Induction of tumor cell resistance to therapeutics has been a major obstacle in cancer therapy. Targeting of the death receptors by a natural ligand, tumor necrosis factor–related apoptosis-inducing ligand (TRAIL), or agonistic monoclonal antibodies against TRAIL receptor 1 (TRAIL-R1) or TRAIL receptor 2 (TRAIL-R2) has been thought to be a promising cancer therapy. To determine whether tumor cells are able to generate a resistance to apoptosis induced by an anti-TRAIL-R2 antibody, TRA-8, we examined the apoptotic response of human breast and ovarian cancer cell lines after treatment with TRA-8. Our results show that tumor cell resistance to TRA-8 can be induced by repeated treatment of tumor cells with low, non-apoptosis-inducing doses of TRA-8. Interestingly, the induced resistance to apoptosis was not due to a global apoptotic defect in tumor cells but rather a selective defect in the TRAIL-R2 signaling pathway. Whereas TRA-8-treated tumor cells developed a selective resistance to TRAIL-R2-mediated apoptosis, the apoptotic responses induced by TRAIL, an anti-TRAIL-R1 antibody (2E12), and other apoptotic stimuli were not impaired. The expression levels of cell surface TRAIL-R2 were not altered and mutations of TRAIL-R2 were not found in the resistant cells. The induced TRA-8 resistance was due to a selective blockade at the level of the death domain and could be reversed by a wide array of chemotherapeutic agents. Proteomic analysis of death-inducing signaling complex formation during TRA-8 treatment shows that the translocation of TRAIL-R2-associated apoptotic proteins was significantly altered. Our results suggest that the prevention of tumor cell resistance to therapeutic agents that target the death receptors must be taken into consideration. (Cancer Res 2006; 66(17): 8520-8)
- signal transduction and cancer cells
Tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) selectively induces apoptosis of tumor cells ( 1) and has been proposed as a potential anticancer therapy ( 2– 5). At least five receptors for TRAIL have been identified: TRAIL receptor 1 (TRAIL-R1; DR4; ref. 6), TRAIL receptor 2 (TRAIL-R2; DR5; ref. 7), TRAIL receptor 3 (DcR1; ref. 8), TRAIL receptor 4 (DcR2; ref. 9), and OPG ( 10). Among these receptors, TRAIL-R1 and TRAIL-R2 are the agonistic receptors responsible for inducing cell death, whereas others serve as decoy receptors that antagonize TRAIL-mediated apoptosis. We have developed previously agonistic antibodies specific for TRAIL-R1 (2E12; ref. 11) and TRAIL-R2 (TRA-8; ref. 12), which strongly induce apoptosis in tumor cells both in vivo and in vitro. Although TRAIL and agonistic anti–death receptor antibodies are moving to clinical development as anticancer therapies, both spontaneous and inducible resistance to these therapeutic agents is a concern. It is well known that many tumor cells that express high levels of cell surface TRAIL-R2 may not be susceptible to TRAIL-mediated or death receptor–mediated apoptosis ( 12, 13). However, it is still unclear whether tumor cells are also able to develop an inducible resistance to death receptor–mediated apoptosis. In addition, there is accumulating evidence indicating that combination therapy of the anti–death receptor agents and chemotherapy agents synergistically enhances therapeutic efficacy compared with single-agent therapy ( 3, 4, 14– 19), suggesting that chemotherapy may be used to overcome or reverse the resistance of tumor cells to anti–death receptor therapy. Here, we report that previously sensitive tumor cells treated with low and repeated doses of TRA-8 become completely resistant to TRA-8-mediated apoptosis. The induced resistance is selective for TRAIL-R2 and can be reversed by combination treatment with most chemotherapeutic agents. Analysis of TRAIL-R2 signal transduction indicates that there is a selective blockade of apoptotic signaling at the level of the death domain of TRAIL-R2 by altered translocation of TRAIL-R2-associated proteins.
