Tumor Necrosis Factor-related Apoptosis-inducing Ligand Receptors Signal NF-κB and JNK Activation and Apoptosis through Distinct Pathways*

Tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) is a member of the TNF family that interacts with several receptors, including TRAIL-R1, TRAIL-R2, and TRAIL-R4. TRAIL-R1 and TRAIL-R2 can induce apoptosis of cancer cells and activate the transcription factor NF-κB. TRAIL-R4 can activate NF-κB and protect cells from TRAIL-induced apoptosis. Here we show that TRAIL-R1-, TRAIL-R2-, and TRAIL-R4-induced NF-κB activation are mediated by a TRAF2-NIK-IκB kinase α/β signaling cascade but is MEKK1 independent. TRAIL receptors also activate the protein kinase JNK. JNK activation by TRAIL-R1 is mediated by a TRAF2-MEKK1-MKK4 but not the TRAF2-NIK/IκB kinase α/β signaling pathway. We also show that activation of NF-κB or overexpression of TRAIL-R4 does not protect TRAIL-R1-induced apoptosis. Moreover, inhibition of NF-κB by IκBα sensitizes cells to tumor necrosis factor- but not TRAIL-induced apoptosis. These findings suggest that TRAIL receptors induce apoptosis, NF-κB and JNK activation through distinct signaling pathways, and activation of NF-κB is not sufficient for protecting cells from TRAIL-induced apoptosis.


Tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) is a member of the TNF family that interacts with several receptors, including TRAIL-R1, TRAIL-R2, and TRAIL-R4. TRAIL-R1 and TRAIL-R2 can induce apoptosis of cancer cells and activate the transcription factor NF-B. TRAIL-R4 can activate NF-B and protect cells from TRAIL-induced apoptosis. Here we show that TRAIL-R1-, TRAIL-R2-, and TRAIL-R4-induced NF-B activation are mediated by a TRAF2-NIK-IB kinase ␣/␤ signaling cascade but is MEKK1 independent. TRAIL receptors also activate the protein kinase JNK. JNK activation by TRAIL-R1 is mediated by a TRAF2-MEKK1-MKK4 but not the TRAF2-NIK/IB kinase ␣/␤ signaling pathway. We also show that activation of NF-B or overexpression of TRAIL-R4 does not protect TRAIL-R1-induced apoptosis. Moreover, inhibition of NF-B by IB␣ sensitizes cells to tumor necrosis factor-but not TRAIL-induced apoptosis. These findings suggest that TRAIL receptors induce apoptosis, NF-B and JNK activation through distinct signaling pathways, and activation of NF-B is not sufficient for protecting cells from TRAIL-induced apoptosis.
Tumor necrosis factor (TNF) 1 -related apoptosis-inducing ligand (TRAIL) is a member of the TNF family, which also includes TNF and FasL (1)(2)(3). Unlike TNF and FasL, which are mainly expressed by activated immune cells, TRAIL is constitutively expressed in most normal tissues (4,5). Previous studies suggest that TRAIL is capable of inducing apoptosis of various cancer cell lines but not of normal cells (3)(4)(5), pointing to the possibility of developing TRAIL as a reagent for cancer treatment.
Currently, the intracellular signaling pathways responsible for TRAIL receptor-mediated NF-B activation are unclear, and the mechanisms responsible for TRAIL receptor-induced apoptosis are controversial. It has been reported that TRADD and FADD, two death domain-containing cytoplasmic proteins involved in TNF-R1 signaling, interact with TRAIL-R1 and TRAIL-R2 and are involved in apoptosis mediated by these receptors (8,21). However, other studies have reached opposite conclusions (6,7,10). Moreover, studies using FADD knockout embryonic fibroblasts suggest that FADD is not required for apoptosis induced by overexpression of TRAIL-R1 (22).
The signaling pathways mediated by TNF receptor family members have been best illustrated by studies with TNF-R1. TNF-R1 is a death domain-containing receptor that can induce apoptosis and activate NF-B and JNK kinase (23)(24)(25)(26). The death domain of TNF-R1 interacts with TRADD in a TNF-dependent process (24,25,27). Once TRADD is recruited to TNF-R1, it functions as an adapter protein to recruit several structurally and functionally divergent proteins, including FADD, RIP, TRAF2, and cIAP1 (25,27,28). The interaction of TRADD with FADD leads to apoptosis through the activation of a caspase cascade (24). The interaction of TRADD with TRAF2 and RIP activates NIK, a member of the mitogenactivated protein kinase kinase kinase family (29). Once NIK is activated, it further activates two downstream kinases, IKK␣ and IKK␤ (29 -34). It has been shown that IKK␣ and IKK␤ form a heterodimer complex that directly phosphorylates IBs (29 -35). Once IBs are phosphorylated, they are degraded, and consequently the active NF-B is released (36,37). * 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.
