Chemotherapy enhances TNF-related apoptosis-inducing ... - Nature

Oncogene (2003) 22, 1807–1816

& 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc

Chemotherapy enhances TNF-related apoptosis-inducing ligand DISC assembly in HT29 human colon cancer cells Sandrine Lacour1,3, Olivier Micheau2,3, Arlette Hammann1, Ve´ronique Drouineaud1, Jurg Tschopp2, Eric Solary1 and Marie-The´re`se Dimanche-Boitreln,1 1 INSERM U517, IFR 100, Faculties of Medicine and Pharmacy, 7 boulevard Jeanne d’Arc, 21033 Dijon Cedex, France; 2Institute of Biochemistry, University of Lausanne, BIL Biomedical Research Center, CH-1066 Epalinges, Switzerland

Cytokines such as Fas-ligand (Fas-L) and Tumor Necrosis Factor-Related Apoptosis-Inducing Ligand (TRAIL) can induce human colon cancer cell apoptosis through engagement of their death domain receptors. All the cancer cells are not sensitive to these cytokines. We have shown recently that low doses of cytotoxic drugs could restore TRAIL-induced cell death in resistant colon cancer cell lines. The present work further explores the death pathway triggered by the cytotoxic drug/TRAIL combination in HT-29 colon cancer cells (www.alexiscorp.com). Clinically relevant concentrations of cisplatin, doxorubicin and 5-fluorouracil synergize with TRAIL to trigger HT-29 cell death. Activation of this pathway leads to apoptosis that involves both caspases and the mitochondria. An increased recruitment of Fas-associated death domain (FADD) and procaspase-8 to the TRAIL-induced death-inducing signaling complex (DISC) was shown in cells exposed to anticancer drugs. Following caspase-8 activation at the DISC level, the mitochondria-dependent death pathway is activated, as demonstrated by the cleavage of Bid, the dissipation of DWm, the release of mitochondrial proteins in the cytosol and the inhibitory effect of Bcl-2 expression. Importantly, besides mitochondrial potentiation, we show here that cytotoxic drugs sensitize HT-29 colon cancer cells to TRAIL-induced cell death by enhancing FADD and procaspase-8 recruitment to the DISC, a novel mechanism whose efficacy could depend partly on Bcl-2 expression level. Oncogene (2003) 22, 1807–1816. doi:10.1038/sj.onc.1206127

Introduction Chemotherapy has demonstrated limited efficacy in the treatment of colon cancer because of intrinsic or acquired drug resistance (Gorlick and Bertino, 1999). Engagement of cell surface death receptors of the Tumor Necrosis Factor (TNF) receptor superfamily was proposed as an alternative strategy to trigger colon *Correspondence: Marie-The´re`se Dimanche-Boitrel; E-mail: [email protected] 3 Both authors contributed equally to this work. Received 29 July 2002; revised 8 October 2002; accepted 9 October 2002

cancer cell death by inducing apoptosis (Baker and Reddy, 1998). These type I transmembrane proteins share homology in their cytoplasmic ‘death domain’, a protein–protein interaction module that ensures downstream cell-death signaling (Baker and Reddy, 1998). Their ligands include TNFa, Fas-ligand (Fas-L), TNFRelated Apoptosis-Inducing Ligand (TRAIL) and Tweak. Although TNFa and Fas-L can be efficient in killing some colon carcinoma cells in vitro, their lifethreatening toxicity when administered systematically currently precludes their clinical use (Chapman et al., 1987; Tanaka et al., 1997). Cell-bound TRAIL or its soluble form was demonstrated to selectively induce apoptosis in some transformed epithelial cells (Wiley et al., 1995; Pitti et al., 1996) and TRAIL administration in SCID and nude mice was effective at reducing growth of colon carcinoma xenografts without any of the toxic effects shown with administration of Fas-L or TNFa (Ashkenazi et al., 1999; Walczak et al., 1999). From these observations, TRAIL appears as a promising cytokine for the treatment of various carcinomas including colon cancers. TRAIL transduces an apoptotic signal by binding to DR4 (TRAIL-R1) and DR5 (TRAIL-R2), which both contain a death domain in their cytoplasmic tail (Golstein, 1997). The death signaling pathway involves the recruitment to the DR4 and DR5 death domain of the adaptor molecule Fas-associated death domain (FADD) in a death-inducing signaling complex (DISC) (Chaudhary et al., 1997; Schneider et al., 1997). In turn, FADD recruits procaspase-8 and -10 in the DISC through homophilic interactions between death effector domains (DED) contained in the amino-terminus of FADD and the prodomain of procaspase-8 and -10 (Bodmer et al., 2000; Kischkel et al., 2001). Two complementary pathways were shown to lead from caspase-8 activation in the DISC to apoptosis (Scaffidi et al., 1998). The first one, in which caspase-8 directly activates downstream effector caspases, does not involve the mitochondria and escapes the resistance mediated by overexpression of Bcl-2 and related anti-apoptotic proteins. The second pathway involves the cleavage of Bid, a BH3-domain-only proapoptotic protein that belongs to the Bcl-2 family (Li et al., 1998). The truncated Bid (tBid) induces mitochondrial changes, including the translocation of proapoptotic proteins

