Tumor Necrosis Factor (TNF) and Phorbol Ester Induce TNF-related Apoptosis-inducing Ligand (TRAIL) under Critical Involvement of NF- (cid:1) B Essential Modulator (NEMO)/IKK (cid:2) *

We show that tumor necrosis factor (TNF) and phorbol 12-myristate 13-acetate (PMA) induce TNF-related apoptosis-inducing ligand (TRAIL) in T cells. In cells deficient for NF- (cid:1) B essential modulator (NEMO)/IKK (cid:2) , an essential component of the NF- (cid:1) B-inducing I- (cid:1) B kinase (IKK) complex, induction of TRAIL expression was completely abrogated but was recovered in cells restored for IKK (cid:2) expression. In cells deficient for re-ceptor-interacting protein expression TNF, but not PMA-induced TRAIL expression was blocked. Inhibition of protein synthesis with cycloheximide blocked PMA, but not TNF-induced up-regulation of TRAIL. As both TNF and PMA rapidly induce NF- (cid:1) B activation this suggests that NEMO/IKK (cid:2) -dependent activation of the NF- (cid:1) B pathway is necessary but not sufficient for up-regulation of TRAIL in T cells. The capability of the NF- (cid:1) B pathway to induce the potent death ligand TRAIL may explain the reported proapoptotic features of this typically antiapoptotic pathway.

Tumor necrosis factor (TNF) 1 -related apoptosis-inducing ligand (TRAIL)/Apo2L belongs to the TNF ligand family and induces apoptosis in a broad range of tumor cells with apparently no cytotoxic activity on most nontransformed cells. TRAIL mRNA was found in a variety of tissues, in particular in activated T cells, and is often induced by interferons (1). TRAIL interacts with five different members of the TNF receptor superfamily: TRAIL-R1/DR4, TRAIL-R2/DR5/TRICK2/Killer, TRAIL-R3/TRID/DcR1/LIT, TRAIL-R4/TRUNDD/DcR2 and osteoprotegerin (2). TRAIL-R1 and TRAIL-R2 contain a death domain and mediate the apoptotic response toward TRAIL. TRAIL-R3, lacking a cytoplasmic domain, is anchored to the cell surface by modification with glycophospholipids and is regarded as a decoy receptor for TRAIL. TRAIL-R4 has an incomplete cytoplasmic death domain and is therefore incapable of signaling apoptosis but is still able to activate the antiapoptotic NF-B pathway. Thus, TRAIL-R4 is maybe involved in regulation of the apoptotic TRAIL response at the cellular level. The soluble receptor osteoprotegerin, which otherwise has a critical role in regulation of osteoclastogenesis, may act as a systemic negative regulator of TRAIL-induced apoptosis (1).
TRAIL is involved in the induction of cell death by activated CD4 ϩ T cells (3,4), dendritic cells (5), and NK cells (6). TRAIL may also play a role in activation-induced T cell death during HIV infection (7)(8)(9). Inhibition of endogenous TRAIL enhances proliferation of autoreactive lymphocytes or synovial cells, thereby contributing to the exacerbation of arthritic inflammation and joint tissue destruction (10). These data argue for an in vivo function of TRAIL in the maintenance of immune homeostasis by counteracting autoimmune responses. However, in contrast to FasL, TRAIL inhibits autoimmune inflammation by inhibition of cell cycle progression instead of induction of T cell apoptosis (10). In accordance with its remarkable tumoricidal activity in vitro, several studies point to a role of TRAIL in anti-tumoral immunity. In this regard, it has been recently shown that TRAIL is the sole principle mediating antitumorigenic activity of interferon-activated human monocytes (11). Moreover, up-regulation of TRAIL plays a central role in IFN␥mediated antimetastatic effects of interleukin-12 and ␣-galactosylceramide (12,13). Further, aside from the capability of dendritic cells to acquire antigen from apoptotic cells to crosspresent these antigens to cytotoxic T cells (14,15), dendritic cells directly mediate cellular apoptosis via TRAIL after stimulation with IFN␣ or IFN␥ (5). Of interest, a recent study indicates that type I IFNs induce the rapid maturation of monocytes into short-lived TRAIL-expressing dendritic cells (16). Moreover, type I IFNs up-regulate TRAIL expression on anti-CD3-stimulated human peripheral blood T cells suggesting that TRAIL contributes to the indirect cytotoxic activity of type I IFNs used in cancer therapy for the treatment of chronic myelogenous leukemias and renal cell carcinomas (17). In accordance with an antitumoral function of TRAIL, the death domain-containing TRAIL receptors 1 (18) and 2 (19) have been identified as targets of p53. In this study we show that TRAIL is induced by TNF and treatment with phorbol ester under critical involvement of NEMO/IKK␥, an essential component of the NF-B signaling pathway, may account for some of the apoptotic effects described for NF-B in the literature (20 -24).
