The role of TRADD in TRAIL-induced apoptosis and signaling

Abstract

Tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) is a member of the TNF superfamily. TRAIL is promising for anticancer therapy because it induces apoptosis in cancer cells with little or no toxicity to normal cells; hence, TRAIL-receptor agonists are currently undergoing clinical trials for cancer treatment. However, many molecular signaling mechanisms in TRAIL signaling are not completely characterized. The functions of adaptor proteins, including TNF-receptor-associated death domain protein (TRADD) and receptor-interacting protein-1 (RIP1) in TRAIL signaling have been controversial. We demonstrate that while wild-type mouse embryonic fibroblasts (MEFs) are completely resistant to TRAIL-induced apoptosis, MEFs derived from Tradd^−/− mice are hypersensitive to TRAIL (IC₅₀∼0.5 nM rmTRAIL, 24 h), an effect also seen in primary keratinocytes treated with TRAIL/CHX. Restoration of TRADD in Tradd^−/− MEFs restores TRAIL resistance, indicating that TRADD plays a survival role in TRAIL signaling. We show that TRADD is recruited to the TRAIL-receptor complex, and RIP1 recruitment is mediated by TRADD. While early activation of the MAP kinase ERK is deficient in Tradd^−/− cells, the main mechanism for enhanced TRAIL sensitivity is likely due to increased recruitment of FADD to the receptor complex, indicating that TRADD may limit FADD binding within the receptor complex and also mediate RIP1-dependent nonapoptotic signaling events, thus reducing caspase activation and subsequent apoptosis. These novel findings have potential implications for cancer therapy using TRAIL-receptor agonists.—Cao, X., Pobezinskaya, Y. L., Morgan, M. J., Liu, Z. The role of TRADD in TRAIL-induced apoptosis and signaling.

Tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL), also known as Apo2L (or officially as TNFSF10), is a member of the TNF superfamily. TRAIL has emerged as a promising anticancer therapy because it induces apoptosis in cancer cells with little or no toxicity to normal cells. TRAIL initiates the extrinsic (death receptor) pathway in target cells by binding to its receptors, TRAIL-R1 and TRAIL-R2 [also known as DR4 (TNFRSF10A) and DR5 (TNFRSF10B) in humans; mice have only one receptor, TRAIL-R (Tnfsf10)], resulting in receptor aggregation and recruitment of the adapters to form the death-inducing signaling complex (DISC).

Like the other death receptors, such as TNF-R1 and Fas, TRAIL-R1 and TRAIL-R2 also contain a cytoplasmic death domain (DD) that associates with DD-containing adaptor proteins. Recruitment of these adaptor proteins to death receptors, such as Fas-associated protein with death domain (FADD), TNF-receptor-associated death domain protein (TRADD), or receptor-interacting protein-1 (RIP1), triggers multiple cell signals, including the activation of caspases, mitogen-activated protein kinases (MAPKs), and nuclear factor-κB (NF-κB) (1). The canonical TNF signaling pathway through TNFR1 consists of a receptor-associated TRADD-TRAF2-RIP complex (complex I) and a subsequently dissociated TRADD-TRAF2-RIP-FADD apoptotic complex (complex II). Complex I is thought to be responsible for activating signaling pathways such as NF-κB and JNK, while complex II activates the caspase cascade to trigger apoptosis.

Unlike in TNFR1 signaling, FADD is recruited directly to the endogenous TRAIL-receptor disc (2–4) apparently through direct interactions with the TRAIL receptors (5–7), and FADD is thought to be essential for TRAIL-induced apoptosis (2, 8). FADD has both a DD and a death effector domain (DED), which enables caspases 8 and 10 to bind via their respective DEDs. The proximity of caspase-8 molecules to each other allows their autoactivation; they then directly cleave and promote the activation of caspase-3, and induce Bid cleavage and activation as well, which acts to release proapoptotic proteins from the mitochondria. Activated caspases, in turn, cleave numerous cellular proteins, resulting in the biochemical and morphological hallmarks of apoptosis.

