Drug-induced Apoptosis in Lung Cancer Cells Is Not Mediated by the Fas/FasL (CD95/APO1) Signaling Pathway¹

Lung cancer is the leading cause of cancer death in industrialized countries (1), and two major types can be identified:NSCLC⁴ and SCLC. NSCLC, the most prevalent subtype, is relatively resistant to chemotherapy and radiation. In contrast, SCLC is initially highly sensitive to chemotherapy and radiotherapy, although resistant relapses occur in the majority of patients. Defining the molecular determinants of sensitivity or resistance to chemotherapy in lung cancer would have important implications for treatment.

Cytotoxic drugs, irrespective of their intracellular target, have been shown to cause cell death in sensitive cells, at least partly, by inducing apoptosis (2). Nevertheless, the precise molecular requirements that trigger the apoptotic pathway remain largely unknown (3) and have not been extensively studied in lung cancer cells.

Fas (CD95/APO1) is a M_r 45,000 type I transmembrane glycoprotein that belongs to the nerve growth factor/tumor necrosis factor receptor superfamily (4). Cross-linking of Fas by either the natural ligand or an agonistic antibody transduces a signal that results in rapid induction of apoptosis in susceptible cells (5). FasL (CD95-L) is a M_r 40,000 type II transmembrane protein that is a member of the tumor necrosis family of cytokines (6, 7). FasL also exists in a soluble form, released from the cell surface after cleavage by metalloproteinases (8, 9). After the cross-linking by the ligand, the cytoplasmic region of the Fas receptor, called death domain (10),recruits proteins present in the cytoplasm, including MORT1/FADD (11, 12). Subsequently, caspase-8, or MACH/FLICE (13, 14), is activated, and this cysteine protease plays an essential role in the proteolytic cascade that finally leads to apoptosis mediated by Fas and other death receptors (15). A role for the Fas system as a mediator of the drug-induced apoptosis has been proposed (16, 17, 18, 19, 20, 21, 22, 23). Chemotherapy would induce an up-regulation of FasL (19, 23), leading to an autocrine/paracrine activation of Fas signaling, and this may constitute a potential mechanism in the mediation of anticancer drug-induced apoptosis. Alternatively, other studies have shown that drug-induced apoptosis occurs independently from the Fas signaling (24, 25, 26, 27, 28, 29). As a consequence, the exact role of the Fas system in apoptosis induced by chemotherapy remains unsettled.

We previously described cellular changes indicative of apoptosis in lung cancer cell lines treated with topotecan and gemcitabine, two active drugs in lung cancer (30). The present work was designed to study the Fas system in the context of its expression,functional status, and possible relation with chemotherapy-induced apoptosis in lung cancer cell lines. Our results show that the Fas/FasL pathway is not involved in chemotherapy-induced apoptosis in lung cancer cells but indicate that the occurrence of caspase-8 activation induced by chemotherapy does not require Fas/FasL signaling.

Drugs were provided as pure substances. Topotecan (Smithkline-Beecham Pharmaceuticals, Herts, United Kingdom) was diluted in water. Gemcitabine (Eli Lilly Research Laboratories, Indianapolis, IN) and cisplatin (Bristol-Myers Squibb, Woerden, the Netherlands) were diluted in PBS, and paclitaxel (Bristol-Myers Squibb, Syracuse, NY) was diluted in ethanol. The drugs were freshly diluted to the final concentration in culture medium before each experiment.

Human NSCLC (NCI-H460 and NCI-H322) and SCLC (NCI-H187 and NCI-N417)cell lines were kindly provided by Dr. A. Gadzar. Both GLC4 and the mrp-overexpressing GLC4/ADR (generated by continuous culture in the presence of doxorubicin) were obtained from Dr. E. De Vries. The Jurkat human T-cell leukemia cell line was obtained from Dr. T. van Lopik. Cells were cultured in RPMI 1640 supplemented with 10%heat-inactivated FCS (Life Technologies, Inc., Breda, the Netherlands),2 mm l-glutamine, 50 IU/ml penicillin, and 50μg/ml streptomycin and incubated at 37°C in a humidified atmosphere with 5% CO₂. The cell lines were tested regularly for the presence of Mycoplasma infection and found to be negative. Cells from exponentially growing cultures were used for all of the experiments.