In addition to NIK, TRAF2 and RIP can also activate MEKK1, another member of the mitogen-activated protein kinase kinase kinase family (26). Although it has been suggested that overexpression of MEKK1 activates NF-B (38), it is believed that under physiological conditions, MEKK1 mediates TNF-R1-induced JNK but not NF-B activation (35).
In this study, we investigated the mechanism of downstream signaling by TRAIL receptors. The results indicate that TRAIL-R1-, TRAIL-R2-, and TRAIL-R4-induced NF-B activation are mediated by a TRAF2-NIK-IKK␣/␤-dependent signaling cascade, whereas TRAIL-R1-induced JNK activation is mediated by a TRAF2-MEKK1-MKK4 dependent signaling cascade. We also show that inhibition of TRAIL-R1-induced NF-B and JNK activation pathways does not block TRAIL-R1-induced apoptosis. In addition, our data indicate that NF-B activation is not sufficient for protecting cells from TRAIL-induced apoptosis. TRAIL-R4 expression vector was constructed by replacing the TRAIL-R1 cDNA in the pCMV1-Flag-DR4 (TRAIL-R1) vector (6) with a polymerase chain reaction product of TRAIL-R4 cDNA. The parent vector contains a DNA fragment encoding a signal peptide at 5Ј of the Flag tag, and therefore the DNA fragment encoding the native Nterminal signal peptide of TRAIL-R4 was omitted by polymerase chain reaction.
Cell Transfection and Reporter Gene Assays-The human embryonic kidney 293 cell line was maintained in high glucose Dulbecco's modified Eagle's medium containing 10% fetal calf serum, 100 g/ml penicillin G, and 100 g/ml streptomycin (Life Technologies, Inc.). For reporter gene assays, ϳ2 ϫ 10 5 cells/well were seeded on 6-well (35 mm) dishes. Cells were transfected the following day by the standard calcium phosphate precipitation method (39). Luciferase reporter assays were performed using a luciferase assay kit (Pharmingen) following the manufacture's protocols.
Western Blotting-Western blots for detection of Flag-tagged TRAIL-R1, TRAIL-R2, and TRAIL-R4 were performed with a monoclonal anti-Flag following previously described procedure (27,40).
Yeast Two-hybrid Screenings-The cDNA encoding the intracellular domain of TRAIL-R1 was inserted in frame into the Gal4 DNA-binding domain vector pGBT9 (CLONTECH). The human leukocyte, spleen, and 293 cell two-hybrid cDNA libraries were also from CLONTECH. The isolation of positive clones and subsequent two-hybrid interaction analyses were carried out as described (23,24,28,40).
Apoptosis Assays-293 cells were transfected with 0.1 g of pCMV-␤-galactosidase plasmid and various amounts of indicated plasmids. ␤-Galactosidase co-transfection assays for determination of cell death were performed as described (23,24,28,40). Transfected cells were stained with X-gal as described previously (41). The number of blue cells from four viewing fields in one well of a 35-mm dish was determined by counting under a microscope. The average number from one representative experiment in which each transfection was done in duplicate is shown.
Solid-Phase Kinase Assays-Cytokine-treated or -transfected cells were lysed with 600 l of ice-cold lysis buffer (20 mM Tris (pH 7.5), 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 10 g/ml aprotinin, 10 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 10 mM NaF, 2 mM Na 3 VO 4 , and 1 mM dithiothreitol). The lysate was mixed with 15 l of 1:1 slurry of GST-c-Jun-Sepharose beads, and the mixture was incu-bated at 4°C for 1 h. The beads were then washed twice with lysis buffer and once with kinase assay buffer (20 mM Hepes, pH 7.5, 10 mM ␤-glycerophosphate, 10 mM p-nitrophenylphosphate, 10 mM MgCl 2 , 1 mM dithiothreitol, 50 M Na 3 VO 4 ). The washed beads were resuspended in 30 l of kinase assay buffer containing 1 l of [␥-32 P]ATP (10 Ci/L, 1 Ci ϭ 37 GBq) and incubated at 30°C for 30 min. The reaction was terminated by the addition of 30 l of 2 ϫ Laemmli sample buffer and boiled for 5 min. The samples were fractionated by SDS-polyacrylamide gel electrophoresis. The gels were washed in fixing buffer (10% acetic acid, 30% methanol) three times, each for 10 min, and then dried. Autoradiography was performed for 5-30 min. Fold induction of JNK kinase activity was determined by phosphoimaging analysis.
