c-Jun NH2-terminal Kinase Activation Leads to a FADD-dependent but Fas Ligand-independent Cell Death in Jurkat T Cells*

Persistent c-Jun NH2-terminal kinase (JNK) activation induces cell death. Different mechanisms are ascribed to JNK-induced cell death. Most of the JNK-apoptosis studies employ stress stimuli known to activate kinases other than JNK. Here we used overexpression of mitogen-activated protein kinase kinase 7 (MKK7) to activate selectively JNK in T lymphoma Jurkat cells. Similar to that reported previously, Fas ligand (FasL) expression was up-regulated by JNK activation. Dominant negative-FADD and caspase-8 inhibitor benzyloxycarbonyl-Ile-Glu-Thr-Asp effectively inhibited MKK7-induced cell death, supporting a major involvement of FADD cascade. However, MKK7-induced cell death was not prevented by antagonist antibody ZB4 and Fas-Fc, indicating that Fas-FasL interaction is minimally involved. Confocal microscopy revealed that persistent JNK activation led to clustering of Fas. Our results suggest that, in contrast to that reported previously, JNK alone-induced death in Jurkat cells is FADD-dependent but is not triggered by Fas-FasL interaction.

JNK 1 activation is always linked to cell death induced by stress. Apoptosis triggered by UV, ␥-irradiation, and cytotoxic drugs is correlated with activation of JNK, and the cell death is prevented by inhibition of JNK activation (1)(2)(3)(4)(5). The pivotal role of JNK is further illustrated by JNK activation induced by active mitogen-activated protein kinase kinase kinase 1 (MEKK1) (2,6), active Cdc42 (7), and apoptosis signal-regulating kinase 1 (8) that either initiates the apoptotic process or potentiates cell death triggered by low dose stress stimuli (9). JNK activation is implicated in growth factor deprivation-induced cell death (10 -12), in class I major histocompatibility complex ligation-triggered apoptosis (13), or possibly in anoikis (14). The critical role of JNK is also supported by the lack of apoptosis on hippocampal neurons in JNK3-deficient and in JNK1/JNK2 double knockout mice (15,16).
The exact molecular mechanism how JNK induces cell death remains largely elusive. Different apoptotic molecules have also been attributed to JNK-triggered cell death. Activation of c-Jun by JNK seems to mediate part of the apoptotic events (17). MEKK1 or c-Jun induces FasL expression and the subsequent FasL-Fas interaction and cell death (6,12,18,19). p53 and Bax may also mediate JNK-induced apoptosis following p75 neurotrophin receptor activation (11). Alternatively, apoptosis could be induced by translocation of JNK into mitochondria followed by phosphorylation and inactivation of Bcl-2 and Bcl-xL (20,21). In addition, the UV-induced mitochondrial death pathway is abrogated in the absence of JNK, further supporting mitochondria as the target of JNK (22).
Most of JNK-inducing signals such as UV and cytotoxic drugs activate signals other than JNK. Even for the selective expression of MEKK1, Cdc42, or apoptosis signal-regulating kinase 1, activation of JNK is accompanied by stimulation of p38 and/or IB kinase. In addition, the contribution of p38 to stress-activated apoptosis has been demonstrated (5,10,23,24). In this study, we used transient expression of MKK7 to activate JNK in Jurkat T cells. MKK7 selectively activates JNK but not other kinases (25)(26)(27)(28)(29). We confirmed the previous notion that JNK activation leads to increased FasL expression in Jurkat cells. Blockage of FADD-initiated apoptotic pathway effectively prevented JNK-induced cell death. However, blockage of FasL-Fas interaction by antagonizing antibody or Fas-Fc did not affect MKK7-induced apoptosis, suggesting that FasL is minimally involved in JNK-mediated cell death. Our results clearly suggest that JNK induces apoptosis by a FADD-dependent but FasL-independent mechanism in Jurkat cells.
Transfection-Jurkat T cells (1 ϫ 10 7 ) were washed and resuspended in 0.6 ml of RPMI medium containing 1% glucose, 10% fetal calf serum, and 10 -20 g of plasmid DNA. The electroporation was performed in Bio-Rad Gene Pulser II at 260 mV and 975 microfarads. The cuvette was left on ice for 15 min, washed twice with phosphate-buffered saline, and incubated for the indicated time for cell death and biochemical analysis.