Cells were stained with the anti-Fas (APO-1) mAb CLB-CD95/15, raised in our group (31), or anti-FasL NOK-1 mAb (PharMingen, San Diego, CA) for 45 min at 4°C. To control for aspecific binding, an IgG2a antibody (Dako, Santa Barbara, CA) for FasR or an IgG1 for the ligand (Dako) were used as isotype-matched, nonbinding antibodies. The equivalent to 1 μg of protein of anti-Fas and FasL mAb antibodies or the respective isotype antibody was used in each sample. Cells were then washed twice with cold PBS and incubated with FITC-conjugated goat-antimouse antibody in the dilution 1:50 at 4°C in the dark for 30 min. Two additional washing steps with cold PBS were performed before the cells were analyzed by FACScalibur using CELLQuest software(Becton Dickinson, Mount View, CA). The same instrument settings were used for all of the experiments, and 5000 events were analyzed. MFI ratio was defined as: MFI of gated live cells stained with anti-Fas/MFI of cells stained with isotype-matched antibody.

A 100-μl suspension of 20 × 10³ cells was added to each well of flat-bottomed, 96-well plates (Costar, Corning,NY). After 24 h, various concentrations of the drugs (topotecan,gemcitabine, cisplatin, and paclitaxel) or the Fas agonistic mAb CLB-CD95/15 (31) was added to the cells. After 72 h,the cells were incubated with a solution of 5 mg/ml of MTT (Sigma Chemical Co., St. Louis, MO) dye for 4 h. The MTT crystals were solubilized with 100 μl of DMSO, and absorbance was measured at 540 nm, using Spectra Fluor (Tecan, Salzburg, Austria). Absorbance values were expressed as a percentage of untreated controls, and concentrations resulting in cell growth inhibition of 50%(IC₅₀) and 80% (IC₈₀) were calculated.

Cells were plated at a density of 5 × 10⁶cells in 75-cm² tissue culture flasks (Costar,Cambridge, MA) 24 h before treatment. For determination of Fas-induced apoptosis, cells were incubated with 5, 10, or 50 μg/ml of CLB-CD95/15 mAb or 100 ng/ml of recombinant FasL (Alexis Biochemicals, San Diego, CA) for 4–72 h. Drug-induced apoptosis was analyzed after incubation of the cells for 4–72 h with IC₅₀ and IC₈₀concentrations of topotecan, gemcitabine, cisplatin, or paclitaxel. The inhibitory anti-Fas mAb CLB-CD95/2 (10 μg/ml; Ref. 31)or neutralizing anti-FasL mAb NOK-2 (1 μg/ml; PharMingen, San Diego,CA) was applied 1 h before drug treatment. The extent of cell death was determined by PI staining of hypodiploid DNA, and the measurement of early apoptotic events was performed by annexin V staining. For the PI staining, 3 × 10⁵cells were resuspended in Nicoletti buffer as described (32) and analyzed by FACScan (Becton Dickinson, Mount View, CA). The fraction of cells with sub-G₁ DNA content was assessed by the Lysis program (Becton Dickinson). For annexin V staining (33), 3 ×10⁵ treated cells were washed, resuspended in 300μl of a binding buffer containing 2 mmCa²⁺, and incubated at room temperature for 15 min in the dark with 3 μl of annexin V-phycoerythrin (Nexins Research, Kattendijk, the Netherlands). Analysis was performed on FACScalibur using CELLQuest software (Becton Dickinson). The percentage of specific apoptosis was calculated subtracting the percentage of spontaneous apoptosis of the relevant controls from the percentage of total apoptosis.