RNase Protection Assay-Cells (10 ϫ 10 6 ) were treated as indicated, washed with ice-cold phosphate-buffered saline, and total RNAs were isolated with the RNA INSTAPURE kit (Eurogentech, Seraing, Belgium) according to the manufacturer's recommendations. Transcripts of the indicated genes were detected using the Multi-Probe template set hApo3c (PharMingen, Hamburg, Germany). Probe synthesis, hybridization and RNase treatment were performed with the RiboQuant Multi-Probe RNase Protection Assay System (PharMingen). Protected transcripts were resolved by electrophoresis on denaturing polyacrylamide gels (5%) and visualized on a PhosphorImager with the ImageQuant software (Molecular Dynamics, Sunnyvale, CA).
Western Blotting-For detection of TRAIL protein cells were treated as indicated, and after the given time supernatants were collected or lysates were prepared in RIPA buffer supplemented with a protease inhibitor mixture stock solution (Roche Molecular Biochemicals). For preparation of cell lysates, cell debris was removed by centrifugation (10,000 ϫ g, 10 min), and the protein concentration of the lysate was determined using the Bradford assay. To relieve Western blot detection of TRAIL in TNF-treated Jurkat cells, TRAIL was concentrated by immunoprecipitation with TRAIL-R2-Fc and protein A-Sepharose (Amersham Pharmacia Biotech). For this purpose Jurkart cells (25 ϫ 10 6 ) were lysed in 1 ml RIPA buffer and incubated for 1 h with 2 g/ml TRAIL-R2-Fc at 4°C, and finally TRAIL/TRAIL-R2-Fc complexes were precipitated with 25 l of protein A-Sepharose (1 h, 4°C) and washed three times in RIPA buffer. TRAIL/TRAIL-R2-Fc complexes, cell lysates (50 g), and cell supernatants were resolved by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes by electroblotting. Blots were blocked for 1 h at room temperature in Trisbuffered saline containing 0.05% Tween 20 and 3% (w/v) dry milk, washed, and incubated with anti-TRAIL mAb mix (mAbs RIK-2, 2G9, 2E11; each 0.3 g/ml) for at least 1 h at room temperature. TRAIL⅐mAb complexes were visualized with alkaline phosphatase-conjugated goat anti-mouse-IgG (Sigma) and nitro blue tetrazolium and 5-bromo-4chloro-3-indolyl phosphate as substrate.