In addition to inducing apoptosis, TRAIL-receptor stimulation also leads to activation of IKK, ERK, p38, and JNK. However, there is considerable debate regarding how these signaling pathways are activated, and the functions of adaptors in the process. TRADD and RIP1 have been reported to interact with TRAIL receptors in overexpression systems, particularly in the presence of overexpressed FADD (5, 6, 9). In some studies examining the endogenous cellular components of the TRAIL DISC, neither protein was detected (2, 4), while other groups found alternatively that RIP1 but not TRADD was detected within the TRAIL DISC (9), or that RIP1 was found associated with FADD only in secondary complexes but could not be detected in association with the TRAIL receptors (10, 11). It has been suggested that in TRAIL signaling, IKK, p38, and JNK are activated in FADD-dependent secondary complexes (10). RIP1 has been shown to be necessary for TRAIL-initiated activation of IKK (9, 10), as well as the activation of p38 (10). The literature is inconsistent on whether RIP1 is required for TRAIL-induced JNK activation; however, TNF-associated factor 2 (TRAF2) appears to be required for full JNK activation in response to TRAIL (9, 10). Knockdown of RIP1 and TRAF2 by siRNA increases TRAIL-induced cell death, indicating that these molecules are protective (10). The role of TRADD in these TRAIL-initiated pathways has not been addressed.

We have previously shown that TRADD was critical for TNF-α-induced signaling and cell death (12). Here, we examined the role of TRADD in TRAIL-initiated cell death and signaling by comparing wild-type (WT) and Tradd^−/− MEFs. Remarkably, we found that TRADD-deficient MEFs were hypersensitive to TRAIL, while WT MEFs were resistant. Significantly, we were able to detect endogenous TRADD in the TRAIL-stimulated receptor complex. Reconstitution of TRADD in Tradd^−/− MEFs not only reduced the recruitment of FADD to the TRAIL-receptor complex but also restored the binding of RIP1, suggesting that TRADD may compete with FADD in binding to the TRAIL receptor and that it may also be responsible for RIP1 recruitment to the TRAIL-receptor complex. ERK and IκBα phosphorylation was also impaired in Tradd^−/− MEFs in response to TRAIL treatment. At the same time, competition for the receptor in WT cells likely prevents sensitivity to FADD-mediated apoptosis.

MATERIALS AND METHODS

Reagents and antibodies

Soluble recombinant murine TRAIL (1121-TL-010) was purchased from R&D Systems (Minneapolis, MN, USA). Glutathione S-transferase (GST)-rhTRAIL was a kind gift from S. Lipkowitz [U.S. National Institutes of Health (NIH), Bethesda, MD, USA]. Glutathione-agarose (G4501) and cycloheximide (C1988) were from Sigma (St. Louis, MO, USA). zVAD (03FK109) was from MP Biomedicals (Solon, OH, USA). Antibody to phosphorylated JNK (44-682) was from Biosource (San Jose, CA, USA); anti-TRADD (H-278) was from Santa Cruz Biotechnology (Santa Cruz, CA, USA); anti-JNK (G151-666), anti-mouse TRAIL-R1 (557868), and anti-RIP1 (G322-2) were from BD Pharmingen (San Diego, CA, USA); anti-actin (AC-40) was from Sigma; antibodies to phosphorylated ERK (9101), p38 (9215), and IκBα (14D4) and anti-ERK (9106), anti-p38 (9212), and anti-caspase 3 (9662) were from Cell Signaling (Danvers, MA, USA); anti-human FADD (F36620) was from BD Bioscience (San Jose, CA, USA); anti-mouse FADD was a kind gift from Jianke Zhang (Thomas Jefferson University, Philadelphia, PA, USA); anti-human TRAIL-R1 (200-401-982) was from Rockland Biotech (Gilbertsville, PA, USA).