Chemotherapy enhanced TRAIL DISC assembly S Lacour et al

1808

such as cytochrome c, SMAC/DIABLO and apoptosisinducing factor (AIF) from the mitochondria to the cytosol where they activate the caspase cascade or to the nucleus where they trigger DNA fragmentation and chromatin condensation (Li et al., 1998; Kroemer and Reed, 2000; Deng et al., 2002). Other receptors to TRAIL have been identified, including DcR1 (TRAIL-R3) and DcR2 (TRAIL-R4) that have no cytoplasmic tail and a truncated form, respectively (Golstein, 1997). These two receptors bind TRAIL with an affinity comparable to that of DR4 and DR5, but do not mediate apoptosis upon ligation and were proposed to act as decoy receptors whose expression determines whether a cell is resistant or sensitive to TRAIL (Sheridan et al., 1997). The apoptotic response to TRAIL is also modulated by cFLIP, a protein with a structure similar to that of procaspase-8 but lacking the catalytic site and competitively interacting with FADD and procaspase-8 at the DISC (Zhang et al., 1999). We and others have shown previously that subtoxic concentrations of chemotherapeutic drugs could restore the response to Fas-L and TRAIL in cancer cells that are not spontaneously sensitive to these cytokines (Micheau et al., 1997; Lacour et al., 2001). In addition, TRAIL was shown to cooperate synergistically with various chemotherapeutic drugs to induce cancer cell death in vitro and tumor regression in animal models (Ashkenazi et al., 1999). Tumor cell sensitization to TRAIL was initially related to p53-dependent and -independent changes in death receptor expression (Sheikh et al., 1998; Gibson et al., 2000; Nagane et al., 2001). However, we have shown recently that cytotoxic drug-induced sensitization of colon carcinoma cells to TRAIL could occur in the absence of significant changes in the membrane expression of TRAIL receptors, either agonistic or decoy receptors (Lacour et al., 2001). The present study further explores the molecular mechan-

isms of cytotoxic drug-induced sensitization to TRAIL and identifies the death pathway activated by a cytotoxic drug/TRAIL combination in the HT-29 colon carcinoma cell line.

Results Antitumor drugs selectively enhance HT-29 cell sensitivity to death receptor ligands We first compared HT-29 colon carcinoma cell response to TNFa and three TNF-related death ligands including Fas-L, TRAIL and Tweak. We have shown previously that cytotoxic drugs could sensitize HT-29 cells to Fas agonists and TRAIL (Micheau et al., 1997; Lacour et al., 2001). By testing a large range of cytokine concentrations, we show here that a 16-h exposure of HT-29 cells to Fas-L or TRAIL induces limited cytotoxicity, as determined by using a methylene blue assay, while these cells are completely resistant to TNFa and Tweakinduced cytotoxicity (Figure 1). A 24-h pretreatment with nontoxic concentrations of several anticancer drugs including cytotoxic drugs cisplatin (CDDP) (5 mg/ml), doxorubin (DXR) (2.5 mg/ml) and 5-fluorouracil (5-FU) (100 mg/ml) increased HT-29 response to Fas-L and TRAIL, while it demonstrated little effect on HT-29 cell response to TNFa and Tweak (Figure 1). We then focused on the mechanisms of anticancer drug-induced sensitization to TRAIL. Antitumor drugs synergize with TRAIL to trigger caspase-dependent apoptosis of HT-29 cells A pretreatment with anticancer drugs is not required for HT-29 cell sensitization to TRAIL. Indeed, a combination of either CDDP (5 mg/ml), DXR (2.5 mg/ml) or 5FU (100 mg/ml) with TRAIL (200 ng/ml with 2 mg/ml

Figure 1 Cytotoxic drugs specifically sensitize HT-29 cells to Fas-L- and TRAIL-induced cytotoxicity: HT-29 cells were either left untreated (open diamond), or exposed to 5 mg/ml CDDP (black squares), 2.5 mg/ml DXR (black triangles) or 100 mg/ml 5-FU (black circles) for 24 h, then treated with indicated concentrations of TRAIL, Fas-L, TNFa or TWEAK for 16 h. Viability was measured by using a methylene blue assay. Results are expressed as the percentage of viability compared to cells that were not exposed to the cytokine Oncogene


1809

Figure 2 Synergistic induction of caspase-dependent apoptosis by the cytotoxic drug/TRAIL combination: (a) HT-29 cells were either left untreated (NT) or treated for 24 h with either cisplatin (CDDP; 5 mg/ml), doxorubicin (DXR; 2.5 mg/ml), 5-fluorouracil (5-FU; 100 mg/ml), TRAIL (200 ng/ml with 2 mg/ml anti-Flag Ab) or indicated combinations of a cytotoxic drug and TRAIL. Apoptosis was determined by staining nuclear chromatin with Hoechst 33342. Results are the mean7s.d. of three independent experiments. (b) Immunoblot analysis of indicated proteins. Numbers are molecular weights in kD. Asterisks (*) indicate cleaved peptides. HSC70 expression was studied for loading control. One representative of three independent experiments is shown. (c) Cleavage activity of lysates from cells treated as above was studied using indicated peptides and calculated as described in Materials and methods. One representative of three independent experiments is shown

anti-Flag M2 Ab) for 24 h induces apoptosis in 40–70% of cells, as determined by Hoechst 33342 staining of the nuclear chromatin, while all these agents induce less than 12% apoptosis when tested alone at the studied concentrations (Figure 2a). To determine whether apoptosis induced by the combination of chemotherapeutic agents and TRAIL was mediated by caspase activation, we performed Western blot experiments (Figure 2b). Activation of caspase-3, -8 and -9 was easily identified in HT-29 cells exposed to any of the cytotoxic drug/TRAIL combinations, as evidenced by the decreased expression of the proform and the appearence of cleavage products (Figure 2b). The cytosolic BH3 domain-only protein Bid, which is a target for caspase-8, and the nuclear enzyme poly(ADPribose) polymerase, which was described as a caspase-3 target, were cleaved in HT-29 cells exposed to the cytotoxic agent/TRAIL combination. In contrast, none of these events could be identified in HT-29 cells exposed to TRAIL alone or a cytotoxic drug alone (Figure 2b). Caspase activation in cells exposed to the cytotoxic drug/TRAIL combination was further supported by the ability of cell lysates to cleave DEVD-AMC, IETDAFC and LEHD-AFC peptide substrates that mimic caspase-3, -8 and -9 target sites, respectively. In agreement with the immunoblotting results, cytotoxic drugs and TRAIL did not generate a peptide