Materials and Methods
Cell lines, antibodies, and reagents. The human breast cancer cell line MDA-MB-231 was purchased from the American Tissue Culture Collection (Manassas, VA). The human ovarian cancer cell line UL-3C was kindly provided by Dr. Taylor (University of Louisville). Cells were maintained in DMEM or RPMI 1640 supplemented with 10% heat-inactivated FCS, 50 μg/mL streptomycin, and 50 units/mL penicillin (Cellgro, Mediatech, Inc., Herndon, VA). Anti-human TRAIL-R1 (clone 2E12) and anti-human TRAIL-R2 (clone TRA-8) monoclonal antibodies were described previously ( 11, 12). Anti-human TRAIL-R2 (clone 2B9) for flow cytometry and immunoprecipitation assays was developed in our laboratory. Recombinant soluble TRAIL was purchased from Alexis Biochemicals (San Diego, CA). Polyclonal anti-caspase-3 and anti-caspase-8 antibodies were purchased from BD PharMingen (San Diego, CA). Polyclonal anti-phosphorylated stress-activated protein kinase/c-Jun NH2-terminal kinase (JNK; Thr183/Tyr185), anti-phosphorylated p38 mitogen-activated protein kinase (Thr180/Tyr182), and anti–poly(ADP-ribose) polymerase (PARP) antibodies were purchased from Cell Signaling Technology, Inc. (Beverly, MA). Anti-β-actin antibody was purchased from Sigma (St. Louis, MO). Anti-FADD was purchased from Transduction Laboratories (Lexington, KY). Anti-FLIP was purchased from ProSci, Inc. (Poway, CA). All horseradish peroxidase (HRP)–conjugated secondary reagents were purchased from Southern Biotechnology Associates, Inc. (Birmingham, AL). The fluorogenic peptide derivatives Ac-carbonyl-Ile-Glu-Thr-Asp-7-amido-4-methylcoumarin were purchased from Alexis Biochemicals. Caspase-8 inhibitor (FMKSP01) were purchased from R&D Systems, Inc. (Minneapolis, MN).
Flow cytometric analysis of cell surface expression of TRAIL-R1 and TRAIL-R2. Cells (106) were incubated with 1 μg/mL biotinylated 2E12 and 1 μg/mL phycoerythrin (PE)–conjugated 2B9 on ice for 30 minutes. After washing twice with fluorescence-activated cell sorting buffer (PBS with 5% fetal bovine serum and 0.01%NaN3), cells were incubated with streptavidin-CyChrome. Viable cells (10,000) were analyzed by FACScan flow cytometer (Becton-Dickinson FACScan, San Jose, CA).
Cytotoxicity analysis of tumor cell susceptibility to TRA-8-, 2E12-, and TRAIL-mediated apoptosis. Cells (1,000 per well) were seeded into 96-well plate in triplicate with eight concentrations (2-fold serial dilutions from 1,000 ng/mL) of TRA-8, 2E12, or TRAIL. Cell viability was determined after overnight culture using ATPLite assay according to the manufacturer's instructions (Packard Instruments, Meriden, CT). The results are presented as the percentage of viable cells in treated wells compared with medium control wells.
Induction of tumor cell resistance to TRAIL-R2. Cells (5 × 105/mL) were incubated with a starting dose of 1 ng/mL TRA-8 for 2 days to induce resistance. All of the treated cells were split with fresh medium every 2 days, and TRA-8 doses were doubled until the dose reached 2,000 ng/mL within 5 weeks. All cells were finally treated with 2,000 ng/mL TRA-8 for 2 days, and TRA-8 was withdrawn. After induction of resistance, cells might be cultured in the absence of TRA-8 for 8 to 10 weeks without losing the resistance. Typically, the experiments were done 4 to 8 weeks after withdraw of TRA-8. During this period, cells might be frozen down for the future use. The results indicate that the resistant status of the treated cells remained for at least 10 weeks without TRA-8 treatment. To monitor the development of resistance, at each cycle of TRA-8 treatment, cell viability of control parental and treated cells was determined by ATPLite assay.
Cloning and sequencing of TRAIL-R2. The full-length cDNA of TRAIL-R2 was obtained by PCR using the Platinum Taq DNA Polymerase High Fidelity (Invitrogen, Carlsbad, CA). Resulting cDNA was cloned into the pCR2.1-TOPO vector (Invitrogen). At least five independent clones were selected for sequencing at the DNA Sequencing Facility of the University of Alabama at Birmingham.