Proteins for Western blot analysis were extracted from whole-cell pellets and lysed for 30 min at 4°C with 50 μl solution [1%Triton X-100, 150 mm NaCl, 10 mm Tris-HCl (pH 7.6), and 5 mm EDTA] per 1 ×10⁶ cells. Protease cocktail inhibitor tablets(Boehringer-Mannheim, Almere, the Netherlands) were freshly diluted in PBS before each experiment and added to the lysing solution. Protein concentration was assayed using the Bio-Rad assay (Bio-Rad Laboratories, Richmond, CA). Of each sample, 25 μg of protein/lane were separated on a 10% SDS-PAGE and electroblotted onto nitrocellulose membranes (Amersham, Braunschweig, Germany). Protein loading equivalence was assessed by the expression of β-actin. After blocking for 30 min in PBS supplemented with 5% BSA (Sigma) and 5%nonfat dry milk for 1 h at room temperature, immunodetection was performed using mouse anti-FasR and anti-FasL mAbs (Transduction Laboratories, Lexington, KY) at a concentration of 1:500 or anti-caspase-8 mAb (Immunotech, Marseille, France) at a concentration of 1:1000 at 4°C overnight, followed by horseradish peroxidase-conjugated goat-antimouse antibody. Enhanced chemiluminescence (ECL; Amersham, Braunschweig, Germany) was used for detection, and protein expression was quantified by densitometry of autoradiographs (model GS-690 Imaging densitometer; Bio-Rad, Richmond,CA).

Total RNA was extracted using RNAzol B (CINNA/Biotecx Laboratory, Inc.,Houston, TX). RNA was denatured at 55°C for 15 min in 50% formamide,electrophoresed through a 1% agarose gel containing formaldehyde, and blotted onto Qiabrane nylon membranes (Qiagen, Hilden, Germany). The membranes were hybridized with the full-length cDNA for Fas antigen and FasL (4, 7), kindly provided by Dr. Shigekazu Nagata,Osaka University Medical School, Osaka, Japan. The membranes were washed and autoradiographed for 48 h at −80°C. A cDNA probe for glyceraldehyde-3-phosphate dehydrogenase was used as a control for RNA loading.

The p53 gene of the six lung cancer cell lines was sequenced from exon 5 to exon 9 according to methods described previously (34, 35, 36). In brief, DNA was isolated, and a PCR reaction using primers for exons 5–9 was performed. The product was submitted to a new PCR round using specific γ-³³P end-labeled primers for each of the five exons to be sequenced. Subsequently, the product was separated on a 6% PAGE gel, fixed, and autoradiographed at room temperature overnight.

Caspase-8 activity was assessed using ApoAlert Caspase-8 Fluorescent Assay kit (Clontech Laboratories, Inc., Palo Alto, CA). Whole-cell pellets obtained from 2 × 10⁶ cells were resuspended in 50 μl of chilled lysis buffer and incubated on ice for 10 min. The cell lysates were centrifuged at 12,000 rpm at 4°C for 3 min, and the supernatants were collected. Subsequently, 50 μl of a reaction buffer containing DTT at a concentration of 10 mmand 5 μl of 1 mm IETD-AFC-conjugated substrate were added to the supernatants. The supernatants were then incubated at 37°C for 1 h in a water bath. Fluorescence was detected using a fluorometer equipped with a 400-nm excitation and a 505-nm emission filter. Fold-increase in the protease activity was determined by comparing the levels of the treated cells with the untreated controls.

Quantitative experiments were analyzed by use of Student’s t test. All Ps resulted from the use of two-sided tests and were considered significant when <0.05. Correlative data were analyzed using the Pearson correlation coefficient.

The expression of Fas and FasL was evaluated by FACS analysis using specific mAbs. As shown in Fig. 1, all cell lines revealed Fas and FasL protein expression. In four cell lines(H460, H322, GLC4, and GLC4/ADR), expression was quantitatively similar to that of Jurkat T-leukemia cells, used as a positive control, but H187 and N417 cells showed lower levels of both Fas and FasL (Fig. 1and Table 1). Western blot analysis,using mAb directed against different epitopes (see “Materials and Methods”) and Northern blotting confirmed the FACS results for both Fas and FasL (data not shown).