Screening of TRAIL-resistant Cells--HeLa or MCF7 cells (5 ϫ 10 5 ) were treated with 200 ng/ml recombinant TRAIL for 24 h. Treated cells were switched to new medium containing 200 ng/ml TRAIL for additional 24 h. Surviving cells were then amplified and designated as HeLa-TL-R and MCF7-TL-R, respectively.

Activation of NF-B by TRAIL Receptors Is Mediated through a TRAF2-NIK-IKK␣/␤-dependent Signaling Cascade-Al-
though several studies indicated that TRAIL could induce NF-B activation (8,10,21), one group reported results contrary to this (6). To determine whether TRAIL could activate NF-B, we transfected 293 cells with a NF-B-luciferase reporter construct and performed luciferase reporter gene assays. As shown in Fig. 1, treatment with recombinant soluble TRAIL induced NF-B activity. In this experiment, TRAIL was less potent than TNF in activating NF-B (Fig. 1). To determine the relative contributions of various TRAIL receptors to TRAILinduced NF-B activation, we compared the effects of overexpression of individual TRAIL receptors on NF-B activity in reporter gene assays. As shown in Fig. 2, TRAIL-R1, TRAIL-R2, and TRAIL-R4 all activated NF-B. TRAIL-R1 was found to be more potent than TRAIL-R2 and TRAIL-R4 in activating NF-B in this experiment, at least partially because of its higher expression level than TRAIL-R2 and TRAIL-R4 (Fig. 2).
Because TRAIL receptors have similar biological effects as TNF receptors, we tested whether the cytoplasmic proteins involved in TNF receptor signaling, including TRADD, TRAF2, NIK, MEKK1, IKK␣, and IKK␤, also participate in TRAIL receptor signaling. To do this, we determined whether their dominant negative mutants could block TRAIL receptor-mediated NF-B activation in reporter gene assays. TRADD(296S), a TRADD dominant negative mutant that inhibits TNF-R1mediated NF-B activation (42) (Fig. 2D), did not block TRAIL-R1-, TRAIL-R2-, and TRAIL-R4-induced NF-B activation (Fig.  2), suggesting that TRADD is not involved in TRAIL receptormediated NF-B activation pathways. Consistent with this observation, we failed to detect an interaction between TRADD and TRAIL-R1, TRAIL-R2, or TRAIL-R4 in co-transfection and co-immunoprecipitation experiments (data not shown). Interestingly, overexpression of TRADD(296S) potentiated TRAIL receptor-induced NF-B activation (Fig. 2).
TRAIL Receptors Activate JNK through a TRAF2-MEKK1-MKK4-dependent Pathway-Several TNF receptor family members are capable of activating the JNK kinase pathway. To determine whether TRAIL and its receptors have a similar effect, we performed solid-phase kinase assays with GST-c-Jun as a substrate. As shown in Fig. 4A, TRAIL treatment induced JNK activation in 293 cells. Similarly, overexpression of TRAIL-R1, TRAIL-R2, and TRAIL-R4 activated JNK (Fig. 4B) and was further enhanced by TRAIL treatment. To explore the possible signaling pathways leading to TRAIL receptor-induced JNK activation, we tested whether the dominant negative mutants of TRAF2, NIK, MEKK1, MKK4, and IKK␤ could block TRAIL-R1-induced JNK activation. As shown in Fig. 4C, TRAF2-(87-501), MEKK1(K1255M), and the MKK4 dominant negative mutant MKK4-DN inhibited TRAIL-R1-induced JNK activation, whereas NIK(K429A/K430A) and IKK␤(K44A) had no significant inhibitory effect on TRAIL-R1-induced JNK activation (Fig. 4C). These data suggest that TRAIL-R1-induced JNK activation is mediated by a TRAF2-MEKK1-MKK4 dependent pathway and is independent of the NIK-IKK␣/␤ cascade.