Determination of NF-B Activity by Electrophoretic Mobility Shift Assay and Reporter Gene Analysis-For preparation of nuclear extracts 5 ϫ 10 6 Jurkat cells were seeded in 100-mm cell culture dishes and cultivated overnight. Next day cells were stimulated with TNF or phorbol 12-myristate 13-acetate and ionomycin (PMA/I) for the indicated time, washed once with cold phosphate-buffered saline, resuspended in 0.4 ml of buffer A (10 mM KCl, 10 mM HEPES, pH 7.9, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride) and incubated on ice for 20 min. Subsequently, 25 l of 10% Nonidet P-40 were added for 2 min, and nuclei were pelleted and resuspended in buffer B (20 mM HEPES, pH 7.9, 400 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol). After 20 min of shaking and subsequent centrifugation, the lysates containing the nuclear proteins in the supernatant were used for the electrophoretic mobility shift assay after protein determination (Bio-Rad) with bovine serum albumin as standard. High pressure liquid chromatography-purified NF-B-specific oligonucleotides (5Ј-ATCAGGGACTTTCCGCTGGGGACTTTCCG-3Ј) were end-labeled with [␥-32 P]ATP and electrophoretic mobility shift assays were performed by incubating 10 g of nuclear extracts with 4 g poly(dI-dC) in binding buffer (5 mM HEPES, pH 7.8, 5 mM MgCl 2 , 50 mM KCl, 0.2 mM EDTA, 5 mM dithiothreitol, 10% glycerol). The doublestranded, end-labeled, purified oligonucleotide probe (2 ϫ 10 4 to 5 ϫ 10 4 cpm) was added to the reaction mixture for 10 min at room temperature. The samples were separated by native polyacrylamide gel electrophoresis in low ionic strength buffer, and gels were analyzed using a PhosphorImager with the ImageQuant software (Molecular Dynamics). For monitoring PMA/I-and TNF-induced NF-B activation by reporter gene analysis, Jurkat cells (20 ϫ 10 6 in 800 l of RPMI medium, 10% fetal calf serum) were transiently transfected by electroporation (250 V, 1800 F) with a 3ϫ NF-B-luciferase reporter plasmid (30 g) in 4-mm electroporation cuvettes (Peqlab, Erlangen, Germany) using an Easyject plus Equibio electroporator (Peqlab, Erlangen, Germany). After 1 day of recovery, cells were split (50 ϫ 10 3 /well) into 96-well tissue culture plates and stimulated as indicated with a mixture of TNF (20 ng/ml), PMA/I (100 nM/1 g/ml) and the proteasome inhibitor MG132 (20 M), which inhibits NF-B activation. After 8 h cells were har-vested, washed once in phosphate-buffered saline, and lysed in luciferase lysis buffer (Promega, Mannheim, Germany), and finally luciferase activities were determined using a Lucy2 96-well luminometer (Anthos, Krefeld, Germany).
Cytotoxic Assay-Jurkat cells (50 ϫ 10 3 /well) were seeded in 96-well tissue culture plates in RPMI medium containing 10% hiFCS, and next day the indicated reagents were added. After an additional 18 h of culture metabolic activity was measured by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide method. Cross-linked TRAIL and FasL were generated by mixing the respective concentration of FLAG-tagged TRAIL or FasL with the anti-FLAG monoclonal antibody M2 (Sigma) to a final concentration of 1 g/ml antibody. After 15 min of incubation at room temperature the formed TRAIL-M2 or FasL-M2 complexes were transferred to the cells.