cleavage activity in HT-29 cells when tested alone (Figure 2c). Antitumor agents enhance TRAIL DISC formation We have shown previously that the ability of cytotoxic drugs to sensitize colon carcinoma cells to TRAILmediated cell death could not be related to an increased expression of DR4 and DR5 receptors at the surface of tumor cells, nor to a decreased expression of TRAIL decoy receptors DcR1 and DcR2 (Lacour et al., 2001). This observation did not rule out the possibility that cytotoxic drugs could facilitate the formation of the DISC in response to TRAIL exposure. To check this hypothesis, we performed immunoprecipitation experiments in which HT-29 cells were treated or not with CDDP (5 mg/ml), DXR (2.5 mg/ml) or 5-FU (100 mg/ml) for 24 h. Then, cells were incubated with TRAIL for up to 30 min at 371C. In HT-29 cells that were not exposed to a cytotoxic drug, TRAIL recruits DR4 and DR5, the adaptator molecule FADD, procaspase-8 and the inhibitor FLIP (long and short isoforms) (Figure 3a) but does not recruit procaspase-3, used as a negative control (not shown). In contrast to what had been observed by flow cytometry method in nonpermeabilized cells (Lacour et al., 2001), immunoprecipitation experiments suggested that pretreatment Oncogene


1810

Figure 3 Influence of cytotoxic drugs on the formation of the DISC in response to TRAIL: (a) HT-29 cells were either left untreated (NT) or exposed to CDDP (5 mg/ml) or DXR (2.5 mg/ml) or 5-FU (100 mg/ml) for 24 h, then treated for indicated times with TRAIL (500 ng/ml with 2 mg/ml anti-Flag Ab). DISCs were immunoprecipitated with protein G and analysed for indicated proteins by immunoblotting. In unstimulated controls (0), TRAIL and M2 mAb were added after lysis to immunoprecipitate nonstimulated TRAIL receptors. Numbers are molecular weights in kDa. Asterisks (*) indicate cleaved peptides. One representative of three different experiments is shown. (b) Immunoblot analysis of DISC constituents. Numbers are molecular weights in kD. One representative of three independent experiments is shown. (c) HT-29 cells were either left untreated (NT), or exposed to 5-FU (100 mg/ml) for 24 h, then treated for indicated times with TRAIL (250 ng/ml with 2 mg/ml anti-Flag Ab) (pretreatment) or were simultaneously exposed to TRAIL with 5-FU for indicated times (cotreatment). DISCs were immunoprecipitated as described above, with the exception that TRAIL was not added to lysates of unstimulated cells (time point 0). One representative of three different experiments is shown

with cytotoxic drugs could increase the intracellular amount of DR4 and DR5 (Figure 3a, c). This effect was associated with an enhanced recruitment of FADD and procaspase-8 to DR4 and DR5 that may cause more caspase-8 processing. This enhanced recruitment could not be explained by an increase in intracellular amount of FADD, procaspase-8 or procaspase-10 under treatment by CDDP, DXR or 5FU (Figure 3b). Only 5-FU treatment was observed to decrease the intracellular amount of cellular FLIP in Oncogene

HT-29 cells as soon as 4 h (Figure 3b). Accordingly, less FLIP was recruited to the DISC in cells pretreated with 5-FU (Figure 3a). To determine whether cotreatment with TRAIL and the cytotoxic drug also facilitates the DISC assembly, we compared immunoprecipitation experiments in which HT-29 cells were pretreated with 5-FU (100 mg/ ml) for 24 h and incubated with TRAIL, and those in which HT-29 cells were cotreated with 5-FU and TRAIL up to 30 min. The enhanced DISC formation


1811

was observed in both situations, although it was less important under cotreatment (Figure 3c).

Antitumor drugs synergize with TRAIL to activate the mitochondrial pathway to cell death The truncation of Bid observed in HT-29 cells exposed to the cytotoxic drug/TRAIL combination suggested involvement of the mitochondria in cell death. To confirm the role of the mitochondria, we first analysed their transmembrane potential using a lipophilic cation probe (DePsipher or 5, 50 , 6, 60 -tetrachloro-1, 10 , 3, 30 tetraethylbenzimidazolyl carbocyanine iodide). While neither TRAIL alone nor any of the three cytotoxic drugs tested alone were able to trigger the decrease in red fluorescence (FL-2) that indicates mitochondria membrane potential disturbance, a significant decrease was observed in HT-29 cells exposed to the cytotoxic drug/TRAIL combination and was prevented by addition of the permeant caspase-8 inhibitor z-IETD-fmk (Figure 4a). These changes in DCm were associated with the appearance of three proapoptotic proteins, namely cytochrome c, Smac/DIABLO and apoptosis-inducing factor (AIF), in the cytosol of HT-29 cells exposed to the cytotoxic drug/TRAIL combination (Figure 4b). Cytochrome c relocalization from the mitochondria to the cytosol was further confirmed by confocal microscopy analysis of HT-29 cells treated by a CDDP/TRAIL combination for 24 h (Figure 4c). In addition, HT-29 cells that express Bcl-2 through stable transfection were more resistant to apoptosis induced by the cytotoxic drug/TRAIL combination (Figure 4d) or by drug pretreatment followed by TRAIL addition (data not shown) than HT-29 cells transfected with an empty vector. Expression of Bcl-2 in HT-29 cells that do not express the protein spontaneously also prevented the DEVD-AMC and LEHD-AFC cleavage activities in lysates of cells exposed to the cytotoxic drug/TRAIL combination, while it demonstrated little effect on IETD-AFC cleavage activity in this model (Figure 4e), suggesting that caspase-8 activation was not influenced by Bcl-2.