Cell culture and transfection

WT, Tradd^−/− and FADD^−/− MEFs, as well as HeLa cells, were cultured in DMEM supplemented with 10% calf serum, 2 mM glutamine, and 100 U/ml penicillin and streptomycin. Cells were transfected with Lipofectamine (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's instruction. Primary murine keratinocytes were isolated as done by Lichti et al. (13).

Reconstitution of Tradd^−/− MEFs with mTRADD-Flag

RNA was isolated from WT MEFs and mouse TRADD. cDNA was amplified by RT-PCR using sense primer 5′-TTTTAAGCTTCATGGCAGCCGGTCAG-3′ and antisense primer 5′-TTTTCTCGAGAGTACTAGACTTAGGC-3′, then cloned into plasmid pFLAG-c1 using XbaI and HindIII. Tradd^−/− MEFs were transfected with this plasmid, and puromycin (2 μg/ml)-resistant clones were selected.

Western blot analysis

Cells were seeded into 6-well plates for 24 h before treatment. After treatment, as indicated in the figure legends, the cells were lysed with M2 buffer (20 mM Tris, pH 7; 0.5% Nonidet P-40; 250 mM NaCl; 3 mM EDTA; 3 mM EGTA; 2 mM dithiothreitol; 0.5 mM phenylmethylsulfonyl fluoride; 20 mM β-glycerol phosphate; 1 mM sodium vanadate; and 1 μg/ml leupeptin) at 4°C for 30 min and analyzed by SDS-PAGE, followed by immunoblot analysis using indicated antibodies.

Cytotoxicity assay

Cell death was assessed by tetrazolium dye colorimetric test. After dissolving cells in DMSO, the absorbance of the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Sigma) product was measured with a plate reader at 570 nm.

GST pulldown and DISC analysis

Cells (1×10⁸) were treated with 1 μg/ml GST-rhTRAIL for 15 min or left untreated. Cells were then washed twice with ice-cold PBS and lysed with 1 ml M2 buffer. The soluble fraction was incubated with 30 μl of 50% glutathione-Sepharose beads overnight on a rotator at 4°C. In some experiments, as indicated, GST-TRAIL (1 μg/ml) was added to the lysates prepared from nonstimulated cells (after lysis) to precipitate the nonstimulated receptors. After 3 washes with M2 buffer, the bound proteins were eluted by boiling for 5 min in SDS-PAGE loading buffer, resolved in 4–20% SDS-PAGE, and analyzed by immunoblot.

RESULTS AND DISCUSSION

Consistent with previous studies, TRAIL alone did not induce the death of WT MEFs; however, surprisingly, we observed that Tradd^−/− MEFs were quite sensitive to TRAIL treatment alone (Fig. 1A). When used in combination with CHX, TRAIL triggered cell death efficiently in WT MEFs, but death in Tradd^−/− MEFs was substantially greater (Fig. 1B). In contrast, but in harmony with our previous data (12), the Tradd^−/− MEFs lacked any sensitivity to TNF-α in the presence of CHX, whereas the WT cells predominantly died in response to TNF-α. At lower doses of CHX (0.1 μg/ml), Tradd^−/− MEFs were mostly dead, while WT MEFs were unaffected (Fig. 1C). TRAIL has been reported to induce caspase-independent necrosis in some cell types (14). However, the cell death induced by the TRAIL-CHX combination in MEFs was clearly blocked by pretreatment with the pancaspase inhibitor zVAD at any concentration of CHX used, suggesting that only apoptosis was occurring in these cells (Fig. 1C). Furthermore, caspase-3 cleavage was observed by Western blot only in Tradd^−/− MEFs, not in WT cells, starting ∼2 h after treatment with recombinant murine TRAIL and peaking at 4 h (Fig. 1D). As previously reported, TRAIL-induced death was dependent on FADD, since FADD-deficient MEFs were completely resistant to TRAIL plus CHX (Supplemental Fig. S1A). In addition, using a pulldown of human GST-TRAIL, which was similarly able to induce cell death in Tradd^−/− MEFs or HeLa cells (Supplemental Fig. S1B), we were able to detect FADD directly within the TRAIL-receptor complex (Supplemental Fig. S1C). Therefore, the presence of TRADD in the cell serves to prevent caspase-dependent apoptosis induced by TRAIL through FADD.