Western blot analysis of apoptosis-associated proteins. Tumor cells (3 × 106) were washed twice with cold PBS and lysed with 300 μL lysis buffer containing 10 mmol/L Tris-HCl (pH 7.6), 150 mmol/L NaCl, 0.5 mmol/L EDTA, 1 mmol/L EGTA, 0.1% SDS, 1 mmol/L sodium orthovanadate, and a mixture of protease inhibitors [1 mmol/L phenylmethylsufonyl fluoride (PMSF), 1 μg/mL pepstatin A, and 2 μg/mL aprotinin]. Lysates were sonicated for 10 seconds and then centrifuged for 20 minutes at 12,000 × g. Equal amounts of total protein from each cell lysate were boiled for 5 minutes in SDS-PAGE sample buffer. Total cell lysates were separated on 8%, 10%, or 12% SDS-PAGE gels and electrophoretically transferred to nitrocellulose membranes. The blots were blocked with 5% nonfat dry milk in TBST [20 mmol/L Tris-HCl (pH 7.4), 500 mmol/L NaCl, 0.1% Tween 20] and incubated with primary antibody in blocking buffer at 4°C overnight. The blots were then washed thrice with TBST and probed with HRP-conjugated secondary antibodies for 1 hour at room temperature. After washing four times with TBST, the probed proteins were visualized using the Enhanced Chemiluminescence (ECL) Western blotting detection system (Amersham Biosciences, Piscataway, NJ) according to the manufacturer's instructions.
cDNA array analysis of transcriptional regulation of apoptosis-associated and cell signaling–associated genes. The Human Apoptosis Gene Array (HS-002) and the Human Signal Transduction PathwayFinder Gene Array (HS-008) were purchased from SuperArray, Inc. (Frederick, MD). Total RNA was extracted from cells using the TRIzol protocol (Invitrogen). cDNA probes were synthesized with [32P]dCTP. The cDNAs on the membrane blots were hybridized with the [32P]dCTP-labeled probes at 60°C overnight. The gene expression profiles were analyzed using the Cyclone PhosphorImager.
Coimmunoprecipitation of TRAIL-R1 and TRAIL-R2. Cells (107) were washed with ice-cold PBS and lysed for 15 minutes on ice with lysis buffer [1% Triton X-100, 150 mmol/L NaCl, 10% glycerol, 20 mmol/L Tris-HCl (pH 7.5), 2 mmol/L EDTA, 0.57 mmol/L PMSF, and a protease inhibitor cocktail]. The lysates were then cleared twice by centrifugation at 16,000 × g for 10 minutes at 4°C. The soluble fraction was incubated with either 30 μL TRA-8- or 2B4-conjugated Sepharose 4B at 4°C overnight. After seven washes with lysis buffer and three washes with 10 mmol/L Tris, the bound proteins were eluted by boiling for 3 minutes in SDS-PAGE loading buffer and separated using SDS-PAGE. The presence of caspase-8 and FADD was then determined by Western blot analysis.
Assay of caspase inhibitory activity of TRAIL-R2-associated protein complex. Fluorometric assays were conducted in 96-well clear-bottom plates, and all measurements were carried out in triplicate wells. Assay buffer [100 μL; 10 mmol/L HEPES (pH 7.0), 50 mmol/L NaCl, 2 mmol/L MgCl2, 5 mmol/L EDTA, 1 mmol/L DTT] was added. Active caspase-8 and peptide substrates (Ac-IETD-AMC) were added to each well to a final concentration of 100 ng/μL. The eluted fractions of TRAIL-R2 coimmunoprecipitation of parental sensitive and resistant cells were added into the reaction mix of caspase-8 and its substrate. Assay plates were incubated at 37°C for 1 hour. Fluorescence was measured on a fluorescence plate reader (Bio-Tek, Winooski, VT) set at 355 nm excitation and 440 nm emission. The inhibition of caspase-8 activity by TRAIL-R2 protein complex is presented as a percentage of maximum caspase-8 activity in the absence of TRAIL-R2 protein complex.
Two-dimensional PAGE. After coimmunoprecipitation with 2B4-conjugated Sepharose 4B, the proteins were eluted and desalted with acetone and reconstituted in the IEF sample buffer (Bio-Rad, Hercules, CA). Total protein (160 μg) was loaded on IPG strips (Bio-Rad) at room temperature overnight and then run through isoelectric focusing in the PROTEAN IEF Cell (Bio-Rad). The protein strips were then equilibrated with the ReadyPrep Equilibration Buffers (Bio-Rad) and run on 10% SDS-PAGE gels. On completion, the gels were fixed with a buffer containing 10% methanol and 7% acetic acid and stained with SYPRO Ruby Staining buffer (Bio-Rad). The gels were imaged using the VersaDoc Digital Imaging System (Bio-Rad) and analyzed using PDQuest (Bio-Rad). Mass spectrometry analysis of the spots of interest was done at the Mass Spectrometry and Proteomics Core Facility of University of Alabama at Birmingham. After two-dimensional PAGE, the proteins might be transferred onto nitrocellulose membranes and probed with anti-DDX3 antibody.
