensation and nuclear fragmentation in the JM3A5 cells (data not shown). All of these changes occurred in a time course that could not be distinguished from etoposide-treated parental Jurkat cells (Fig 1). These observations provide additional evidence that etoposide-induced apoptosis does not depend on a functional Fas/FasL pathway.

These results were not limited to etoposide-induced apoptosis. When JM3A5 cells were treated with doxorubicin (0.5 μmol/L, 24 hours), topotecan (0.1 μmol/L, 24 hours), cisplatin (1 μmol/L, 48 hours), methotrexate (100 nmol/L, 24 hours), staurosporine (1 μmol/L, 6 hours), or γ-irradiation (1,200 cGy followed by a 24-hour incubation), apoptosis was comparable to that observed in parental Jurkat cells (Fig 4A). Dose-response curves performed with etoposide and doxorubicin indicated that JM3A5 cells were at least as sensitive as parental Jurkat cells to the induction of apoptosis by these treatments (Fig 4B and C). These results indicated that Fas-resistant JM3A5 cells remain sensitive to a wide variety of chemotherapeutic agents.

Multiple Fas-resistant lines remain sensitive to chemotherapy-induced apoptosis.These results were not unique to the JM3A5 cell line. Results obtained with four additional Fas-resistant Jurkat lines (JM5A5, JM14A5, JM18A1, and JM19A6) are summarized in Fig 5. JM3A5, JM5A5, JM18A1, and JM19A6 express wild-type levels of Fas, whereas Fas is undetectable on the surface of JM14A5 cells (data not shown). All of these cell lines were resistant to anti-Fas–induced apoptosis (Fig 5). Like the JM3A5 cells, however, the JM5A5, JM14A5, and JM18A1 Fas-resistant lines displayed DNA fragmentation after all of the other treatments used in this study (Fig 5 and data not shown). These results confirm and extend the observations made with the JM3A5 cell line.

Interestingly, one Fas-resistant cell line, JM19A6, was less sensitive to several chemotherapeutic agents (Fig 5). This line nonetheless retained its ability to undergo apoptosis after treatment with topotecan, indicating that the resistance observed in this cell line was drug-specific rather than a common feature of all chemotherapeutic agents.

In the present report we have shown that (1) chemotherapeutic agents and Fas ligation trigger apoptotic pathways that converge on the same effector proteases17 and (2) induction of apoptosis by chemotherapeutic agents is not dependent on Fas/FasL interactions. These results have important implications for current understanding of apoptotic pathways and of drug resistance.

The conclusion that the downstream pathways activated by chemotherapeutic agents and Fas ligation converge is based on multiple types of experiments. Immunoblotting experiments (Fig 1C) demonstrated that etoposide and anti-Fas both lead to selective cleavage of certain procaspases (eg, procaspases-3 and -7) but not others (eg, procaspase-2), a result that agrees with recent reports that procaspases-3, -6, and/or -7 are cleaved after treatment with anti-Fas or various chemotherapeutic agents.9,16,23,29,30 In experiments that extend these studies, DEVD-AFC and VEID-AFC cleavage assays showed that multiple distinct caspase activities11 become detectable after treatment with either anti-Fas or etoposide (Fig 1D). Moreover, affinity labeling of caspase active sites showed that etoposide and anti-Fas result in activation of the same spectrum of downstream caspases (Fig 1E). Collectively, all of these results suggest that the late stages of apoptosis after treatment with anti-Fas or etoposide proceed through a common caspase pathway in Jurkat cells.

Although anti-Fas and etoposide activate common downstream effectors, two lines of experimentation provide evidence that chemotherapeutic agents do not depend on Fas/FasL interactions to trigger apoptosis. In an initial series of experiments, parental Jurkat cells were treated with etoposide or doxorubicin in the presence of ZB4 MoAb. This antibody inhibited apoptosis induced by anti-CD3, which upregulates FasL, as well as anti-Fas–induced apoptosis (Fig 2). Nonetheless, this antibody did not inhibit etoposide- or doxorubicin-induced apoptosis (Fig 2), suggesting that etoposide and doxorubicin trigger apoptosis in a Fas/FasL-independent manner. In a second series of experiments, five independently derived Fas-resistant Jurkat lines were examined for their ability to undergo apoptosis after treatment with multiple chemotherapeutic agents. Four of the five lines characterized were as sensitive to doxorubicin-induced apoptosis as the parental Jurkat cells from which they were derived (Figs 4 and 5). These four doxorubicin-sensitive lines included at least one line, JM14A5, in which resistance to anti-Fas was associated with a lack of Fas expression, as well as three lines in which normal levels of immunoreactive Fas were expressed. The same four Fas-resistant lines also underwent apoptosis after treatment with etoposide, topotecan, cisplatin, methotrexate, staurosporine, or γ-irradiation (Fig 5 and data not shown). These results again indicate that a functional Fas/FasL pathway is not required for chemotherapy-induced apoptosis.

One of the five Fas-resistant cell lines, JM19A6, was cross-resistant to the induction of apoptosis by etoposide and cisplatin (Fig 5). This cell line retained the ability to undergo apoptosis after treatment with topotecan, again indicating that an intact Fas/FasL pathway is not required for drug-induced apoptosis. The phenotype displayed by this cell line suggests that one or more mutations affecting drug-specific pathways might have been acquired during the process of mutagenesis and selection.

Our observation that chemotherapy-induced apoptosis can occur in the absence of Fas/FasL interactions differs from that of Friesen et al.31 This disparity might reflect differences between cell lines used in the two studies. However, additional experiments performed in our laboratories have indicated that the Fas/FasL-independent induction of apoptosis by chemotherapeutic agents is not limited to Jurkat cells. For example, we have observed that HL-60 cells, which undergo apoptosis after exposure to a wide range of chemotherapeutic agents,4 fail to express Fas and fail to undergo apoptosis after treatment with anti-Fas antibodies (C.M.E. and T.J.K., unpublished observations, July 1996).

Although the present observations indicate that chemotherapeutic agents initiate apoptosis through a pathway separate from the Fas receptor, at some point downstream of Fas both pathways apparently converge. Because the molecules essential for either Fas-induced or etoposide-induced apoptosis have not been fully elucidated, it is unknown how far upstream from caspases-3, -6, and -7 or how far downstream from Fas the two pathways converge. The Fas-resistant Jurkat mutants described in the present work should provide unique reagents that will be helpful in determining which transducers of the apoptotic signal are unique to the Fas/FasL pathway, which are unique to the chemotherapy-activated pathway, and which are features of the common downstream pathway.

We thank Ivan Lieberburg for making this collaboration possible, Daniel J. McCormick for synthesizing VEID-AFC, Robert T. Abraham for helpful conversations, Keith Bible for advice regarding electron microscopy, and Deb Strauss for secretarial assistance.