Figure 1.

Tradd^−/− MEFs are hypersensitive to TRAIL-induced apoptosis. A–C) MTT assay shows the viability (average±se) of cells treated for 4 h. Pretreatment with inhibitors CHX and zVAD, where indicated, was for 1 h. Cells were treated with or without murine TRAIL (20 ng/ml) or TNF (30 ng/ml), as indicated: no further treatment (A) or treatment in the presence of CHX (10 μg/ml) or CHX (0.1 μg/ml) (B), or CHX (0.1 μg/ml) plus zVAD (20 μM) (C). D) Immunoblot analysis of total caspase 3, including procaspase3 (Procasp3) and cleaved caspase3 (cleaved casp3) in WT and Tradd^−/− MEF lysates treated for indicated times with murine TRAIL (20 ng/ml).

To confirm that the different responses of WT and Tradd^−/− MEFs to TRAIL treatment were due to the loss of TRADD, we transfected Tradd^−/− MEFs with plasmid expressing Flag-tagged mTradd, and selected cells stably expressing Flag-mTradd. Expression of TRADD in this clone was within 2-fold of the expression level of endogenous TRADD in WT cells (Fig. 2A). Reconstitution of TRADD in Tradd^−/− MEFs not only restored sensitivity to TNF-α plus CHX, as predicted (Fig. 2B), but also eliminated much of their sensitivity to TRAIL (Fig. 2C), confirming that TRADD protects MEFs from TRAIL-induced apoptosis.

Figure 2.

Reconstitution of TRADD in Tradd^−/− MEFs restores sensitivity to TNF-α and resistance to TRAIL. A) Immunoblot analysis of TRADD and β-actin in lysates of WT, Tradd^−/−, and TRADD-reconstituted MEFs. B, C) MTT assay of Tradd^−/− and Flag-TRADD-reconstituted MEFs treated for 24 h with CHX (10 μg/ml) or CHX plus murine TNF (30 ng/ml) (B) or no treatment or 1 μg/ml GST-rhTRAIL (C).

In addition to the activation of caspases, TRAIL can also induce other signaling pathways, such as the activation of NF-κB and MAPKs (1). To determine how TRADD prevents TRAIL-induced apoptosis in MEFs, we examined the components of the TRAIL DISC in Tradd^−/− MEFs and the activation of the other signaling pathways.

Using a GST pulldown of GST-TRAIL, we were able to detect TRADD in the death-receptor complex after TRAIL treatment (Fig. 3A), thus confirming that TRADD has a direct role in TRAIL signaling. As RIP1 has also been shown to be important in many of these signals in both the TNF-α and TRAIL signaling pathways (9, 15), we checked whether TRADD deficiency affected the recruitment of RIP1 in the same GST pulldown assay. In Tradd^−/− MEFs, the recruitment of RIP1 to the TRAIL receptor was substantially less than in the reconstituted MEFs (Fig. 3A). This result is consistent with previous experiments, which showed that the overexpression of TRADD facilitated RIP1 recruitment (9). This also suggests that TRADD may play a similar role for TRAIL-induced signaling as it does for the TNF-α signaling pathway in MEFs, although the activation of many signaling pathways by TRAIL is less strong than by TNF-α, as demonstrated in several other cell lines (10).

Figure 3.