To investigate whether the constitutional expression of Fas in lung cancer cell lines would be sufficient for Fas-induced cytotoxicity, we exposed the cells to different concentrations of the Fas agonistic mAb CLB-CD95/15. Growth inhibition in lung cancer cells varied from none to 50%, whereas Jurkat T leukemia cells, used as a positive control,showed 100% growth inhibition (Fig. 2,A). To investigate whether this growth inhibition could be translated into apoptosis, cells were again exposed to different concentrations of the Fas agonistic mAb, and apo-ptosis was analyzed at different time points (4–72 h) by PI-stained hypodiploid DNA. As shown in Fig. 2,B, Fas-induced apoptosis was observed in four of six lung cancer cell lines. Compared with Jurkat cells, apoptosis in lung cancer cell lines occurred at a lower extent (not exceeding 25%) and started at later time points(24–48 h; Fig. 2,B). The complete resistance to the Fas agonistic mAb found in the growth inhibition assay in H187 and N417 lines was also detected when Fas-induced apoptosis was analyzed in these cells (Fig. 2). No clear difference in Fas sensitivity between NSCLC and SCLC cell lines was found, but a correlation(R² = 0.95; P = 0.06)between Fas expression and sensitivity to Fas-induced cytotoxicity was observed among the cell lines (Table 1). Fas-induced apoptosis was also assessed by annexin V staining and forward/side scatter analysis,leading to similar results (data not shown). Because the functionality of the Fas pathway is a prerequisite to investigate its role in drug-induced apoptosis, the four Fas-sensitive lung cancer lines (H460,H322, GLC4, and GLC4/ADR) were selected for the following experiments.

An involvement of the Fas signaling in drug-induced apoptosis would require an up-regulation of FasL, potentiated by an increase in Fas expression, to lead to the fratricide and autocrine mechanism of cell death (19, 23). Therefore, we assessed the levels of expression of these two molecules during drug-induced apoptosis in the Fas-sensitive lung cancer cell lines. We selected four anticancer agents active in lung cancer and directed against different intracellular targets (cisplatin, topotecan, gemcitabine, and paclitaxel). On the basis of our previous results (30), we exposed the cells to their respective IC₅₀ and IC₈₀ values, which were established for each of the four drugs (Table 2). The results were analyzed by Western blotting and Northern blotting at different time points (4–72 h). In none of the cell lines, regardless of the drug or concentration used, was FasL expression significantly increased when compared with the expression levels observed for the untreated controls (Fig. 3,A). Fas expression was assessed during the same experiments described above; a 2–3-fold increase in the Fas expression at protein and RNA levels was found in H460 but not in H322 cells (Fig. 3,B). GLC4 and GLC4/ADR cell lines showed similar results to H322 cells (data not shown). To have a better insight into the influence of p53 status on the up-regulation of the Fas levels upon drug exposure, the p53 gene was submitted to direct sequencing. H460 was found to be the only wt-p53 cell line (Table 3), whereas the other cell lines were shown to harbor p53 mutations. Similar results of FasL and Fas expression upon exposure to anticancer drugs were obtained using FACS analysis (data not shown).

Assuming that drug-induced cytotoxicity would be mediated by the Fas pathway, a correlation between drug and Fas-induced growth inhibition and apoptosis should be found, because they would share the same pathway to induce cytotoxicity. Therefore, we compared the induction of growth inhibition and apoptosis by anticancer drugs and by Fas agonistic mAb. No correlation was observed between sensitivity of the cells to anticancer drugs, regardless of their pharmacological mechanism of action, and Fas-induced cytotoxicity. In contrast, GLC4,the most sensitive cell line to the four drugs tested, was the least sensitive to Fas-induced cytotoxicity (Table 2). Furthermore, the doxorubicin-resistant cell line GLC4/ADR, also cross-resistant to topotecan and paclitaxel (Table 2), was five and three times more sensitive than the parental cell line GLC4 to Fas-induced growth inhibition and apoptosis, respectively (Table 2). A discrepancy between drug and Fas-induced cytotoxicity was also observed between the NSCLC cell lines H460 and H322 (Table 2). The correlation coefficient(R²) between the sensitivity of the cells to IC₈₀ concentration of cisplatin and to 50 μg/ml of the CLB/CD95–15 analyzed at 24–72 h was <0.05. Similar values were obtained when the other drugs (topotecan, gemcitabine, and paclitaxel) were used.