Caspase-8 activation, Bid cleavage and mitochondrial transmembrane potential decrease are early events in apoptosis triggered by the cytotoxic drug/TRAIL combination We then studied the kinetics of the previously described molecular events that lead from TRAIL exposure to HT-29 cell death. In these experiments, HT-29 cells were exposed to the DXR/TRAIL combination. We observed that caspase-8 activation, Bid cleavage into its active form tBid, cytochrome c release (Figure 5a) and the decrease in DCm (Figure 5b) could be detected about 6 h after the beginning of cell treatment. A faint band corresponding to active caspase-9 is also observed after 6 h of treatment. Caspase-3 and PARP cleavage were detected at a later time point (Figure 5a).

Discussion The discovery that ligands for endogenous cell surface receptors can induce cell death provided excitement regarding the potential interest of these ligands for selectively inducing apoptosis of cancer cells, particularly those that are otherwise resistant to chemotherapeutic drugs. In this setting, TRAIL proved unique in that its systemic administration does not induce the toxicity observed with Fas-L and TNFa in animal models including mice and nonhuman primates (Ashkenazi et al., 1999; Walczak et al., 1999). In addition, TRAIL frequently induces apoptosis of malignant cells, both in vitro and in xenograft murine models, without causing the death of normal cells (Ashkenazi et al., 1999). A series of recent reports has demonstrated that clinically relevant concentrations of various anticancer drugs could restore the apoptotic response to TRAIL in cancer cells that are not spontaneously sensitive to this cytokine (for review, see Nagane et al., 2001), including colon cancer cell lines (Lacour et al., 2001). All the chemotherapeutic drugs are not equal in sensitizing tumor cells to TRAIL-induced apoptosis, for example, DXR and 5-FU are potent sensitizers, while methotrexate can be inefficient (Keane et al., 1999). The molecular mechanisms of this sensitization and the death pathways activated by the TRAIL/cytotoxic drug combinations remain a matter of controversy. We show here that, in HT-29 colon cancer cells, cytotoxic drugs enhance the recruitment of FADD and procaspase-8 to the DISC whose formation is induced by TRAIL. We also show that the TRAIL/cytotoxic drug combination activates a death signaling pathway that involves the mitochondria, thus remaining dependent on Bcl-2 protein level expression. HT-29 human colon carcinoma cells are spontaneously resistant to the four tested TNF superfamily cytokines including Fas-L, TRAIL, TNFa and Tweak. We show that several commonly used anticancer drugs at clinically relevant concentrations specifically sensitize these cells to Fas-L- or TRAIL-induced cytotoxicity, but demonstrate little influence on TNFa and Tweakmediated apoptosis. We have previously reported that cytotoxic drug-induced sensitization of HT-29 cells to Fas-L was associated with an increased expression of Fas receptor at the cell surface (Micheau et al., 1997). In contrast, anticancer drugs do not induce any significant change in the expression of TRAIL receptors at the surface of HT-29 cells (Lacour et al., 2001). Cytotoxic drug-induced changes in TRAIL receptor expression through p53-dependent and -independent mechanisms have been identified in other cell types (Sheikh et al., 1998). However, it was shown recently in glioma cell lines that such events may not account for the cytotoxic drug/TRAIL synergy (Rohn et al., 2001). Three years ago, Scaffidi et al. (1998) have reported two distinct Fas signal transduction pathways leading to apoptosis: mitochondria-dependent in type II cells and -independent in type I cells. Overexpression of Bcl-2 inhibited the mitochondria-dependent pathway, but not the mitochondria-independent pathway. It has been Oncogene


1812

Figure 4 Apoptotic pathway triggered by the cytotoxic drug/TRAIL combination in HT-29 cells involves the mitochondria: (a) Polarization of the mitochondria membrane was studied using the DePsipher probe and analysed by flow cytometry in cells left untreated (control) or treated with TRAIL alone (200 ng/ml with 2 mg/ml anti-Flag Ab), each cytotoxic drug alone (CDDP; 5 mg/ml. DXR; 2.5 mg/ml. 5-FU; 100 mg/ml) or a combination of both for 24 h. A decrease in FL-2 fluorescence indicates changes in the mitochondrial membrane polarization. One representative of three independent experiments is shown. Numbers indicate the percentage of cells exhibiting a decrease in DCm. (b) The expression of indicated proteins was studied by immunoblotting in cytosolic extracts prepared from cells treated as described in Figure 2. HSC70 expression was used as a loading control. Numbers are molecular weights in kDa. One representative of three different experiments is shown. (c) HT-29 cells were left untreated (control) or treated with CDDP alone, TRAIL alone or the combination of both for 24 h, then labeled for confocal laser scanning microscopy. Mitochondrial Hsp70 (mHSP70) was used as a control. (d) One HT-29 cell clone transfected with the empty vector (Co) and two Bcl-2-expressing clones through stable transfection (c9, dark gray bars and c10, light gray bars) were treated as described in Figure 2 before measuring the percentage of Hoechst 33342-stained cells with an apoptotic phenotype. Results are the mean7s.d. of three independent experiments. (e) Caspase activities were measured by determining the ability of cell lysates from control-transfected (black bars) and Bcl-2-expressing (c10; dark gray bars) HT-29 cells treated as in Figure 2 to cleave indicated peptide substrates. One representative of three independent experiments is shown

recently shown that the TRAIL receptor- and Fasmediated apoptosis share the same death signal pathway and involve both mitochondria-dependent and -independent pathways (Suliman et al., 2001). Involvement of the mitochondrial pathway in response to the TRAIL/ cytotoxic drug combination has been assessed by Oncogene

caspase-8 activation, Bid cleavage and mitochondrial transmembrane potential decrease. All these events occurring 6 h after the beginning of HT-29 cell treatment, preceded the activation of downstream caspases such as caspase-3, and are inhibited by the permeant inhibitor z-IETD-fmk that targets mainly caspase-8.