Generation of monoclonal antibodies against apoptosis-associated proteins. Monoclonal anti-human caspase-2 and caspase-8 antibodies and monoclonal anti-human Bcl-2, Bcl-xL, Bax, cellular inhibitor of apoptosis (cIAP) 1, cIAP2, X-linked IAP (XIAP), and survivin antibodies were prepared in our laboratory. The full-length recombinant human Bcl-2 and Bcl-xL proteins, the NH2-terminal portion of human Bax (amino acids 1-120), the full-length human caspase-2 and caspase-8, the full-length human cIAP1, cIAP2, XIAP, and survivin, and the full-length human DDX3 were expressed in Escherichia coli and purified with nickel column with a 6-histine tag. BALB/c mice were immunized with each purified protein. After five immunizations, the lymph node cells were fused with NS1 myeloma cells. The primary screenings were done for selection of positive clones to each immunized antigen by ELISA. The secondary screenings were done to rule out the cross-reactivity among the members of a protein family using ELISA. After establishment of the specificity of each positive clone, the screenings were further done to confirm their utility and specificity in Western blot analysis. The characteristics of these monoclonal antibodies are summarized in Table 1 . Monoclonal anti-human caspase-2 (clone 2A3) and anti-human caspase-8 (clone 5G7) were selected for Western blot as they recognize both proforms and activated forms of caspase-2 and caspase-8, respectively. 3E2 anti-DDX3 antibody recognizes an epitope at the NH2-terminal portion (amino acids 1-100) of DDX3, whereas 3E4 anti-DDX3 antibody recognizes an epitope of DDX3 between amino acids 100 and 150.
Induction of selective resistance to TRAIL-R2-mediated apoptosis. A human breast cancer cell line, MDA-MB-231, and a human ovarian cancer cell line, UL-3C, were selected for induction of TRAIL-R2 resistance because they coexpress high levels of cell surface TRAIL-R1 and TRAIL-R2 as determined by two-color flow cytometry analysis using anti-TRAIL-R1 (2E12) and anti-TRAIL-R2 (2B9) antibodies ( Fig. 1A ). The two tumor cell lines were susceptible to apoptosis induced by agonistic anti-TRAIL receptor antibodies, 2E12 and TRA-8, as well as TRAIL itself as determined by in vitro cytotoxicity assays ( Fig. 1B), indicating that both receptors for TRAIL are functional in the two tumor cell lines. To determine whether resistance to apoptosis caused by TRA-8 could be induced in these tumor cell lines, cells were treated initially with a nonapoptotic dose (1 ng/mL) of TRA-8 for 2 days. Treated cells and control cells were then given doses doubling every 2 days to a final dose of 2,000 ng/mL before the withdrawal of TRA-8. Cell viability of both induced and untreated control cells was measured at each dosage. At <10 ng/mL TRA-8, there was no significant cell death in either treated or untreated cells. When TRA-8 doses were increased to ≥50 ng/mL, a dose-dependent reduction of cell viability was observed in control cells. In contrast, there was no significant cell death observed in treated cells at doses of TRA-8 up to 2,000 ng/mL ( Fig. 1C). These results indicate that the repeated treatment of tumor cells with low, nonapoptotic doses of TRA-8 may induce resistance to apoptosis and that the induced resistance is not due to a selective process removing apoptosis-sensitive cells.
Four weeks after the withdrawal of TRA-8, both parental cells (MDA-MB-231P and UL-3CP) and treated cells (MDA-MB-231R and UL-3CR) were tested for their susceptibility to apoptosis induced by TRA-8, 2E12, or TRAIL. Compared with ∼100% cell death of the parental cells after treatment with 1,000 ng/mL TRA-8, no significant cell death was induced in either MDA-MB-231R or UL-3CR cells over a range of concentrations of TRA-8 ( Fig. 2A ), indicating that the cells continued to be highly resistant to TRA-8-induced apoptosis. In contrast, the susceptibility of TRA-8-resistant tumor cells to 2E12-induced apoptosis remained unchanged ( Fig. 2B). The induced TRA-8-resistant cells were still susceptible to TRAIL-mediated apoptosis, albeit less so than parental cells ( Fig. 2C). These results indicate that TRA-8-induced apoptosis resistance is selective for TRAIL-R2. After withdrawal of TRA-8, the cells remained resistant to TRA-8 for at least 3 months, and sensitivity to TRA-8 was slowly restored to ∼30% of parental cell levels by 4 months ( Fig. 2D), indicating that the induced resistance to TRA-8 was long-lasting but partially reversible.