To further assess the role of the Fas pathway in drug-induced apoptosis in lung cancer cell lines, we also tested the ability of the anticancer drugs to induce apoptosis in the absence of a functionally active Fas signaling. Cells were preincubated either with the FasR-blocking specific mAb CLB-CD95/2 or the FasL-neutralizing mAb NOK-2 1 h preceding drug treatment. These two antibodies blocked almost completely apoptosis induced, respectively, by Fas agonistic mAb and recombinant FasL in Jurkat and lung cancer cell lines (Table 4). Nevertheless, the addition of these antibodies failed to protect from apoptosis induced by cisplatin,topotecan, gemcitabine, and paclitaxel in all of the lung cancer cell lines tested, regardless of the drug used. Similar results of the impact of Fas signaling blockade were obtained when different time points (4, 24, and 48 h) and lower drug concentrations(IC₃₀) were analyzed or if another FasL-neutralizing mAb (NOK-1) was used (data not shown). To further validate our results, we tested Jurkat T-cells (a prototype of a Fas-sensitive line) in similar experiments; the application of the blocking antibodies did not protect Jurkat T cells from drug-induced apoptosis (Table 4).

Caspase-8 (FLICE/MACH5) is an upstream cysteine protease, involved in apoptosis mediated by Fas and other death receptors (15). To obtain further insight into the Fas pathway during drug-induced apoptosis in lung cancer cells, we also assessed caspase-8 status. As described for other caspases, the activation of caspase-8 results in cleavage products. Although we observed that drug-induced apoptosis occurred without Fas/FasL signaling, we observed chemotherapy-induced caspase-8 activation, represented by intermediate cleavage products in the Western blotting (Fig. 4,A). Comparable kinetic cleavage was obtained with all lung cancer cell lines and also different drugs. To further confirm that the processed caspase-8 was active, its proteolytic activity was assessed upon drug exposure. A concentration- and time-dependent increase in the caspase-8 proteolytic activity up to 9-fold was observed, compared with the untreated controls, corroborating the results of the Western blot (Fig. 4 B). The increase in the caspase-8 proteolytic activity was similar, regardless of the drug used. Regarding the other cell lines,GLC4 and GLC4/ADR showed similar results to H460, but in H322 cells, a less intense increase in the proteolytic activity was observed (data not shown). Similar results in both Western blot and proteolytic activity assay were observed when lung cancer cells were incubated with Fas-antagonistic mAb CLB-CD95/2 prior to drug exposure in all lung cancer cell lines.

The expression of Fas and FasL has been demonstrated in a number of solid tumors and hematological malignancies (20, 37, 38, 39). The physiological role of the Fas/FasL pathway outside the immune system remains unclear, although a role for the Fas system has been suggested in drug-induced apoptosis in some cell types (16, 17, 18, 19, 20, 21, 22, 23). The present study demonstrates that both Fas and FasL were expressed in all of our six lung cancer cell lines (H460,H322, GLC4, GLC4/ADR, N417, and H187). N417 and H187 cells are characterized in flow cytometry by a relatively higher fluorescence background when stained with isotype-matched antibodies compared with other lung cancer cell lines. A possible influence of that fact on the results of the Fas and FasL expression is unlikely, because the FACS findings regarding the expression of Fas and FasL in the panel of lung cancer cells were confirmed by Western blotting (data not shown).