Protein Kinase Assay-Jurkat T cells were transfected with pcDNA3, ⌬MEKK1, or pcDNA3-MKK7␣. Cell lysates were prepared 24 h after transfection, and 100 -200 g of lysate was precipitated with 1 l of anti-JNK1 antibody 101 (31) or anti-p38 (32), followed by 20 l of protein A-Sepharose. The kinase activity of the immune complexes was determined by using GST-c-Jun-(1-79) or myelin basic protein as substrates. The reaction mixtures were resolved on SDS-polyacrylamide gel electrophoresis, followed by autoradiography, and quantitated by PhosphorImager (Molecular Dynamics).
Cell Death Measurement-Apoptosis in bulk population was determined by propidium iodide (PI) staining. At the indicated times after treatment, cells were harvested and washed in phosphate-buffered saline twice and resuspended in hypotonic fluorochrome solution (50 g/ml PI, 0.1% sodium citrate, 0.1% Triton X-100) (33). Cells were placed at 4°C in the dark overnight, and DNA content was analyzed by FACScan (Becton Dickinson, Mountain View, CA). The fraction of cells with sub-G 1 DNA content was assessed using the CELLFIT program (Becton Dickinson). For apoptosis in cells transiently transfected with MKK7␣ or ⌬MEKK1, green fluorescence protein expression vector pGreen Lanten-1 (Life Technologies, Inc.) was cotransfected. Cells were harvested at the indicated times, fixed with paraformaldehyde, and terminal dUTP nick-end labeling reaction was performed using FlowTACS kit (R & D Systems). The incorporated biotin-dUTP was labeled with Tri-Color-streptavidin (Caltag, Burlingame, CA). The green cells (GFP-positive) were then gated on FACScan, and the fraction of cell stained with Tri-Color was quantitated. Alternatively, PI staining was also used to determined the fraction of subdiploid cells in GFP-positive population.

Expression of Active MKK7␣ and MEKK1
Led to JNK-dependent Cell Death-Jurkat cells were transfected with active MKK7␣ or MEKK1 by electroporation. We chose electroporation because the transfection efficiency was close to 30% as determined by cotransfection with GFP (Fig. 1A, R1). To assess the cell death induced by JNK activation, the population expressing GFP was gated in fluorescence-activated cell sorter, and the fraction of apoptotic cells labeled with biotin-dUTP was quantitated (Fig. 1A). Activation of JNK by MKK7␣ and MEKK1 in Jurkat cells led to 50% death 24 h after transfection (Fig. 1B). We next examined whether the cell death observed was JNK-dependent. MKK7␣ and MEKK1 were equally effective in the JNK induction ( Fig. 2A). The specificity of MKK7␣ was further confirmed by its inability to activate p38 mitogenactivated protein kinase, in contrast to the effective induction of p38 by MEKK1 and MKK3 (Fig. 2B). The activation of JNK by MEKK1 or MKK7␣ was prevented by cotransfection of SEK-AL ( Fig. 2A). The exact mechanism how SEK-AL prevents MKK7␣-induced activation of JNK is not completely clear. Presumably, the binding of SEK-AL with JNK (34) would compete with the interaction of JNK with MKK7␣. Inhibition of JNK activation by SEK-AL prevented MKK7␣-or MEKK1induced cell death (Fig. 2C), indicating that the observed cell death is JNK-specific. Because MKK7␣ is a more specific activator of JNK, in the following experiments mainly the results of MKK7␣ transfection are shown. In all the criteria evaluated, MEKK1-induced apoptosis displayed an identical character.