1813

Figure 5 Time course of the molecular events triggered by the DXR/TRAIL combination: (a) Immunoblot analysis of indicated proteins in HT-29 cells left untreated (NT) or exposed for 24 h to DXR (2.5 mg/ml), TRAIL alone (200 ng/ml with 2 mg/ml anti-Flag Ab) or treated for indicated times with the DXR/TRAIL combination. Numbers are molecular weights in kDa. Asterisks (*) indicate cleaved peptides. HSC70 was studied as a loading control. One representative of three different experiments is shown. (b) HT-29 cells were treated for indicated times with the DXR/TRAIL combination before studying the mitochondria membrane polarization by using the DePsipher probe as in Figure 4. One representative of three independent experiments is shown. Numbers indicate the percentage of cells exhibiting a decrease in DCm

This latter caspase was recently demonstrated to be required for TRAIL-induced apoptosis (Bodmer et al., 2000; Seol et al., 2001) and to cleave Bid in response to TRAIL exposure (Yamada et al., 1999). Only in type II cells, apoptosis can be blocked by overexpressed Bcl-2 or Bcl-XL. The consequences of Bcl-2 expression in HT-29 cells further argue for a mitochondria involvement in cell death triggered by the cytotoxic drug/TRAIL combination as well as by drug pretreatment followed by TRAIL addition. The bcl-2 gene was first identified in follicular lymphomas as a consequence of a t(14;18) chromosomal translocation (Bisiau et al., 1995). Bcl-2 was also found to be overexpressed in various tumor types including colon cancers in the absence of gene rearrangements (Buglioni

et al., 1999). Bcl-2 is localized on the external mitochondrial membrane and prevents the release of proapoptotic molecules such as cytochrome c from the mitochondria to the cytosol (Gottlieb, 2001). Since most chemotherapeutic drugs mediate apoptosis mainly by inducing mitochondrial dysfunction, Bcl-2 behaves as a potent multidrug resistance factor. We show here that Bcl-2 expression partially prevents caspase-9 and -3 activation induced in HT-29 cells by the combination of TRAIL with cytotoxic agents, while it has little effect on caspase-8 activity. These results are in accordance with previous observations made in several other human tumor cell lines (Munshi et al., 2001; Suliman et al., 2001; Sun et al., 2001; Fulda et al. 2002) and further indicate that the cytotoxic drug/TRAIL combination Oncogene


1814

activates the mitochondrial pathway to cell death in colon cancer cells. Importantly, we show here that the ability of subtoxic concentrations of chemotherapeutic drugs to overcome the resistance of human colon cancer cells to TRAILinduced apoptosis is related to an enhanced assembly of the TRAIL-induced DISC. A 24 h exposure of HT-29 cells to low concentrations of either CDDP or DXR or 5-FU increases the recruitment of FADD and procaspase-8 in the complex whose formation is provoked by interaction of TRAIL with either DR4 or DR5. The increased recruitment of procaspase-8, which occurs as soon as 2 min after TRAIL addition in HT-29 cells pretreated with an anticancer drug, may facilitate caspase-8 activation by autocatalytic processing (Martin et al., 1998) that is followed by cell death. The increased recruitment of procaspase-8 is not related to a druginduced upregulation of procaspase-8 or downregulation of procaspase-10. Such an increased recruitment of procaspase-8 is also observed in HT-29 cells simultaneously cotreated with 5-FU and TRAIL, although the expression of DR4 and DR5 was not increased in these experimental conditions. The mechanism by which the DISC formation is facilitated by cytotoxic drugs remains poorly understood. While we initially failed to detect an increased expression of DR4 and DR5 at the surface of tumor cells exposed to cytotoxic drugs (Lacour et al., 2001), we show here that these agents increase the intracellular amount of both receptors, as detected by immunoprecipitation, which may contribute to facilitate the DISC formation. Another important determinant of the response to TRAIL in colon carcinoma cells is c-FLIP (Burns and El-Deiry, 2001). The long isoform of c-FLIP are observed to be recruited to the DISC in response to TRAIL receptor engagement in HT-29 cells. A decreased recruitment of c-FLIP was identified in cells exposed to 5-FU, as recently reported in cancer cells exposed to ligands of peroxysome proliferatoractivated-gamma (Kim et al., 2002), while it was not detected in those exposed to either CDDP or DXR. This observation indicates that c-FLIP recruitment to the DISC is not the sole determinant of HT-29 cell sensitization to TRAIL by cytotoxic drugs and that distinct mechanisms exist. One of these mechanisms could be the facilitation of TRAIL receptor aggregation by anticancer drugs, as demonstrated previously for Fas receptor (Micheau et al., 1999b). Such an aggregation, in which plasma membrane microdomains could play a role (Ikonen, 2001; Hueber et al., 2002), may account for the enhanced TRAIL DISC assembly observed in HT29 cells after a 24-h drug pretreatment as well as after a TRAIL/cytotoxic drug cotreatment. Additional events that were associated with cytotoxic drug-mediated sensitization to TRAIL include the suppression of NF-kB induction (Yamanaka et al., 2000; Micheau et al., 2001), the downregulation of glutathione S-transferase-p (Mizutani et al., 1999) and the transcriptional upregulation of FADD and some procaspases (Droin et al., 1998; Micheau et al., 1999a). Present understanding of TRAIL signaling pathways appears Oncogene

incomplete, suggesting that so far unidentified molecular actors of these pathways could also account for our observations. In conclusion, our study demonstrates that several anticancer drugs sensitize HT-29 colon cancer cells to TRAIL-induced apoptosis by facilitating the formation of the DISC and the activation of caspase-8. This enzyme then cleaves the proapoptotic Bid and recruits the mitochondrial amplifier to activate a downstream caspase cascade that leads to the cell dismantle. The synergy between anticancer drugs and TRAIL may help to overcome most resistance mechanisms to chemotherapeutic agents. However, the efficacy of this combination in colon cancer treatment still depends on Bcl-2-related family members, which could impair its cytotoxic efficacy.

