Dominant-Negative Fas Mutation Is Reversed by Down-expression of c-FLIP

Fas triggering by agonistic antibodies or by its cognate ligand, FasL, induces apoptotic cell death, whereas mutation in the Fas death domain is associated with lymphoma progression. On prolonged culture in the presence of an agonistic anti-Fas antibody, we raised a Jurkat cell line resistant to agonistic antibodies but still sensitive to soluble FasL, which carried at the heterozygous state, a point mutation into the Fas death domain. Down-modulation of c-FLIP expression reversed the blockade of the Fas pathway. We show that the activation threshold for the Fas receptor is more easily overcome by multimeric FasL than by agonistic antibodies and that the increase of this threshold due to mutation in the Fas death domain can be overcome by acting on a downstream effector of the Fas signal, c-FLIP. These findings put forward a new approach to eradicate Fas-resistant tumor cells. [Cancer Res 2007;67(1):108-15]


Introduction
Fas is a transmembrane receptor that belongs to the tumor necrosis factor (TNF) receptor (TNFR) superfamily.Its cognate ligand, FasL, is a transmembrane protein found in a soluble trimeric form after cleavage by metalloproteinases (1,2).This soluble homotrimeric FasL complex is not able to induce cell death unless it is produced as a homohexameric complex (3).
Fas carries an intracellular conserved stretch of 80 amino acids called the death domain, which serves as a docking platform to trigger an intracellular signal.Indeed, on binding of FasL or multivalent agonistic antibodies, the Fas death domain recruits the adaptor molecule Fas-associated death domain protein (FADD), which in turn aggregates the initiator caspase-8 and caspase-10.This intracellular complex is called death-inducing signaling complex (DISC).The fact that caspases are brought together in close vicinity leads to their autocleavage and activation of the downstream apoptotic signal.In parallel, Fas triggers nonapoptotic signals, such as nuclear factor-nB (NF-nB) and mitogen-activated protein kinase pathways (4).A caspase-like protein termed c-FLIP (cellular FADD-like interleukin-1h-converting enzyme inhibitory protein) impairs the Fas-mediated cell death signal (5).c-FLIP is expressed mainly in a long (c-FLIP L ) form and a short spliced form (c-FLIP S ).The short form consists of exclusively two tandem repeats of death effector domain (DED), which inhibits procaspase-8 activation (6).The longer isoform contains two DED in its NH 2 -terminal region and shares extensive homology with the caspase-8 catalytic domain in its COOH-terminal region.However, c-FLIP L sequence lost the conserved Cys residue in the catalytic site and a closely positioned His residue, which are essential for the caspase-8 catalytic activity (7).Recently, the inhibitory role of c-FLIP L has been challenged because depending on its expression level, Chang et al. (8) showed that it can improve the apoptotic signal.
In humans, heterozygous expression of a Fas mutation interferes dominantly to inhibit the apoptotic signal because the transduction of Fas-mediated death signal requires a specific stoichiometric homotrimeric protein complex (9) that is inhibited by the incorporation of a mutated receptor.Mutations in the Fas death domain have been described both in mice (Lpr cg mice; ref. 10) and in patients developing lymphoma (11,12) or suffering of type Ia autoimmune lymphoproliferative syndrome (ALPS; refs.[13][14][15].Dysregulation of the Fas apoptotic pathway has been proposed as a mechanism of oncogenesis by providing a survival advantage to potentially malignant cells and has also been implicated in the relapse after chemotherapeutic treatment (16).
We reported previously that agonistic antibodies and FasL trigger a qualitatively different Fas signaling pathway (17).In the context of a mild induction of Fas signal (Fas mutation-expressing cell and agonist antibody-mediated Fas signal), herein we observe that c-FLIP acts as a potent checkpoint molecule and that its down-expression can restore cell death in Fas mutant-expressing cells.
Cell lines and stable transfectants.T leukemia Jurkat cell line called Jurkat 77 was obtained from Dr. P. Anderson (Dana-Farber Cancer Institute, Boston, MA).The murine Fas-deficient WR19L cell line was kindly provided by Prof. S. Nagata (Osaka Bioscience Institute, Osaka, Japan; ref. 20).Cells were cultured in RPMI 1640 supplemented with 8% heat-inactivated FCS and 2 mmol/L L-glutamine in a 5% CO 2 incubator at 37jC.Cells (5 Â 10 6 in 0.3 mL) were electroporated with 10 Ag vector at 300 V with one pulse (10 ms duration) using the BTM 830 electroporation generator (BTX Instrument Division, Holliston, MA).Twenty-four hours later, G418 (Life Technologies, Cergy-Pontoise, France) was added to the cultures at final concentrations of 2 mg/mL for Jurkat and 1.5 mg/mL for WR19L.Then, the G418-resistant cells were cloned by limiting dilution.Transient transfections of Jurkat cells were done identically, except that 24 h after transfection, the living cells were isolated using a Ficoll gradient centrifugation method and immediately used in the different assays.