RESULTS AND DISCUSSION
In a search for apoptosis-related genes regulated by TNF or during T-cell activation, we analyzed Jurkat T-cells treated with TNF or a mixture of PMA/I by RNase protection analysis. Both TNF and PMA/I induced up-regulation of TRAIL mRNA but with clearly distinct kinetics (Fig. 1A). While TNF-induced half-maximal TRAIL expression already after 1 h and reached a plateau after 3 h (Fig. 1A, right panel), TRAIL induction by PMA/I was delayed for several hours and reached maximum levels after 15 h (Fig. 1A, left panel). Similarly delayed kinetics of TRAIL induction was also found in DII/23-7 T-cells (Fig. 1B). There was no evidence for an up-regulation of TRAIL in the B cell lines BJAB and Raji suggesting that TRAIL up-regulation by TNF and PMA/I is cell type-specific (Fig. 1C). PMA/I and TNF showed a capacity to induce TRAIL expression in the same range as the well established TRAIL inducers IFN␣/␤/␥ (Fig. 1D). In all cases two protected TRAIL-specific fragments of slightly different sizes were obtained, suggesting that TRAIL-mRNA exists in at least two different splice forms. In accordance with the barely detectable TRAIL-mRNAs levels in untreated Jurkat cells (Fig. 1A), we were unable to detect TRAIL protein in non-stimulated cells or cell culture supernatants derived thereof. However, three anti-TRAIL-mAb-reactive protein bands were detectable by Western blotting in lysates of PMA/I-stimulated Jurkat cells whereas no TRAIL was detectable in the supernatants. When Jurkat cells were treated with PHA during the last 30 min of PMA/I stimulation a significant amount of all three TRAIL-mAb-reactive proteins were found in the supernatant (Fig. 1E). TRAIL protein was also detectable in lysates of TNF-treated cells after immunoprecipitation with TRAIL-R2-Fc and protein A-Sepharose (Fig.  1F). The size of the smallest anti-TRAIL mAb reactive band (19 kDa, Fig. 1, E and F) corresponds to the size of the already described soluble TRAIL product comprising the extracellular receptor-binding domain of the ligand (27). The sizes of the two slower migrating anti-TRAIL-mAb reactive proteins (32 and 34 kDa) are in good accordance with that of complete TRAIL, suggesting that these two bands represent two isoforms of full-length TRAIL. PHA-stimulation lead to the secretion of all three anti-TRAIL-mAb reactive proteins, suggesting that PHA induces the release of TRAIL containing microvesicles as described elsewhere (28) rather than inducing proteolytic processing of cell-bound TRAIL. In PMA/I-treated T cells we also observed up-regulation of FasL and TRAIL-R2 mRNA, and in the B-cell lines PMA/I induced TRAIL-R1 and TRAIL-R2 (Fig.  1). Analyses of Jurkat cells exclusively stimulated with PMA or ionomycin revealed that PMA alone is sufficient for the induction of TRAIL (data not shown).
TNF and PMA/I are both potent inducers of the NF-B pathway in Jurkat cells ( Fig. 2A). We therefore analyzed whether this pathway contributes to TRAIL induction in T cells. For this purpose, we analyzed a mutant Jurkat cell line deficient in expression of IKK␥/NEMO (25), an essential part of the NF-B-inducing I-B kinase (IKK) complex (29). These IKK␥/NEMO-deficient Jurkat cells are unable to respond with NF-B activation to a variety of NF-B inducers including TNF and PMA, but show normal activation of NF-AT and AP1 (30). As shown in Fig. 2B both TNF and PMA/I-induced up-regulation of TRAIL was completely prevented in this mutant Jurkat cell line. However, in a clone derived from the IKK␥/NEMOdeficient Jurkat cell line, in which NF-B responsiveness had been restored by retransfection with an IKK␥ expression construct (30), responsiveness of the TRAIL gene toward PMA/I and TNF was partly restored (Fig. 2A, right panel). Further, PMA/I-induced up-regulation of TRAIL-R2, but not FasL induction, was significantly inhibited in the IKK␥/NEMO-deficient Jurkat cells (Fig. 2A). Involvement of the NF-B pathway in PMA/I-induced TRAIL-R2 expression is in good agreement with some recent studies showing a critical role of the cRel subunit of NF-B in expression of TRAIL-R1 and TRAIL-R2 (24). Lack of requirement of the NF-B pathway for FasL gene induction in IKK␥/NEMO-deficient Jurkat cells has already been described elsewhere (30) and suggests that the death ligands TRAIL and FasL are up-regulated during T cell activation by at least partially distinct signaling pathways. In the case of TNF-induced TRAIL expression we analyzed the involvement of the NF-B pathway in a second independent genetic model, i.e. Jurkat T-cells deficient in RIP expression (26). The RIP kinase was originally isolated as a Fas interacting death domain-containing kinase, but is also part of the TNF-R1 signaling complex and plays an essential role in TNF-R1-mediated NF-B activation (31). Lack of RIP expression impaired TNF-induced up-regulation of TRAIL (Fig. 2C) but had no significant effect on PMA/I-induced up-regulation of TRAIL and TRAIL-R2 (data not shown).