Materials and methods Cell lines and apoptotic assays The HT-29 human colon carcinoma cell line was obtained from ATCC (Rockville, MD, USA) and cultured in Eagle’s minimum essential medium (EMEM) (Biowhittaker Co., Fontenay sous Bois, France) complemented with 10% fetal calf serum (GibcoBRL, Life Technologies, Cergy Pontoise, France) and 2 mm l-glutamine (Biowhittaker Co.). Stable transfection of HT-29 cells with psFF-Bcl-2 containing fulllength human Bcl-2 cDNA (kindly provided by Dr J Breárd INSERM 461, France) or the empty vector was obtained by the use of lipofectamine Plus reagent (GibcoBRL). Single clones were picked 2–3 weeks after transfection and selected in a medium containing 1 mg/ml geneticin (Sigma-Aldrich, St Quentin Fallavier, France). For apoptotic assays, cells (7 105) were seeded in 6-well flat-bottomed plates for 24 h before treatment. Apoptosis was identified by staining nuclear chromatin with 1 mg/ml Hoechst 33342 for 15 min at 371C, then identifying morphological changes by fluorescence microscopy. Cell viability was assessed by methylene blue as described (Micheau et al., 1997). When used, the caspase inhibitor z-IETD-fmk was incubated for 1 h prior to the addition of cytotoxic drugs and/or TRAIL. Chemicals and antibodies The Cytotoxic drugs cisplatin (CDDP), doxorubicin (DXR) and 5-fluorouracil (5-FU) were obtained from Roger Bellon Chemical Co. (Neuilly, France), Sigma-Aldrich and Dakota pharm (Gentilly, France), respectively. The caspase inhibitor z-IETD-fmk was from Calbiochem (San Diego, CA, USA) and the recombinant human soluble Flag-tagged TRAIL, Fasligand, TWEAK and TNFa from Alexis Biochemicals (Coger, Paris, France) (www.alexis-corp.com). Mouse anti-human procaspase-8 antibodies (Abs) were either obtained from Immunotech (Coulter, Marseille, France) or MBL (France Biochem, Meudon, France), anti-human procaspase-10 from MBL, anti-HSC70 from Santa Cruz (Tebu, Le Perray en Yvelines, France), anti-Flag (M2) from Sigma-Aldrich, antiprocaspase-3 from Santa Cruz and anti-FADD from Transduction Laboratories (Interchim, Montluçon, France). AntiBid Abs were obtained from R&D Systems (Abington, UK) and kindly provided by Dr X Wang (Howard Hughes Medical Institute, Dallas, TX, USA). The rabbit polyclonal anticaspase-9 Ab was from MBL, anti-TRAIL-R1 (antibody DR4CT) and anti-TRAIL-R2 (antibody DR5ID) from Alexis Biochemicals. The rat anti-human FLIP was purchased from


1815 Alexis Biochemicals. All other chemicals were of reagent grade and purchased from local sources. Analysis of mitochondrial membrane potential Mitochondrial transmembrane potential (DCm) was studied by using the DePsipher kit (R&D Systems). Briefly, 106 cells were incubated for 20 min at 371C in the presence of 5 mg/ml DePsipher solution before washing in PBS. Cells treated for 30 min with 100 mm of the uncoupling agent carbonyl cyanide m-chlorophenylhydrazone (Sigma-Aldrich) were used as a positive control. Analysis was performed by using a FACscan flow cytometer (Becton-Dickinson, Le Pont de Claix, France). DePsipher probe aggregates upon membrane polarization and forms a red fluorescent derivative. Mitochondrial potential disturbance prevents the dye to access the membrane space. Thus, a decrease in red fluorescence (FL-2) reflects a decrease in DCm. Cell fractionation Mitochondrial and cytosolic fraction for cytochrome c, AIF and SMAC release studies were prepared by resuspending the cells in ice-cold buffer A (250 mm sucrose, 20 mm HEPES, 10 mm KCl, 1.5 mm MgCl2, 1 mm EGTA, 1 mm EDTA, 1 mm DTT, complete protease inhibitor cocktail (Roche Biochemicals, Meylan, France)) before passing them through an icecold cylinder cell homogenizer. Unlysed cells and nuclei were pelleted via 10 min, 750 g spin. The supernatant was spun at 10 000 g for 25 min. The pellet was resuspended in buffer A and represents the mitochondrial fraction. The supernatant was spun at 10 000 g for 1 h. The final supernatant represents the cytosolic fraction.