MKK7␣ Expression-induced FasL Expression-Consistent with previous reports (6,18), there was a significant increase of FasL expression after transfection of MKK7␣ and MEKK1 as determined by immunoblots (Fig. 3A). Despite a similar degree of JNK activation (Fig. 2), the extent of FasL expression was higher for MEKK1 transfection than MKK7␣ transfection. A likely cause is because MEKK1 also activates NF-B and p38 mitogen-activated protein kinase, and both contribute to activation of the FasL promoter (35)(36)(37)(38). Despite the increase of total cellular FasL, there was little increase in the surface FasL  1-79). B, p38 activities were determined by precipitation with anti-p38 (32) and phosphorylation of myelin basic protein (MBP). C, apoptosis was determined 24 h after transfection as described in Fig. 1. TC-SA, Tri-Color-streptavidin. expression after MKK7␣ expression (Fig. 3B, dark curve). As a positive control, TPA/A23187 treatment significantly promoted the surface FasL expression (Fig. 3B, light curve). The expression of Fas is already high in Jurkat cells. Expression of active MEKK1 or MKK7␣ added little to the surface Fas expression (not shown).
MKK7␣-induced Cell Death Was Not Prevented by Fas-antagonizing Antibody and Fas-Fc-To determine the role of Fas-FasL engagement in MKK7␣-induced cell death, the antagonistic anti-Fas antibody ZB4 was used. Preincubation with ZB4 (250 ng/ml) effectively suppressed CH11-induced apoptosis of Jurkat cells (Fig. 5A). In contrast, the extent of Jurkat cell death resulting from MKK7␣ overexpression was indistinguishable in the presence or absence of ZB4 (Fig. 5B). We also used soluble Fas-Fc fusion protein to block the interaction of Fas and FasL. Fas-Fc (200 ng/ml) prevented FasL-induced cell death (Fig. 5A), yet Fas-Fc failed to interfere with MKK7␣induced apoptosis in Jurkat cells (Fig. 5B). Because ⌬MEKK1 induced higher expression of FasL (Fig. 3), we further examined whether ⌬MEKK1-induced cell death could be inhibited by Fas-Fc or ZB4. Neither Fas-Fc nor ZB4 prevented apoptosis induced by ⌬MEKK1 (not shown). The observations that Fas-Fc and ZB4 did not protect Jurkat cells from apoptosis suggest that the apoptosis induced by JNK is not mediated through Fas-FasL interaction.
MKK7-induced Fas Aggregation on Jurkat Cells-Stress stimuli such as UV, cycloheximide, cisplatin, etoposide, vinblastine, and doxorubicin may induce cell death by triggering Fas clustering in a FasL-independent manner (43)(44)(45)(46). We also examined the surface Fas distribution on a macroscopic level before and after JNK activation using confocal laser scanning microscope. Fas was evenly distributed on the surface of the untransfected Jurkat cells (Fig. 6A). Treatment with sFasL led to increased aggregation of Fas on the surface of Jurkat cells (Fig. 6B). For cells transfected with MKK7␣, as those marked by GFP expression (Fig. 6C), there was a similar increased clustering of Fas on the surface of Jurkat cells as compared with the nearby untransfected cells (Fig. 6D). Therefore, constitutive MKK7␣ expression promotes the aggregation of the surface Fas. DISCUSSION JNK has been implicated as the major mediator of cell death induced by stress (1-3), yet the exact apoptotic mechanism triggered by JNK is not fully understood. In the present study, we used a JNK-specific activator MKK7␣ to induce JNK and the subsequent apoptosis. Because MKK7␣ selectively activates JNK (25)(26)(27)(28)(29), we were able to address death specifically activated by JNK in the absence of other signaling such as p38 (Fig. 2B). Our results illustrate that JNK activation induced FasL expression (Fig. 3). The requirement of Fas-FADD pathway is supported by the inhibition of MKK7-triggered cell death by DN-FADD, FLIP, and Z-IETD (Fig. 4). We have also demonstrated, likely for the first time, that JNK-induced cell death in Jurkat cells is FasL-independent (Fig. 5).