Induced TRAIL-R2 resistance is not due to altered cell surface expression, mutations of TRAIL-R2, or an intrinsic apoptotic defect. Because TRA-8-induced resistance to apoptosis was selective for TRAIL-R2, it is possible that expression of TRAIL-R2 might be selectively reduced or a mutation of TRAIL-R2 might occur after induction of TRA-8 resistance. To rule out these possibilities, we examined cell surface expression of TRAIL-R2 and found that there was no alteration in expression levels of TRAIL-R2 in TRA-8-resistant cells compared with their parental cells ( Fig. 3A ). This conclusion was further confirmed by Western blot analysis showing that the two isoforms of TRAIL-R2 were equally expressed in parental and resistant cells ( Fig. 3B). Full-length cDNA clones of TRAIL-R2 were isolated from both TRA-8-resistant cell lines and were sequenced, and no mutations were identified (data not shown). These results indicate that the TRA-8 resistance is not due to alterations of TRAIL-R2 itself.
Because TRAIL-R2-mediated apoptosis may be regulated by several proteins, such as the IAP family ( 20– 25) and the Bcl-2 family ( 26– 31), we examined the expression levels of cIAP1, cIAP2, XIAP, survivin, Bcl-2, Bcl-xL, and Bax by Western blot analysis using a panel of newly developed monoclonal antibodies. Although MDA-MB-231 and UL-3C expressed variable levels of these proteins, there was no significant difference between parental and resistant cells ( Fig. 3B), indicating that the expression levels of these proteins are unlikely to be involved in the induction of TRA-8 resistance. A more broad screening for potential transcriptional alterations among a panel of apoptosis-associated and cell signaling–associated genes was done using membrane cDNA arrays (SuperArray), which tested for >200 well-known apoptosis-related genes ( Fig. 3C, top) and cell signaling genes ( Fig. 3C, bottom). Comparison between MDA-MB-231 parental and resistant cells indicates that there was no significant alteration in the expression profile of these genes after induction of TRA-8 resistance.
Selective blockade of TRAIL-R2 signal transduction in TRA-8-resistant tumor cells. Sequential activation of caspase-8 and caspase-3 are key events in TRAIL-R2 signal transduction. Thus, we examined time-dependent activation of these two caspases. As shown previously, the treatment of parental MDA-MB-231 cells with TRA-8 induced activation of caspase-8 ( Fig. 4A, top ) and caspase-3 ( Fig. 4A, middle) as shown by the generation of cleaved fragments of the caspases after TRA-8 treatment. PARP, a substrate of caspases used to monitor their activation, was also quickly cleaved ( Fig. 4A, bottom). However, activation of caspase-8 and caspase-3 and subsequent cleavage of PARP did not occur in the resistant cells after TRA-8 treatment. This failure of caspase activation is not due to an intrinsic defect in the caspase pathways, as caspase activation triggered by 2E12-mediated activation of TRAIL-R1 was not impaired in TRA-8-resistant cells ( Fig. 4A, left). These results suggest that the TRAIL-R2-associated caspase cascade is selectively blocked at the level of the upstream caspase-8 after induction of TRA-8 resistance.
Previous studies in our laboratory showed that caspase-8-dependent activation of the JNK/p38 kinase pathway plays a critical synergistic role in TRAIL-R2-mediated apoptosis ( 32, 33). The activation of the JNK/p38 kinases was measured by Western blot analysis of their phosphorylation states during TRA-8 treatment. Correspondent to caspase-8 activation, JNK ( Fig. 4B, top) and p38 ( Fig. 4B, bottom) were quickly phosphorylated in a time-dependent fashion. However, in TRA-8-resistant cells, only 2E12, and not TRA-8, was able to induce phosphorylation of the JNK/p38, indicating that the JNK/p38 kinase pathways are selectively inhibited in TRA-8-resistant cells.