Plasmid construction.The pEGFP plasmid was purchased from Clontech (Mountain View, CA).The pcDNA3-FlagFas plasmid was kindly provided by Dr. Ruberti (Institute of Cell Biology, Rome, Italy) and described elsewhere (21).Fas Q257K mutation was generated using the pcDNA3-Fas wild-type (WT) vector and the other mutants were generated using pcDNA3-FlagFas plasmid.These templates were modified using a site-directed mutagenesis kit following the manufacturer's protocol (QuickChange II, Stratagene, La Jolla, CA).
Detergent lysis and Western blot analysis.Cells were lysed for 30 min at 4jC in lysis buffer [25 mmol/L HEPES, 1% Triton X-100, 150 mmol/L NaCl (pH 7.4)] supplemented with a mix of protease and phosphatase inhibitors (Sigma, St. Louis, MO).The lysate was centrifuged at 15,000 rpm for 10 min and protein concentration in the supernatant was determined using the bicinchoninic acid method (Sigma) according to the manufacturer's protocol.Proteins were separated by SDS-PAGE on 12% gels in reducing conditions and transferred to a polyvinyldifluoride membrane (Amersham, Buckinghamshire, England).The membrane was blocked for 1 h with TBST [50 mmol/L Tris, 160 mmol/L NaCl, and 0.1% Tween 20 (pH 8)] containing 5% dried skimmed milk, and all subsequent steps were done in this buffer.The indicated specific antibody was then incubated overnight at 4jC.After washes, the peroxidase-labeled anti-mouse (Clinisciences, Montrouge, France), anti-goat (Vector Laboratories, Burlingame, CA), or anti-rabbit (Zymed, Invitrogen, Cergy Pontoise, France) secondary antibody was added for 1 h.Then, the proteins were visualized with the enhanced chemiluminescence (ECL) substrate kit (Pierce, Brebieres, France).
DISC analysis.Jurkat cells (10 8 ) were either treated with 1 Ag/mL anti-Fas mAb clone APO-1-3 for 15 min at 37jC, washed once in cold PBS, and lysed in lysis buffer (15-min condition) or lysed before anti-CD95 was added (0-min condition).Protein A-Sepharose beads (Sigma) were then added to immunoprecipitate Fas and bound proteins for 3 h at 4jC.The immunoprecipitates were washed four times with 1 mL lysis buffer.Immune complexes were separated by SDS-PAGE.
Flow cytometry analysis.Cells were washed with PBS/bovine serum albumin and incubated for 30 min at 4jC with the FITC-conjugated goat anti-mouse IgG.After washing, cells were resuspended in 0.2 mL PBS and immediately analyzed with a FACSCalibur flow cytometer (Becton Dickinson, Mountain View, CA).
Quantification of DNA fragmentation as a specific measure of apoptosis was carried out by nucleus staining with propidium iodide as described elsewhere (22).Briefly, 50,000 cells were treated for 18 h with the indicated stimulus, centrifuged, and resuspended in 300 AL buffer containing 0.1% sodium citrate, 0.1% Triton X-100, and 50 Ag/mL propidium iodide for 4 h.Cells were directly analyzed by flow cytometry and sub-G 1 population was considered as DNA-fragmented cells.