Because the above data indicated that TRAIL-R1-induced NF-B and JNK activation pathways might bifurcate at TRAF2, we examined whether TRAF2 could directly interact with TRAIL-R1 by co-transfection and co-immunoprecipitation experiments. We found that TRAF2 did not interact with TRAIL-R1, TRAIL-R2, and TRAIL-R4 (data not shown), suggesting that unidentified adapter molecule(s) may link TRAF2 to TRAIL receptors.
FADD Is Involved in TRAIL-R1-and TRAIL-R2-induced Apoptosis Pathway-Previously, it has been reported that TRAIL-R1 and TRAIL-R2 induce apoptosis through a FADDdependent pathway (8,21). In contrast, other reports have suggested that FADD is not involved in TRAIL-R1-and TRAIL-R2-induced apoptosis (6, 7, 10, 23). We screened several human cDNA libraries using the yeast two-hybrid system with TRAIL-R1 intracellular domain as bait. These screenings identified FADD as a protein that specifically interacted with TRAIL-R1 (data not shown). Consistent with the physical interaction, FADD-(80 -205), a dominant negative mutant of FADD that inhibits TNF-R1-induced apoptosis (24), significantly inhibited TRAIL-R1-induced apoptosis in a well established apoptosis assay (23, 24, 28, 40) (Fig. 5A). These data suggest that FADD is involved in a TRAIL-R1-induced apoptosis pathway.

Activation of NF-B Is Not Capable of Protecting Cells from TRAIL-R1-induced Apoptosis-Although TRAIL induces apoptosis of various cancer cells, some cancer cells and normal
cells are resistant to TRAIL-induced apoptosis, even though these cells express TRAIL-R1 and TRAIL-R2 (3). Because TRAIL-R1, TRAIL-R2, and TRAIL-R4 can activate NF-B, one of the possible mechanisms responsible for the resistance of a cell to TRAIL may be because of a dominant effect of NF-B activation, which has been shown to protect cells from TNFinduced apoptosis (18 -20). In this context, it has been suggested that TRAIL-R4, which can induce NF-B activation but not apoptosis, can protect cells from TRAIL-induced apoptosis (16). To investigate whether NF-B activation is responsible for TRAIL-R4-mediated protection of cells from TRAIL-induced apoptosis, we tested the effects of overexpression of TRAIL-R4 and up-regulation of NF-B activity on TRAIL-R1-mediated apoptosis. As shown in Fig. 5B, overexpression of TRAIL-R4 did not protect TRAIL-R1-induced apoptosis. Co-transfection of 293 cells with expression vectors for TRAIL-R1 and NIK or IKK␤ greatly increased NF-B activity in comparison to TRAIL-R1 transfection alone (Fig. 5C). However, up-regulation of NF-B activity by overexpressing NIK and IKK␤ had no protective effect on TRAIL-R1-induced apoptosis (Fig. 5B). These data suggest that overexpression of TRAIL-R4 and activation of NF-B do not protect cells from TRAIL-R1-induced apoptosis.
Inhibition of NF-B Activation Potentiates TNF-, but Not TRAIL-induced Apoptosis-IB␣(S32A/S36A) is an IB␣ mutant that has more potent inhibitory effect on NF-B activation than its wild type counterpart (36,37). To test whether inhibition of NF-B activation can sensitize cells to TRAIL, we determined the effect of IB␣(S32A/S36A) on TRAIL-induced apoptosis. To do this, we first isolated TRAIL-resistant HeLa and MCF7 cells. In these experiments, 90% of HeLa cells and 30% of MCF7 cells were killed. TRAIL-resistant cells, designated as HeLa-TL-R and MCF7-TL-R, respectively, were then amplified and transfected with expression vectors for IB␣(S32A/S36A) and a ␤-galactosidase reporter gene. Fourteen hours after transfection, the cells were treated with TNF, TRAIL, or left untreated for 10 h and then stained by X-gal. As shown in Fig.  6, transfection of IB␣(S32A/S36A) sensitized both HeLa-TL-R and MCF7-TL-R cells to TNF-but not TRAIL-induced apoptosis. These data are consistent with the hypothesis that activation of NF-B protects TNF-but not TRAIL-induced apoptosis. DISCUSSION TRAIL stimulation induces three distinct biological effects: apoptosis, NF-B, and JNK activation. These effects of TRAIL are mediated through three signaling receptors, including TRAIL-R1, TRAIL-R2, and TRAIL-R4. In this report, we investigated the mechanisms of downstream signaling by the three mentioned TRAIL receptors.