Selective blockade of TRAIL-R2 death domain function in TRA-8-resistant tumor cells. Because the caspase-8-initiated caspase cascade is selectively blocked at the level of TRAIL-R2 in TRA-8-resistant cells, we suspected that a blockade that prevents the recruitment and formation of the death-inducing signaling complex (DISC) might be generated at the death domain of TRAIL-R2 after induction of TRA-8 resistance. As FADD and caspase-8 are recruited to the death domain of TRAIL-R2 and are major DISC components, we examined the capability of forming a DISC at TRAIL-R1 and TRAIL-R2 in both parental and resistant cells using coimmunoprecipitation assays. In MDA-MB-231 parental cells treated with TRA-8 or 2E12, there was a time-dependent increase of FADD ( Fig. 5A, top ) and caspase-8 ( Fig. 5A, middle), which were coimmunoprecipitated with TRAIL-R2 ( Fig. 5A, left) or TRAIL-R1 ( Fig. 5A, right), respectively. In TRA-8-resistant cells, there was no FADD or caspase-8 coimmunoprecipitated with TRAIL-R2 during TRA-8-mediated apoptosis, but coimmunoprecipitation of FADD and caspase-8 with TRAIL-R1 after 2E12 treatment was not affected. Furthermore, to determine whether cFLIP, a competitive inhibitor of caspase-8 recruitment to the death domain, plays a role in the blockade of DISC formation, coimmunoprecipitation of cFLIP with TRAIL-R1 and TRAIL-2 was also examined. In a similar time-dependent pattern, cFLIP was coimmunoprecipitated with TRAIL-R2 during TRA-8-mediated apoptosis in the parental cells but not in the resistant cells ( Fig. 5A, bottom). The coimmunoprecipitation of cFLIP with TRAIL-R1 during 2E12-induced apoptosis in parental and TRA-8-resistant cells was not different. Because there were similar levels of total protein expression of FADD, caspase-8, and cFLIP, the failure of the recruitment of these death domain–associated proteins is not due to defective expression of these proteins. These results indicate that the induced TRA-8 resistance is likely due to a selective defect in the ability of TRAIL-R2 to recruit FADD and caspase-8 to begin DISC formation after TRA-8 treatment.
Failure of DISC assembly at the death domain of TRAIL-R2 in TRA-8-resistant cells suggests that the function and composition of TRAIL-R2 protein complex might be altered. For instance, a newly generated or functionally altered protein might associate with TRAIL-R2 and prevent the recruitment of FADD and caspase-8 to the death domain of TRAIL-R2. First, to determine whether there is a functional alteration, we examined the effect of the TRAIL-R2-associated protein complex isolated from parental and resistant cells on caspase-8 activity. The TRAIL-R2-associated protein complex was isolated from TRA-8-sensitive parental and resistant cells by TRAIL-R2 coimmunoprecipitation with 2B9 anti-TRAIL-R2 monoclonal antibody and was incubated with recombinant, active form of human caspase-8 in the presence of a caspase-8 fluorescent substrate. The TRAIL-R2-assocaited protein complex from the resistant MDA-MB-231 and UL-3C cells exhibited a higher degree of inhibitory activity against caspase-8 than that from the corresponding parental cells ( Fig. 5B). These results suggest that the properties of the TRAIL-R2-associated protein complex might be altered after induction of resistance. To further test this hypothesis, we compared the proteomic profiles of TRAIL-R2-associated proteins in TRA-8-sensitive parental and TRA-8-resistant MDA-MB-231 cells before and after TRA-8 treatment using two dimensional electrophoresis and mass spectrometry. Two-dimensional gels of TRAIL-R2-associated proteins were analyzed using PDQuest, which led us to focus on three protein spots that were altered between parental and resistant cells following TRA-8 treatment ( Fig. 5C). Spots 1 and 2, representing proteins with molecular weights of 50 and 20 kDa, respectively, appeared only in MDA-MB-231 parental cells after TRA-8 treatment but not in untreated cells or TRA-8-treated resistant cells, suggesting that these proteins might be recruited to the TRAIL-R2 during TRA-8-mediated apoptosis. Based on their molecular weights and isoelectric points, spot 1 was confirmed as caspase-8 ( Fig. 5D) and spot 2 as FADD ( Fig. 5E) by Western blot. The proteins in spot 3 were interesting because they were constantly associated with TRAIL-R2 but shifted from a higher molecular weight protein to a lower molecular weight protein during TRA-8-mediated apoptosis ( Fig. 5F). This conversion seemed to be relevant to the induced TRA-8 resistance, as it was only observed in TRA-8-treated MDA-MB-231 parental cells but not in resistant cells. Furthermore, mass spectrometry analysis identified both spots as being derived from DDX3, a member of the DEAD-box RNA helicase family. This was further confirmed by Western blot analysis of two-dimensional PAGE with DDX3-specific monoclonal antibody, which reveals a molecular weight shift of DDX3 during TRA-8-mediated apoptosis of TRA-8-sensitive parental cells ( Fig. 5G). Because a higher molecular weight form of DDX3 constantly associated with TRAIL-R2 in TRA-8-resistant cells, it might be a factor that prevents the recruitment of FADD and caspase-8 to the death domain of TRAIL-R2. Therefore, it would be interesting to further determine the role of TRAIL-R2-associated DDX3 in the induction of TRA-8 resistance.