Thus, our data clearly argue for an essential involvement of IKK␥/NEMO and therefore most likely the NF-B pathway in TNF-and PMA/I-induced TRAIL expression. However, the data described so far let open the question whether TRAIL induction occurs directly or not. We therefore analyzed the capability of TNF and PMA/I to induce TRAIL-mRNA in the presence of high concentrations of the translation inhibitor cycloheximide (CHX). While PMA/I-induced TRAIL expression was almost completely inhibited in the presence of 50 g/ml CHX (Fig. 3A) TNF-mediated up-regulation was not affected by this treatment (Fig. 3B). Thus, while PMA/I engages TRAIL transcription indirectly, TNF induces TRAIL expression directly. As interferons are major inducers of TRAIL, we looked for an involvement of endogenous interferons in PMA/I-induced up-regulation of TRAIL. However, we found no evidence for an involvement of interferons in PMA/I-induced up-regulation of TRAIL. So, neutralizing mAbs against IFN␣ receptor chain 2 or interferon-␤ failed to inhibit TRAIL-induction (data not shown). Moreover, IFN␥ is most likely also not involved in PMA/I-induced up-regulation of TRAIL because even saturating concentrations of recombinant IFN␥ only marginally upregulate TRAIL-mRNA (Fig. 1E). As both PMA/I and TNF rapidly activate NF-B under essential involvement of IKK␥/ NEMO (Fig. 2, A and B), the differential need for protein synthesis challenges the concept of NF-B dependent up-regulation of TRAIL (see above). A possible explanation could be the existence of a second pathway that synergistically acts with the

FIG. 1. Treatment with PMA and ionomycin or TNF induces TRAIL in T but not B cell lines. A-C, Jurkat (A) and DII/23-7 (B) T cells and BJAB (C)
and Raji (C) B cells, respectively, were treated for 6 h or the indicated times with a mixture of PMA (100 nM) and ionomycin (1 g/ml) or TNF (20 ng/ml). Total RNAs were isolated for RNase protection analysis and 10 g of each RNA sample were analyzed with a Multi-Probe template set to detect the indicated mRNAs. L32 and glyceraldehyde-3-phosphate dehydrogenase were included in the template set as internal controls. D, Jurkat cells were treated for 6 h with IFN␣ (50 ng/ml), IFN␤ (15 ϫ 10 3 units/ml), IFN␥ (20 ng/ml), TNF (20 ng/ml), or PMA/I (100 nm/1 g/ml). Total RNAs were isolated and analyzed as in A-C. E, Jurkat cells were stimulated for 15 h with PMA/I or were left untreated. In the indicated groups PHA was added during the last 30 min of incubation. Cleared supernatants and cell lysates (50 g) were prepared and separated on a 15% polyacrylamide gel. F, Jurkat cells were stimulated for 15 h with TNF (25 ng/ml) or were left untreated. Cells were lysed in 1 ml RIPA buffer as described under "Materials and Methods," and TRAIL was concentrated by immunoprecipitation with TRAIL-R2-Fc (2 g/ml) and protein A beads (25 l/ml). Washed immunoprecipitates were separated on a 15% polyacrylamide gel. TRAIL was detected by Western blotting using a mixture of TRAIL-specific mAbs (E, F).