Results
Selection of agonistic anti-Fas antibody-resistant cells.Treatment of Jurkat cells for 24 h with the anti-Fas IgM agonistic antibody 7C11 at a concentration that triggers maximal cell death (100 ng/mL) reproducibly allowed a few percentage (5-10%) of cells to escape death (see Fig. 1A).On the other hand, the whole population of Jurkat cells was sensitive to high concentration of FasL because all cells died in the presence of 20 ng/mL sFasL (Fig. 1A).To isolate the subpopulation of Jurkat cells sensitive to sFasL and resistant to antibody-mediated cell death, the culture of these cells was prolonged in the presence of the mAb.After 3 weeks, we obtained a resistant population from which we isolated clones by limiting dilution.One clone raised was completely resistant to cell death mediated by the 7C11 agonistic antibody (Fig. 1C).Nevertheless, it expressed the same amount of Fas as the parent cell line as determined by flow cytometry (Fig. 1B) as well as similar quantity of the DISC components Fas, c-FLIP, FADD, and caspase-8 and of the mitochondrion-dependent cell death inhibitors Mcl-1 and Bcl-2 when compared with the parent cell line (Fig. 1B).TRAIL, as FasL, belongs to the TNF family and it induces cell death through engagement of receptors termed death receptor (DR) 4 [TRAIL receptor 1 (TRAIL-R1)] and DR5 (TRAIL-R2).It has been described that these receptors trigger cell death via a FADD-dependent apoptosis pathway similar to the Fas signal (23,24).The acute T leukemia cell line Jurkat does not express DR4 and the membrane expression of DR5 was unchanged in the resistant clone when compared with the parent cell line (data not shown).Both cell lines displayed an identical sensitivity to TRAILinduced apoptosis, which indicated that the intracellular apoptotic signaling pathway was fully functional in our 7C11-resistant cell line (Fig. 1C).
We therefore reasoned that the resistance phenotype could be associated to a specific mutation in the Fas receptor.Indeed, sequencing the PCR-amplified Fas cDNA from this Jurkat clone revealed the presence of two distinct nucleotides at position 1011, either a C as in the WT Fas cDNA or a A (Fig. 1D).To confirm that the selected Jurkat cell was heterozygous for the Fas allele, we cloned the amplified Fas cDNA and we sequenced several clones (n = 6).Half of the sequenced clones harbored the mutation (A in position 1011) and half had the WT sequence (data not shown), showing that the isolated 7C11-resistant Jurkat cell line was heterozygous for the Fas-mutated allele.This point mutation resulted in the replacement of the glutamine located at position 257 in the fourth a-helix of the Fas death domain with a lysine (Q257K; Fig. 1D).
Fas mutation is less efficient in blocking multimeric FasL than antibody-mediated cell death.The 7C11-resistant cell clone (Jurkat-R) was also completely insensitive to two additional agonistic antibodies, the IgM mAb CH11 and the IgG3 mAb APO-1-3 (Fig. 2A).In contrast, the cells remained sensitive to FasLmediated cell death (Fig. 2B).FasL killed more efficiently the parent than the resistant cells, but it is noteworthy that the whole population of resistant Jurkat cells remained fully sensitive to high amounts of sFasL (Fig. 2B).Therefore, in the Jurkat lymphoma cell line, the heterozygous expression of a death domain-mutated Fas abrogates apoptosis triggered by agonistic antibodies but does not inhibit FasL-induced signal.
To show that the newly identified Fas Q257K mutation behaves as Fas mutants described in patient suffering from ALPS type Ia or lymphoma (11,15,25), we generated some of these mutants by site-directed mutagenesis and transfected them in the parental Jurkat cell line.All the Fas mutants inhibited efficiently cell death mediated by the agonistic antibody 7C11 but not by sFasL (Fig. 2C).This finding indicated that the phenotype observed with Fas Q257K expression was not an exception and can be generalized to the Fas death domain mutants encountered in various patients.
A potential target responsible for the 7C11-mediated cell death inhibition was the Fas-associated phosphatase-1 (FAP-1).Indeed, FAP-1 has been reported to protect cells against Fas-induced apoptosis (26) through its binding to the COOH terminus of Fas, distally to the death domain (27).To rule out FAP-1 involvement in the inhibition of the 7C11-induced cell death, we expressed a mutant described previously in a patient suffering from a diffuse large B-cell lymphoma.The cells from this patient harbored a nonsense Fas mutant termed L199X, which lacks the entire death domain and the COOH-terminal region (11).Jurkat cells were transiently transfected with this construct.Around 30% of the cells expressed the Flag-tagged Fas mutant (Fig. 2D).Following incubation with 7C11 or sFasL, most of the transfected cells (i.e., 55-70%) became resistant to 7C11-triggered cell death, whereas, in contrast, only 15% to 18% were resistant to sFasL (see Fig. 2D).Thus, L199X point mutations conferred a more pronounced dominant-negative effect toward the antibody-mediated cell death signal than toward sFasL, indicating that FAP-1 phosphatase is not involved in the blockade of the 7C11-mediated apoptotic signal.