In addition, the Fas system was shown to be functional in four of the cell lines analyzed (H460, H322, GLC4, and GLC4/ADR), because apoptosis was induced in these cells after exposure to Fas agonistic antibody. Despite showing similar levels of Fas expression as compared with Jurkat cells, the lung cancer cells (H460, H322, GLC4, and GLC4/ADR)showed less sensitivity to Fas-induced apoptosis. A possible explanation for that is the expression of natural inhibitors of apoptosis in these lung cancer cell lines, such as decoy receptor 3 and FLICE-inhibitory protein (40, 41). These inhibitors block the Fas pathway acting, respectively, at the level of FasL (40) or FADD (41) and have been shown to be highly expressed in lung cancer tumors (40, 41).

Despite the presence of a functional Fas pathway, we observed that drug-induced apoptosis in lung cancer cells occurred independently from Fas/FasL signaling. This conclusion was substantiated by different results: (a) drug-induced apoptosis was not accompanied by a significant FasL up-regulation compared with untreated controls,irrespective of the compound, cell line, or drug concentration;(b) lack of correlation between sensitivity to chemotherapy and Fas-induced growth inhibition and apoptosis; (c)blockade of the Fas signaling system, either by a Fas-antagonistic mAb or a FasL-neutralizing mAb failed to inhibit chemotherapy-induced apoptosis.

The lack of FasL induction above the level observed for the untreated controls suggests no activation of the Fas/FasL signaling upon drug exposure. The absence of FasL up-regulation after drug exposure also found at the RNA level rules out a possible posttranscriptional or posttranslational mechanism that could mask an increase in FasL levels after drug exposure. This finding is in contrast with results reported in different cell lines (19, 20, 21, 22, 23, 29) but is in accordance with other reports (28, 39). This discrepancy may be attributable to different cell types or distinct experimental conditions. The results showing up-regulation of FasL protein and RNA levels in both treated and untreated cells (Fig. 3,A) suggest that the levels of this molecule may vary physiologically in these cells or be induced by changes in cell culture conditions and highlight the importance of one control for each time point analyzed. We observed drug-induced Fas up-regulation in the wt-p53 H460 cells; however,because it occurred in the absence of FasL induction, we can speculate that it was probably not sufficient to activate the proposed autocrine/paracrine death loop (19) in this cell line. In fact, this Fas up-regulation might constitute an epiphenomenon, being a result of p53 up-regulation, because we described p53 induction in this cell line upon drug exposure (30). The absence of Fas induction in the mt-p53 cell lines (H322, GLC4, and GLC4/ADR) is in line with previous reports (21, 42) and suggests that the Fas gene is under the transcription control of a functional p53. The discrepancy between drug- and Fas-induced cytotoxicity was evidenced by the fact that the most Fas-sensitive lung cancer cell lines were not necessarily the most drug-sensitive ones. In particular, there was a lack of correlation between drug- and Fas-induced cytotoxicity when GLC4 and the mrp-overexpressing GLC4/ADR were compared. GLC4/ADR has been shown to be at least 10 times more resistant to doxorubicin than the parental line (43), and in our hands, it was also found to be cross-resistant to topotecan and paclitaxel (Table 2). However, GLC4/ADR was three times more sensitive to Fas-induced apoptosis than the parental cell line GLC4. Moreover,because the lack of correlation between Fas and chemotherapy-induced cytotoxicity was irrespective of the drug used, the possibility of a misleading analysis attributable to a drug-specific mechanism of resistance is unlikely. These findings contrast with previous reports that showed cross-resistance between Fas and drug-induced apoptosis, as well as reduced Fas antigen expression in the resistant cell lines in comparison with the parental ones (20, 44). Our findings suggest that chemotherapy and Fas-induced apoptosis do not share the same pathway. This, however, does not rule out that both pathways may converge at some downstream point, as suggested recently (24). No protection from drug-induced apoptosis was provided by the blockage of the Fas signaling system, either by Fas-antagonistic mAb or a FasL-neutralizing mAb, in any of the cell lines tested. The same results were obtained irrespective of the drug used, excluding the possibility of the results being attributable to cell line, drug concentration, or anticancer agent-specific effect. Moreover, the observation of similar findings when Jurkat T cells (a prototype of a Fas-sensitivity line) were used exclude that the results obtained in lung cancer cell lines were attributable to a relatively low Fas sensitivity of these cell lines.