TRADD deficiency results in decreased recruitment of RIP1 and increased recruitment of FADD to TRAIL receptor on TRAIL treatment. A) Tradd^−/− and TRADD-reconstituted MEFs were treated with or without 1 μg/ml GST-rhTRAIL for 15 min. GST-rhTRAIL-bound protein complexes were pulled down by glutathione-agarose beads and analyzed for the presence of RIP1, TRADD, and TRAIL-R by Western blot. B) Tradd^−/− and TRADD-reconstituted MEFs were treated with 1 μg/ml GST-rhTRAIL for 15 min or left untreated. GST-rhTRAIL (1 μg/ml) was added after lysis to the lysates of unstimulated cells in order to precipitate unstimulated TRAIL receptors. The DISC was examined for the presence of FADD and TRAIL-R by immunoblot. C) NIH ImageJ 1.38 software was used to quantitate the amount of FADD and TRAIL-R proteins precipitated from the TRAIL-receptor complex of as described in panel B. FADD/TRAIL-R ratios of 4 experiments were calculated and normalized, with the ratio of knockout cells defined as 1. **P = 0.005357.

We next examined whether TRADD deficiency had an effect on FADD recruitment to the DISC. We found that endogenous FADD was recruited to the DISC in a greater abundance in Tradd^−/− MEFs than in Flag-TRADD-reconstituted MEFs, while we detected even less FADD in WT MEFs (Fig. 3B). RIP1 recruitment followed the opposite pattern, with the most RIP1 being recruited in WT cells, with decreasing recruitment of RIP1 in reconstituted cells and almost none in Tradd^−/− MEFs (Supplemental Fig. S1D). The reduction in FADD recruitment in the presence of TRADD was reproducible over the course of several experiments. One such experiment is shown in Supplemental Fig. S1E. When the FADD/TRAIL-receptor protein ratios from GST pulldowns were quantitated for 4 experiments, the reconstitution of TRADD in Tradd^−/− MEFs showed an average 40% decrease in FADD binding, with a value of P = 0.005357 (Fig. 3C). Thus, the presence of TRADD increases RIP1 recruitment and also seems to interfere to some extent with the efficiency of FADD recruitment. Since enhanced sensitivity to TRAIL-induced cell death is associated with enhanced recruitment of FADD to the DISC (16, 17), this may explain, in large part, why Tradd^−/− MEFs are sensitive to TRAIL-induced apoptosis.

We next examined whether there were differences in ERK signaling, given that TRAIL-induced ERK activation has been shown to prevent TRAIL-mediated apoptosis, since it inhibits the processing of caspase-8 and Bid (18, 19). In the short term, ERK activation peaked at ∼15 min after TRAIL treatment and decreased substantially by ∼30 min after treatment in WT MEFs (Fig. 4A). In the Tradd^−/− MEFs, ERK activation by TRAIL was absent, while in the Tradd^−/− MEFs reconstituted with Flag-TRADD, ERK phosphorylation was restored (Fig. 4A). The treatment of WT cells, however, with a pharmacological inhibitor of the ERK pathway failed to sensitize WT MEFs to TRAIL-induced cell death (Supplemental Fig. S1F). Since this compound completely abolishes ERK activation in response to death-receptor stimulation (unpublished results), this lack of sensitivity suggests that lack of ERK activation in TRADD-deficient cells does not solely explain their sensitivity to cell death when compared to WT cells.

Figure 4.

TRAIL-induced signaling is altered in TRADD-deficient MEFs. Immunoblot analysis of lysates of WT, Tradd^−/−, and TRADD-reconstituted MEFs untreated or treated for different periods with murine TRAIL (20 ng/ml) alone or murine TRAIL plus zVAD (20 μM; 4 h; D, F, H), as indicated. Signaling was examined with phospho-ERK (A, E, F), phospho-IκBα (B), phospho-JNK (C, D), or phospho-p38 antibodies (G, H). Blots of actin and ERK, p38, or JNK indicate loading of lanes. Fold of phospho-IκBα increase was shown to be significant in WT MEFs (3.3±0.5, P=0.0031) and TRADD-reconstituted MEFs (2.2±0.6, P=0.0438) compared to Tradd^−/− MEFs when the experiment was repeated 3 times.