Reversal of TRA-8 resistance by chemotherapeutic agents. Previous studies in our laboratory showed that chemotherapeutic agents synergistically enhance TRA-8-mediated apoptosis both in vitro and in vivo ( 32– 34), particularly in TRA-8-resistant cells. To determine whether chemotherapeutic agents are able to reverse induced TRA-8 resistance, we examined the effect of a group of chemotherapeutic agents (Adriamycin, Taxol, Cisplatin, and Bisindolylmaleimide VIII) on TRA-8-mediated apoptosis of cells with induced resistance. In the presence of indicated concentrations of the chemotherapeutic agents, a TRA-8 dose-dependent response was restored in both TRA-8-resistant MDA-MB-231 and UL-3C cells ( Fig. 6A ), indicating that a wide range of chemotherapeutic agents are able to reverse TRA-8-induced resistance.
Because the death receptors and chemotherapeutic agents may trigger different caspase pathways to lead to tumor cell apoptosis, we raised a question as to whether chemotherapeutic agents trigger an alternative caspase pathway that compensates disabled caspase-8 pathway in the resistant cells. Thus, we examined the activation of two major initiator caspases, caspase-2 and caspase-8, in MDA-MB-231-resistant cells after combination treatment with Adriamycin and TRA-8. Caspase-2 has been shown previously to be a major initiator caspase in cytotoxic stress-induced apoptosis ( 35). Treatment of TRA-8-sensitive parental cells (MDA-MB-231P) with TRA-8 alone or TRA-8 plus Adriamycin led to activation of both caspase-2 and caspase-8 ( Fig. 6B, left). In contrast, treatment of TRA-8-resistant cells (MDA-MB-231R) with TRA-8 alone did not induce activation of neither caspase-2 nor caspase-8 ( Fig. 6B, right). However, treatment of TRA-8-resistant cells with Adriamycin alone or TRA-8 plus Adriamycin resulted in caspase-2 activation. In the presence of Adriamycin, TRA-8 was able to trigger caspase-8 activation in the resistant cells. These results suggest that Adriamycin-induced caspase-2 activation plays a critical role in restoration of TRAIL-R2 apoptosis signal transduction. As we have identified DDX3 as a critical candidate in generation of a blockade at the death domain, we further examined the TRAIL-R2-associated DDX3 during the combination treatment. Correlated with activation of caspase-2 and caspase-8 in TRA-8-sensitive parental cells, the amount of the TRAIL-R2-associated DDX3 was reduced after treatment with TRA-8 alone or TRA-8 plus Adriamycin as shown by an anti-DDX3 antibody (3E2), which recognizes the NH2-terminal portion of DDX3. Furthermore, with another anti-DDX3 antibody (3E4), which recognizes the COOH-terminal portion of DDX3, the cleavage of DDX3 was revealed ( Fig. 6C, left). Interestingly, the cleavage of DDX3 was well correlated with the recruitment of FADD to the death domain after treatment of TRA-8-sensitive parental cells with TRA-8 alone or TRA-8 plus Adriamycin. In TRA-8-resistant cells, however, the profile was changed. TRA-8 alone was unable to activate caspase-2 and caspase-8 ( Fig. 6B, right); therefore, there was no corresponding cleavage of DDX3 ( Fig. 6C, right). Adriamycin alone was able to trigger activation of caspase-2 and cleavage of DDX3, but in the absence of TRA-8, there was no FADD recruitment. In the presence of TRA-8, however, FADD recruitment occurred. These results suggest that chemotherapeutic agent induced caspase-2 activation and the following cleavage of DDX3 is a key event in the restoration of the DISC formation.
Chemoresistance of tumor cells to therapeutic agents has been a major obstacle in the development of more effective anticancer therapies. Although TRA-8 is a strong agonistic anti-TRAIL-R2 antibody and exhibits excellent anticancer efficacy in preclinical human tumor models ( 12, 34), both spontaneous and induced apoptosis resistance to TRA-8 is still a concern. Our results show that following treatment with low and repeated doses of TRA-8, tumor cells quickly develop an inducible and selective resistance to subsequent treatment with TRA-8. Thus, the prevention and reversal of resistance to death receptor–mediated apoptosis is an important consideration in future clinical trials of anti–death receptor strategies. These results suggest that the doses and treatment schedules must be carefully considered to prevent the development of resistance. Frequent, suboptimal doses may impede subsequent treatment. We show here that the addition of other anticancer drugs might be necessary to prevent the development of resistance to anti–death receptor therapy.