Caspase activities were measured by monitoring fluorescence continuously in a dual luminescence fluorimeter (LS 50B Perkin-Elmer, France) using specific excitation and emission wavelength for each peptide. Enzyme activities were determined as initial velocities and expressed as relative intensity/ min/mg compared to the control. Confocal laser scanning microscopy analysis of cytochrome c release HT-29 cells were seeded in Lab-Tek chamber slide (Nunc S/A, Polylabo, Strasbourg, France), exposed to treatment, then fixed in 2% paraformaldehyde (Sigma) for 10 min, washed twice with PBS and permeabilized with PBS/0.5% BSA/0.1% saponin (PBS/BSA/saponin) for 10 min. Cells were incubated with a primary Abs diluted 1/100 in PBS/BSA/saponin for 2 h, washed twice in PBS, then incubated for 45 min with a Texas Red-conjugated secondary anti-mouse Ab (Jackson ImmunoResearch Laboratories) diluted 1/50 in PBS/BSA/saponin. Fluorescent labeling was analysed using a confocal laser scanning microscope (Leica, Centre de Microscopie Appliqueé a` la Biologie, Dijon). DISC analysis

Measurement of caspase activities

HT-29 cells pretreated by anticancer drugs were resuspended in 2 ml complete RPMI medium (5 107 cells/ml) at 371C, then stimulated for an indicated time with 500 ng/ml TRAIL and 2 mg/ml anti-Flag (M2) Ab in a final volume of 2 ml. Cotreatment with 100 mg/ml 5-FU (100 mg/ml) and TRAIL were performed in a final volume of 2 ml containing 2 mg/ml anti-Flag (M2) Ab at 371C. In some experiments (Figure 3a), 500 ng TRAIL and 2 mg M2 were added after lysis of unstimulated controls to immunoprecipitate total (membrane-bound and intracytoplasmic) nonstimulated TRAIL receptors. In other experiments (Figure 3c), M2 Ab was added without TRAIL after lysis of unstimulated controls to determine unspecific binding to protein G beads. The reaction was stopped by adding 10 ml of ice-cold PBS. The cells were immediately collected (400 g, 5 min, 41C), washed with 1 ml icecold PBS and lysed in 1 ml lysis buffer (20 mm Tris-HCl, pH 7.4, 150 mm NaCl, 0.2% Nonidet P40, 10% glycerol and complete protease inhibitor cocktail (Roche Biochemicals, Meylan, France)) for 15 min on ice. The lysate was cleared twice by centrifugation at 16 000 g for 10 min at 41C. The soluble fraction was precleared with 20 ml Sepharose-6B (Sigma-Aldrich) for 2 h at 41C and immunoprecipitated with 20 ml protein-G–Sepharose CL-4B (Amersham) at 41C. Beads were recovered by centrifugation and washed four times with 500 ml lysis buffer. Sample buffer was added, beads were boiled for 5 min and samples were analysed by SDS–PAGE and western blotting.

Cell lysates (50 mg) obtained in RIPA buffer (150 mm NaCl, 50 mm Tris-HCl pH 8.0, 0.1% SDS, 1% Nonidet P40, 0.5% sodium deoxycholate, 1 mm paramethylsulfonide, 1 mm benzamidine] were incubated for 30 min at 371C in a caspase assay buffer (100 mm HEPES, pH 7.0, 10% glycerol, 1 mm EDTA, 0.1% CHAPS, 1 mm dithiothreitol) containing 20 mm of either IETD-AFC, LEHD-AFC or DEVD-AMC (Calbiochem).

Acknowledgements This work was supported by grants from the Nie`vre committee (MTDB) and special grants (ES) from la Ligue Nationale Contre le Cancer. SL is supported by la Ligue Nationale Contre le Cancer.

Western blot analysis Cells were incubated for 10 min at 41C in lysis buffer (1% SDS, 10 mm Tris-HCL pH 7.4), then boiled for 5 min. Proteins (50 mg) were separated on a polyacrylamide sodium dodecylsulfate (SDS)-containing gel and transferred to a polyvinylidene difluoride membrane (PVDF) (Biorad, Ivry sur Seine, France). After blocking nonspecific binding sites overnight at 41C by 8% nonfat milk in TPBS (PBS with 0.1% Tween 20), membranes were incubated for 2 h at room temperature with specific Abs. Then, membranes were washed twice with TPBS, incubated for 1 h at room temperature with horseradish peroxidase-conjugated goat anti-mouse, anti-rat or anti-rabbit Abs (Jackson ImmunoResearch Laboratories, Beckman Coulter, Villepinte, France), washed twice with TPBS before revelation using an enhanced chemiluminescence detection kit (Amersham, Orsay, France).

References Ashkenazi A, Pai RC, Fong S, Leung S, Lawrence DA, Marsters SA, Blackie C, Chang L, McMurtrey AE, Hebert A, DeForge L, Koumenis IL, Lewis D, Harris L, Bussiere J,

Koeppen H, Shahrokh Z and Schwall RH. (1999). J. Clin. Invest., 104, 155–162. Baker SJ and Reddy EP. (1998). Oncogene, 17, 3261–3270. Oncogene