Taken together, our results indicate independence of drug-induced apoptosis from the Fas/FasL signaling pathway in lung cancer cells. Our findings are consistent with recent reports showing that chemotherapy-induced apoptosis occurs in the absence of Fas/FasL interaction (24, 25, 26, 27, 28, 29, 45). According to these reports,different approaches to inhibit the Fas pathway, such as antagonistic antibodies (26, 28, 29), overexpression of natural inhibitors like FLICE-inhibitory protein (27), or the use of FADD-null cells (45), do not protect from chemotherapy-induced apoptosis. Nevertheless, our results are in contrast with previous data proposing that the Fas/FasL signaling pathway mediates drug-induced programmed cell death (16, 17, 18, 19, 20, 21, 22, 23). The differences among these reports, especially those in solid tumor lines (21, 22) and ours, might be attributable to differences in the cell type or to difference in the drugs used in the studies, because it has been suggested that not all of the drugs would depend on the Fas pathway to induce apoptosis (46). Nonetheless, the use of a panel of lung cancer cells, composed of both NSCLC and SCLC lines, the confirmation of our findings in lung cancer cells when additional experiments using Jurkat T-cells were performed, in addition to observation of similar results with the use of four anticancer drugs with different pharmacological mechanisms of action, give consistence to our results. Differences in the reagents used, in particular the specificity of the Fas-antagonistic mAbs, should also not be neglected as a possible explanation for the discrepancy in the results.

Because caspase-8 is an essential caspase in the Fas pathway, its status was also analyzed. Although drug-induced apoptosis did not require Fas/FasL signaling, caspase-8 was activated in a drug-, time-,and concentration-dependent fashion in all of the cell lines. This suggests that caspase-8 can also be activated by chemotherapy in a Fas-independent way. To the best of our knowledge, this is the first report of Fas-independent caspase-8 activation induced by conventional chemotherapeutic agents in solid tumor cells, being in line with recent reports obtained in leukemia cells (25, 47, 48, 49, 50). Because caspases form a redundant system, the finding of chemotherapy-induced caspase-8 activation in lung cancer cells certainly requires further investigation to define whether this is relevant for drug-induced apoptosis in solid tumor cells or whether it constitutes only a secondary event. Another issue that requires additional investigation is the mechanism responsible for caspase-8 processing and activation independently of Fas. Activation downstream of caspase-9, as recently described in cell-free extracts (47), or an interaction with apoptotic protease activating factor-1, reported previously (51), may constitute possible scenarios for drug-induced caspase-8 activation in lung cancer cells. Nevertheless, a role for another death receptor, such as tumor necrosis factor-related apoptosis-inducing ligand or tumor necrosis factor, in the processing of caspase-8 cannot be ruled out.

The results presented here may have clinical relevance, not only because we show that lung cancer cells do not depend upon Fas/FasL signaling to undergo drug-induced apoptosis but also because we identify a functional Fas pathway in most of the lung cancer cell lines. The possibility of inducing Fas-mediated apoptosis in vivo has been proposed as a potential approach for anticancer therapy (5, 52). Some reports have suggested a synergistic effect in vitro when chemotherapy and FasL or Fas-agonistic mAbs are combined (28, 53). Because drug and Fas-induced apoptosis involve two different and not completely overlapping pathways in lung cancer cells, the possibility to combine chemotherapy and immunotherapy can be envisaged as a new strategy for anticancer treatment.

We thank Dr. S. Nagata (University of Osaka, Osaka, Japan), who provided Fas and FasL RNA probes. We are especially grateful to Dr. Lucien Aarden (Centraal Laboratorium Bloedtransfusidienst, the Netherlands) for productive discussion.