NF-B pathway to induce TRAIL and is directly induced by TNF but only indirectly triggered by PMA/I. However, we cannot rule out the possibility that TRAIL induction occurs by an IKK␥/NEMO-dependent pathway being distinct from the NF-B pathway, which becomes differentially activated by PMA/I and TNF. With respect to the latter possibility, we found no evidence for a NF-B dependence of an IFN-responsible reporter gene construct containing about 1 kilobase of the TRAIL promoter (data not shown). Although PMA/I stimulation induced the cytotoxic ligand TRAIL and one of its corresponding death receptors the viability of the PMA/I-treated cells was not affected 1 day after treatment. However, upon prolonged incubation time the PMA/I-treated cells started to die. This transient apoptosis resistance observed in the early phase of PMA/I treatment might be due to the concomitant induction of anti-apoptotic factors. Indeed, we have recently shown that PMA/I induces up-regulation of cFLIP, a potent inhibitor of death receptor signaling (Ref. 32, Fig. 3C). Moreover, 6 h of pretreatment with PMA/I protected Jurkat cells from a subsequent 16-h challenge with TRAIL or FasL with an efficiency comparable with the broad range caspase inhibitor Z-VAD-fmk (Fig. 3D). While TRAIL-induced apoptosis was almost completely blocked by z-VAD-fmk and PMA/I pretreatment, FasL-induced cell death was only partially inhibited by both treatments. This is in good accordance with some recent findings showing that in the absence of CHX, TRAIL induces cell death only via the caspase-dependent Fas-associated death domain protein/caspase-8 pathway, whereas FasL in addition signals necrotic cell death via a caspase-independent RIP-mediated pathway (33,34). Although PMA/I treatment efficiently blocked TRAIL-induced apoptosis, PMA/I treatment per se induced cell death in a significant fraction (20 -40%) of the cells (Fig. 3D and Ref. 30). This per se cytotoxicity of PMA/I might reflect killing of a subpopulation of Jurkat cells in which unbalanced amounts of TRAIL/FasL and FLIP (and maybe additional anti-apoptotic factors) were induced. A, Jurkat cells were challenged with TNF (20 ng/ml) or PMA/I (100 nM/1 g/ml), were harvested after the indicated time, and analyzed for NF-B activation by electrophoretic mobility shift assay. For further verification of TNF-and PMA/I-induced NF-B activation, Jurkat cells were transfected with 30 g of a NF-B reporter gene construct by electroporation. After splitting and 1 day recovery, the cells were challenged with TNF or PMA/I. As a control some groups were treated with the proteasome inhibitor Mg132 (20 g/ml), which interferes with NF-B activation. After an additional 8 h cells were lysed and analyzed for luciferase activities. B, the parental Jurkat cell line, RIP-and IKK␥-deficient clones derived thereof, as well as an IKK␥-reconstituted clone of the latter, were treated for 6 h with PMA/I or TNF (20 ng/ml). Total RNAs were isolated for RNase protection analysis and 10 g of each RNA sample were analyzed with a Multi-Probe template set to detect the indicated mRNAs. L32 and glyceraldehyde-3-phosphate dehydrogenase were included as internal controls.
FIG. 3. PMA/I-but not TNF-induced up-regulation of TRAIL requires protein synthesis. A, Jurkat cells were stimulated for the indicated times with PMA/I in the presence or absence of 25 g/ml CHX and were analyzed by RNase protection assay analysis. B, Jurkat cells were stimulated for 6 h with TNF in the presence or absence of 25 g/ml CHX and were analyzed by RNase protection assay analysis. C, Total RNA of PMA/I-treated (6 h) and untreated Jurkat cells were analyzed using a customer template set containing TRAIL and cFLIP-specific probes among others. D, Jurkat cells were seeded in 96-well microtiter plates at a density of 50 ϫ 10 3 cells/well and treated overnight in triplicates with the indicated concentration of cross-linked FasL or cross-linked TRAIL with (open circles) or without (closed circles) z-VADfmk (10 M). In addition, cells were challenged with TRAIL or FasL after 6 h of PMA/I treatment (open squares). Finally, viable cells were quantified by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide method.