Our studies indicate that TRAIL-R1, TRAIL-R2, and TRAIL-R4 are also capable of inducing JNK kinase activity ( Fig. 4). TRAIL-R1-induced JNK activation can be inhibited by dominant negative mutants of TRAF2, MEKK1, and MKK4 but not by those of NIK and IKK␤. Therefore, TRAIL-R1-induced JNK activation is mediated by a TRAF2-MEKK1-MKK4 dependent pathway and is independent of the NIK-IKK kinase cascade. These data suggest that TRAIL-R1-induced NF-B and JNK activation pathways bifurcate at TRAF2. Because TRAF2 does not directly interact with TRAIL-R1, TRAIL-R2, and TRAIL-R4, whereas FADD is not involved in TRAIL receptor-mediated NF-B activation pathway, we believe that unidentified adapter proteins other than TRADD and FADD are required for recruiting TRAF2 to TRAIL-R1, TRAIL-R2, and TRAIL-R4 signaling complexes.
In addition to NF-B and JNK activation, TRAIL-R1 can potently induce apoptosis. The pathway leading to TRAIL-R1induced apoptosis, however, remains controversial. We screened yeast two-hybrid libraries with the intracellular domain of TRAIL-R1 as bait and identified FADD as a protein that specifically interacted with TRAIL-R1. 2 Furthermore, a FADD dominant negative mutant significantly inhibited TRAIL-R1-induced apoptosis in 293 cells (Fig. 5A). Our data support the hypothesis that FADD is involved in TRAIL-R1induced apoptosis pathway. Previously, it has been shown that overexpression of TRAIL-R1 can induce apoptosis in FADD(Ϫ/Ϫ) embryonic fibroblasts (22), suggesting that FADD is dispensable for TRAIL-R1-induced apoptosis. However, these experiments can not exclude the possibility that FADD is required for apoptosis induced by ligation of TRAIL-R1 with TRAIL in untransfected cells. For example, the death domain containing TRAIL-R1, when overexpressed, may artificially interact with other death domain-containing proteins, such as RIP, and therefore induce apoptosis in FADD(Ϫ/Ϫ) cells. Alternatively, a FADD-like molecule, which may have higher affinity with TRAIL-R1, can also transduce the death signal from TRAIL-R1 to the downstream caspase cascades.
Previous studies indicate that TRAIL-R4 can protect cells from TRAIL-induced apoptosis. At least two hypotheses have been proposed to explain this observation. First, TRAIL-R4 may function as a decoy receptor for TRAIL-induced apoptosis, for example, by competing with TRAIL-R1 and TRAIL-R2 for TRAIL binding. This hypothesis, however, cannot explain the observations that some TRAIL-sensitive cells express both TRAIL-R3 and TRAIL-R4, whereas some TRAIL resistant cells do not have detectable TRAIL-R3 and TRAIL-R4 (3). Second, TRAIL-R4 may activate a protective signal to inhibit TRAILinduced apoptosis, for example, by activating NF-B. Previously, it has been shown that activation of NF-B can inhibit TNF-induced apoptosis, probably through transcriptional induction of apoptosis inhibitory genes (20). However, this hypothesis is complicated by the fact that stimulation of TRAIL-R1 and TRAIL-R2 simultaneously activates NF-B and induces apoptosis. One can argue that NF-B activity induced by TRAIL-R1 and TRAIL-2 is too low to antagonize the dominant apoptotic effect induced by TRAIL-R1 and TRAIL-R2. In this study, we found that overexpression of TRAIL-R4 did not protect cells from apoptosis induced by TRAIL-R1. In addition, a dramatic up-regulation of NF-B activity by overexpressing NIK and IKK␤ had no significant effect on TRAIL-R1-induced apoptosis (Fig. 5). Moreover, inhibition of NF-B activation by an IB␣ mutant, IB␣(S32A/S36A), sensitized cells to TNFbut not TRAIL-induced apoptosis (Fig. 6). These data suggest that activation of NF-B is not sufficient for protecting cells from TRAIL-induced apoptosis, and an alternative mechanism other than NF-B activation may account for the protective role of TRAIL-R4 on TRAIL-induced apoptosis.
In conclusion, our data indicate that TRAIL induces apoptosis, NF-B activation, and JNK activation through distinct pathways (Fig. 7). We also conclude that NF-B activation is not sufficient for protecting cells from TRAIL-induced apoptosis.