1816 Bisiau H, Daudignon A, Le Baron F, Pollet JP, Preudhomme C, Duthilleul P and Bastard C. (1995). Nouv. Rev. Fr. Hematol., 37, 241–244. Bodmer JL, Holler N, Reynard S, Vinciguerra P, Schneider P, Juo P, Blenis J and Tschopp J. (2000). Nat. Cell Biol., 2, 241–243. Buglioni S, D’Agnano I, Cosimelli M, Vasselli S, D’Angelo C, Tedesco M, Zupi G and Mottolese M. (1999). Int. J. Cancer, 84, 545–552. Burns TF and El-Deiry WS. (2001). J. Biol. Chem., 276, 37 879–37 886. Chapman PB, Lester TJ, Casper ES, Gabrilove JL, Wong GY, Kempin SJ, Gold PJ, Welt S, Warren RS, Starnes HF. (1987). J. Clin. Oncol., 5, 1942–1951. Chaudhary PM, Eby M, Jasmin A, Bookwalter A, Murray J and Hood L. (1997). Immunity, 7, 821–830. Deng Y, Lin Y and Wu X. (2002). Genes. Dev., 16, 33–45. Droin N, Dubrez L, Eymin B, Renvoize C, Breard J, Dimanche-Boitrel MT and Solary E. (1998). Oncogene, 16, 2885–2894. Fulda S, Meyer E and Debatin KM. (2002). Oncogene, 21, 2883–2894. Gibson SB, Oyer R, Spalding AC, Anderson SM and Johnson GL. (2000). Mol. Cell Biol., 20, 205–212. Golstein P. (1997). Curr. Biol., 7, R750–R753. Gorlick R and Bertino JR. (1999). Semin. Oncol., 26, 606–611. Gottlieb RA. (2001). Biol. Signals Recept., 10, 147–161. Hueber AO, Bernard AM, Herincs Z, Couzinet A and He HT. (2002). EMBO Rep., 3, 190–196. Ikonen E. (2001). Curr. Opin. Cell Biol., 13, 470–477. Keane MM, Ettenberg SA, Nau MM, Russell EK and Lipkowitz S. (1999). Cancer Res., 59, 734–741. Kim Y, Suh N, Sporn M and Reed JC. (2002). J. Biol. Chem., 277, 22 320–22 329 Kischkel FC, Lawrence DA, Tinel A, LeBlanc H, Virmani A, Schow P, Gazda A, Blenis J, Arnott D and Ashkenazi A. (2001). J. Biol. Chem., 276, 46639–46646. Kroemer G and Reed JC. (2000). Nat Med., 6, 513–519. Lacour S, Hammann A, Wotawa A, Corcos L, Solary E and Dimanche-Boitrel MT. (2001). Cancer Res., 61, 1645–1651. Li H, Zhu H, Xu CJ and Yuan J. (1998). Cell, 94, 491–501. Martin DA, Siegel RM, Zheng L and Lenardo MJ. (1998). J. Biol. Chem., 273, 4345–4349. Micheau O, Hammann A, Solary E and Dimanche-Boitrel MT. (1999a). Biochem. Biophys. Res. Commun., 256, 603–607. Micheau O, Lens S, Gaide O, Alevizopoulos K and Tschopp J. (2001). Mol. Cell. Biol., 21, 5299–5305. Micheau O, Solary E, Hammann A and Dimanche-Boitrel MT. (1999b). J. Biol. Chem., 274, 7987–7992.

Oncogene

Micheau O, Solary E, Hammann A, Martin F and DimancheBoitrel MT. (1997). J. Natl. Cancer Inst., 89, 783–789. Mizutani Y, Yoshida O, Miki T and Bonavida B. (1999). Clin. Cancer Res., 5, 2605–2612. Munshi A, Pappas G, Honda T, McDonnell TJ, Younes A, Li Y and Meyn RE. (2001). Oncogene, 20, 3757–3765. Nagane M, Huang HJ and Cavenee WK. (2001). Apoptosis, 6, 191–197. Pitti RM, Marsters SA, Ruppert S, Donahue CJ, Moore A and Ashkenazi A. (1996). J. Biol. Chem., 271, 12 687–12 690. Rohn TA, Wagenknecht B, Roth W, Naumann U, Gulbins E, Krammer PH, Walczak H and Weller M. (2001). Oncogene, 20, 4128–4137. Scaffidi C, Fulda S, Srinivasan A, Friesen C, Li F, Tomaselli KJ, Debatin KM, Krammer PH and Peter ME. (1998). EMBO J., 17, 1675–1687. Schneider P, Thome M, Burns K, Bodmer J, Hofmann K, Kataoka T, Holler N and Tschopp J. (1997). Immunity, 7, 831–836. Seol DW, Li J, Seol MH, Park SY, Talanian RV and Billiar TR. (2001). Cancer Res., 61, 1138–1143. Sheikh MS, Burns TF, Huang Y, Wu GS, Amundson S, Brooks KS, Fornace AJ Jr and el-Deiry WS. (1998). Cancer Res., 58, 1593–1598. Sheridan JP, Marsters SA, Pitti RM, Gurney A, Skubatch M, Baldwin D, Ramakrishnan L, Gray CL, Baker K, Wood WI, Goddard AD, Godowski P and Ashkenazi A. (1997). Science, 277, 818–821. Suliman A, Lam A, Datta R and Srivastava RK. (2001). Oncogene, 20, 2122–2133. Sun SY, Yue P, Zhou JY, Wang Y, Choi Kim HR, Lotan R and Wu GS. (2001). Biochem. Biophys. Res. Commun., 280, 788–797. Tanaka M, Suda T, Yatomi T, Nakamura N and Nagata S. (1997). J Immunol., 158, 2303–2309. Walczak H, Miller RE, Ariail K, Gliniak B, Griffith TS, Kubin M, Chin W, Jones J, Woodward A, Le T, Smith C, Smolak P, Goodwin RG, Rauch CT, Schuh JC and Lynch DH. (1999). Nat. Med., 5, 157–163. Wiley SR, Schooley K, Smolak PJ, Din WS, Huang CP, Nicholl JK, Sutherland GR, Smith TD, Rauch C, Smith CA. (1995). Immunity, 3, 673–682. Yamada H, Tada-Oikawa S, Uchida A and Kawanishi S. (1999). Biochem. Biophys. Res. Commun., 265, 130–133. Yamanaka T, Shiraki K, Sugimoto K, Ito T, Fujikawa K, Ito M, Takase K, Moriyama M, Nakano T and Suzuki A. (2000). Hepatology, 32, 482–490. Zhang XD, Franco A, Myers K, Gray C, Nguyen T and Hersey P. (1999). Cancer Res., 59, 2747–2753.