Although NF-κB activation is only weakly induced by TRAIL, it has been shown to be involved in the TRAIL-induced release of chemokines and inflammatory cytokines (20, 21). We therefore sought to determine whether TRAIL-induced NF-κB activation depended on TRADD. TRAIL-induced phosphorylation of IκBα was increased in WT MEFs and Flag-TRADD-reconstituted cells but did not increase in Tradd^−/− MEFs (Fig. 4B). This suggests that TRAIL-induced activation of IKK and subsequent NF-κB activation may, in part, be dependent on TRADD. Less NF-κB activation could possibly also contribute to the sensitivity of Tradd^−/− MEFs to TRAIL, since NF-κB often regulates prosurvival genes.

While we could not detect early JNK and p38 activation in WT MEFs, Tradd^−/− MEFs, or reconstituted cells, we observed late activation of JNK, p38, and ERK at ∼4 to 8 h after TRAIL treatment in Tradd^−/− MEFs (Fig. 4C–H). While JNK and ERK were also activated in WT cells during these time points, p38 was only weakly stimulated in WT MEFs. Because the late activation of various MAPKs is often associated with the activation of caspases (22), and because we previously showed that caspases are activated by TRAIL in the TRADD-deficient cells, we compared the activation of these MAPKs in presence of the pan-caspase inhibitor, zVAD. TRAIL-induced JNK activation at 4 h was mostly caspase dependent in Tradd^−/− MEFs (Fig. 4D). Likewise, the late activation of ERK at the 4-h time point in Tradd^−/− MEFs was also caspase dependent (Fig. 4F). Interestingly, JNK and ERK are also activated at this late time point in WT MEFs, but their phosphorylation was not inhibited by zVAD. Therefore, the second phase of JNK and ERK activation in MEFs is probably due to secondary signaling that requires TRADD, but in TRADD-deficient cells, the late phase activation of JNK and ERK is dependent on caspases. The phosphorylation of p38 was substantially inhibited by zVAD in Tradd^−/− MEFs, indicating that p38 was primarily activated by caspases (Fig. 4H), which is consistent with the peak of caspase activation at this time point (Fig. 1D).

To verify that TRADD deficiency sensitizes cell types other than MEFs to TRAIL, we measured TRAIL cytotoxicity in primary keratinocytes derived from WT and TRADD^−/− mice. While WT cells were not sensitive to TRAIL alone or in combination with cycloheximide, the treatment of TRADD^−/− keratinocytes with the TRAIL/CHX combination led to significant toxicity (Supplemental Fig. S1G).

In summary, Tradd^−/− MEFs and keratinocytes have an enhanced sensitivity to TRAIL. Based on our data, we hypothesize that one of the most likely reasons for this result is that TRADD and FADD may bind in a competitive manner within the TRAIL-receptor complex. It is unclear whether reduced FADD recruitment is due to a direct competition between the two molecules for the TRADD, a competition between TRADD and FADD for a FADD that is already bound to the TRAIL receptor, or a combination of the two. However, given the difficulty that other researchers have had in detecting endogenous TRADD in the TRAIL-receptor DISC (2, 4) and the fact that overexpression of FADD enhances TRADD binding to the receptor (7), we favor the second hypothesis. In any case, TRADD likely decreases TRAIL-induced apoptosis, at least in part, through a reduction in FADD recruitment. RIP1 recruitment to the TRAIL-receptor complex is also inhibited in Tradd^−/− MEFs and is the likely reason for a lack of early ERK activation in these cells, as well as the late, second-phase caspase-independent activation of JNK and ERK. TRADD deficiency may also slightly decrease the activation of IKK and thus have slightly less induction of protective gene induction.

Resistance to TRAIL-induced apoptosis is a significant problem associated with TRAIL-based therapy of many tumors (23). Acquired TRAIL resistance during TRAIL therapy may shift a patient's treatment from being beneficial to being detrimental, since once cancer cells are resistant to cell death, TRAIL could stimulate the proliferation and metastasis through other pathways. Thus, a greater understanding the molecular mechanisms of TRAIL-induced apoptosis and signaling may benefit personalized therapy using the combination of TRAIL and other drugs. As TRADD can mediate resistance to TRAIL, TRADD may be a potential future target in treating some TRAIL-resistant cancers.