It is well known that the expression levels of cell surface death receptors are not a crucial determinant for death receptor–mediated apoptosis and that the intracellular signaling mechanism is likely to play a key role in the susceptibility of tumor cells to death receptor–mediated apoptosis. The proteins in the Bcl-2 family have been shown to be critical controllers for TRAIL-mediated apoptosis. Selective survival of tumor cells with Bax mutations after treatment with TRAIL was reported ( 36). Proteins of the IAP family have also been proposed to regulate TRAIL-mediated apoptosis. It has been shown that inhibition of the apoptosis-inhibitory function of the IAPs in tumor cells significantly enhances TRAIL-mediated apoptosis ( 24, 25). Many studies have indicated that cFLIP plays an important role in tumor cells resistance to TRAIL. However, our results suggest that cFLIP is not critically involved in TRA-8-induced resistance. Our interpretation is that the apoptosis signal inhibition by cFLIP is downstream of that mediated by DDX3. cFLIP, as a competitive inhibitor of caspase-8, may be recruited to the death domain through its DED domain. The ratio of caspase-8 and cFLIP in the DISC may determine the apoptosis signal outcome. Thus, the recruitment of cFLIP to the death domain is a sign of DISC assembly. In our resistant model, because DDX3-mediated inhibition occurs at an upstream level of DISC assembly, the absence of cFLIP in DISC of resistant cells represents a failure of DISC assembly. Unlike other models of death receptor resistance, which result from an intrinsic apoptotic defect, the highly selective resistance to apoptosis related to TRAIL-R2 in our induced resistance model does not involve a deficiency in the components of the apoptotic mechanism. In fact, total cellular expression of the Bcl-2 and IAP proteins is not altered. Instead, the defect is a functional one, located at the death domain of TRAIL-R2. It is likely that functional alterations of the TRAIL-R2 protein complex may prevent the initial recruitment of FADD and activation of caspase-8. This is supported by the increased caspase-8 inhibitory activity of the TRAIL-R2-associated protein complex in TRA-8-resistant cells ( Fig. 5B) and by the altered proteomic profile of the TRAIL-R2 protein complex during TRA-8-mediated apoptosis between sensitive parental cells and cells with induced resistance ( Fig. 5C). In addition to the failure of FADD and caspase-8 recruitment, a newly identified TRAIL-R2-associated protein, DDX3, was found to be constantly associated with TRAIL-R2. This suggests that TRAIL-R2 signal transduction not only involves the recruitment of certain proteins but also the disassociation of others. Identification of DDX3 as a TRAIL-R2-associated inhibitory protein in TRA-8-resistant cells is interesting as other members, such as MDA5 and RIG1, of this RNA helicase family have been shown to be able to regulate apoptosis through their CARD ( 37– 40). Similar to MDA5 and RIG1, DDX3 has a CARD at its NH2 termini. Thus, the further elucidation of the role of DDX3 in TRAIL-R2-mediated apoptosis, particularly in the development of TRA-8 resistance, would be important.
There is accumulating evidence indicating a synergism between anti–death receptor therapy and chemotherapy, although the synergistic mechanisms are not fully understood. Our studies show that chemotherapeutic agents synergistically enhance TRA-8-mediated apoptosis and are very effective at promoting apoptosis when used in combination with TRA-8 in certain tumor cells that are completely resistant to TRA-8 single-agent therapy. These chemotherapeutic agents are likely to activate an alternative apoptotic pathway, for example, caspase-2 pathway, which complements the induced defects in the death receptor–mediated apoptotic pathway. Furthermore, this study shows that this combination therapy strategy is able reverse induced resistance to TRA-8, suggesting that chemotherapeutic agents might directly act on the death domain protein complex through down-modulating DDX3 to enhance TRA-8-triggered TRAIL-R2 signal transduction. The fully restored caspase cascade in TRA-8-resistant cells by chemotherapeutic agents suggests that chemotherapeutic agents might activate the alternative caspase pathways through the initiator caspase-2. Caspase-2 is an initiator caspase in the stress-induced apoptosis pathway ( 35), which is shown to be involved in TRAIL-mediated apoptosis pathway ( 41). These alternatively activated caspases might help to remove the inhibitory factors that are constantly associated with TRAIL-R2 in TRA-8-resistant cells, such as DDX3, thereby enhancing the death domain function by promoting the recruitment of FADD and caspase-8.
Grant support: Sankyo Co. Ltd. of Japan and NIH grants P50 CA83591, P50 CA89019, U19 AI56542, and P20 CA101955.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Received December 7, 2005.
- Revision received May 2, 2006.
- Accepted June 23, 2006.
- ©2006 American Association for Cancer Research.