Our observations that Fas-Fc and ZB4 were unable to prevent MKK7␣-induced cell death (Fig. 5) are in direct contradiction to the report of Faris et al. (6) that inducible expression of MEKK1 led to cell death which was prevented by soluble Fas and Fas antagonist antibody. We speculate the difference between their study and ours is likely due to the levels of Fas and FasL expression. In the study of Faris et al. (6), the inducible activation of MEKK1 in Jurkat led to an increase of surface FasL levels by 50-fold. This was accompanied by a 10-fold increase of surface Fas (see Fig. 5  are highly unphysiological, as judged from their report that there was a mere 50% increase of surface FasL with a 20% increase of surface Fas expression when Jurkat cells were activated with TPA/A23187 (6). In the present study, despite an induction of total FasL content, surface FasL levels in Jurkat cells were minimally altered by JNK activation (Fig. 3), and cell death proceeded in the absence of FasL binding. Fas-FasL interaction apparently is not essential for JNK-induced cell death in Jurkat cells. We have also repeated our observation in another T lymphoma EL4 (not shown) and reached an identical conclusion.
FADD-dependent but FasL-independent cell death is triggered by UV, cycloheximide, cisplatin, etoposide, vinblastine, and doxorubicin through induced clustering of Fas (43)(44)(45)(46), leading to the association of FADD with Fas and the subsequent activation of caspase-8. Fas aggregation-induced cell death is suppressed by DN-FADD or FADD antisense (44 -46). Similarly, MKK7-induced apoptosis is sensitive to inhibition by DN-FADD, FLIP, and Z-IETD (Fig. 4) and MKK7 overexpression triggered Fas clustering (Fig. 6), suggesting that JNKactivated apoptosis is mediated by Fas aggregation. We do not know the exact cellular process between the JNK activation and Fas clustering. Changes in the microtubule cytoskeleton may activate MEKK1 and JNK activation (47), but whether persistent JNK activation could induce reorganization of the microtubule cytoskeleton that promotes Fas association remains to be determined. It may also be noted that our results on JNK-induced Fas oligomerization are compatible with the recent observation that Fas receptor is in trimer status before FasL binding (48). This is supported by the observation that the engagement of Jurkat cells by FasL led to a visible aggregation of Fas under the microscope (Fig. 6D) as compared with untreated cells, supporting that the Fas aggregation observed by others and us (43)(44)(45)(46) is in a macroscopic level. Therefore, Fas is trimerized in resting Jurkat cells but further aggrega-tion is induced by stress stimuli or JNK activation.
Our results may also help resolve the controversy on the involvement of FasL in chemotherapy-induced apoptosis. Despite the earlier reports that DNA-damaging agent-induced cell death is inhibited by Fas antagonizing antibody or Fas-Fc (32,49), many studies have shown that apoptosis induced by UV and genotoxic drugs is not prevented by neutralizing antibodies for Fas and FasL (45, 50 -52). Our demonstration that JNK triggers FADD-dependent apoptosis through induction of Fas aggregation, which does not require FasL in Jurkat cells, suggests one of the FasL-independent mechanisms that may be involved in the killing of Fas-sensitive cancer cells by chemotherapeutic agents is known to activate JNK.
Our observation on JNK-induced apoptosis in Jurkat cells, however, is not necessarily applicable to all types of cells. In neuronal cells, with the well established role of JNK and c-Jun in apoptosis induction (15,16), Fas-FasL interaction mediates part of JNK-activated apoptosis induced by growth factor withdrawal (12). The difference in the degree of the FasL participation in JNK-induced cell death could be due to the different nature between T lymphocytes and cerebella granule neurons.
Notably, a pivotal role of FADD in JNK-triggered apoptosis does not exclude other apoptotic mechanisms. MKK7-induced cell death was largely inhibited by FADD and Z-IETD (Fig. 4), yet a small fraction of death could still be detected even in the excess of DN-FADD and Z-IETD (not shown), suggesting the presence of apoptotic pathway not mediated by caspase-8. JNK is known to induce apoptosis in Fas-independent manner by phosphorylation and inactivation of Bcl-2 and Bcl-xL (20,21). Together with the results from the present study, persistent activation of JNK is capable of triggering apoptotic pathways initiated by both mitochondria (22) and death receptor. We speculate that the exact contribution from mitochondria and Fas pathway in JNK-mediated apoptosis would be determined by variables such as type of stress, type of cell, expression of Fas, and cellular sensitivity to Fas. Further characterization will help understand the exact molecular process triggered by JNK apoptotic signal in different cells.