Homozygous Fas mutants do not transmit any signal.The murine lymphoma cell line WR19L that is devoid of endogenous Fas becomes sensitive to FasL and anti-Fas antibodies on de novo expression of the WT human Fas receptor (18).Therefore, we expressed the Q257K Fas mutant in this cell line to analyze whether it was able to transmit any apoptotic signal.Although our mutant could be expressed to levels comparable with that of WT Fas (Fig. 3A, top), it did not confer any sensitivity to agonistic antibody or sFasL (Fig. 3A, bottom).Therefore, Fas Q257K by itself was not able to induce any apoptotic signal on its own.Mutated Fas inhibits more efficiently antibody-mediated than FasL-mediated cell death.To rule out the possible implication of another genetic or functional intrinsic abnormality of the selected resistant cell line, as well as to formally show that the Q257K mutation was entirely responsible for the resistance to mAb 7C11, we overexpressed it in the parent Jurkat cell line.We selected among clones displaying increased Fas expression two distinct stable clones (named Q257K1 and Q257K3), which displayed high and low levels of Fas on the cell surface, respectively (Fig. 3B).As expected, both were dramatically resistant to mAb 7C11 (Fig. 3B), and resistance was correlated to the amount of mutated Fas on the cell surface (Fig. 3B).In addition, Fas Q257Kexpressing cells were less sensitive than the parent cell line to sFasL, but they were still killed efficiently when incubated with a high concentration of sFasL (Fig. 3B).Therefore, the resistance toward anti-Fas antibody-mediated cell death can be accounted for by the sole presence of Fas mutation and confirmed that the heterozygous status of this mutation was not able to completely block the signal triggered by FasL.One possible explanation would be that another FasL receptor was expressed by the T-cell line.We ruled out this hypothesis because high concentration of 7C11 competed and completely inhibited the apoptotic signal triggered by FasL treatment (data not shown).Therefore, agonistic antibody was less potent than soluble or membrane FasL (mFasL) to induce Fas-mediated cell death, and a weaker extracellular signal associated with a Fas mutation revealed the presence of a
Mutation of Fas inhibits both apoptotic and nonapoptotic signals.All these findings pointed out that in cells heterozygous for a Fas mutant, the antibody-induced apoptotic signaling pathway is blocked at some step and to define this step, we investigated the formation of Fas microaggregates, one of the earliest events in Fas signal transduction (28).The kinetic and the intensity of formation of these Fas-containing multimeric complexes were identical between parent and 7C11-resistant Jurkat cells (see Fig. 4A).Downstream to these microaggregates, we next analyzed the activation of caspase-8, the most proximal caspase in the Fas cascade.Caspase-8 activation was slightly delayed but not inhibited in the 7C11-resistant Jurkat cell treated with sFasL Figure 2. Fas-mediated cell death triggered by agonistic antibodies is completely inhibited but is mildly impaired on sFasL binding.Cells were incubated for 18 h with the indicated inducers, and cell death was measured using the MTT assay.A, the Jurkat-R cell line is not sensitive to CH11 or APO-1-3 anti-Fas agonistic antibodies.B, the Jurkat-R cell line remains sensitive to apoptosis induced by sFasL and mFasL.C, various Fas mutations reported in ALPS-or lymphoma-suffering patients inhibit more pronouncedly 7C11-induced than FasLinduced cell death.Jurkat cell line was cotransfected with the indicated constructs and a green fluorescent protein (GFP)containing vector at the DNA ratio of 3:1.Twenty-four hours after transfection, living cells were purified and treated or untreated with 0.5 Ag/mL anti-Fas or 100 ng/mL FasL.Percentage of cell death was measured as follows: 100 À (treated % GFP-expressing cells / untreated % GFP-expressing cells) Â 100.D, parent Jurkat cells were transiently transfected with empty (control ) or Flag-tagged Fas L199X-containing plasmid.Living cells (Ficoll gradient isolation) were stained with anti-Flag mAb and their sensitivity to Fas triggering was tested against agonistic antibody and sFasL.The percentage of apoptotic inhibition was calculated as follows: (cell death for empty vector transfectant À cell death for Fas L199X vector transfectant) / percentage of cells expressing Fas L199X.Columns and points, mean of at least three independent experiments; bars, SD. (Fig. 4B).In contrast, the caspase-8 cleavage was completely abrogated in the Jurkat Fas Q257K incubated with the agonistic antibody 7C11 (Fig. 4B).
We showed previously that in heterozygous Lpr or Lpr cg mice, as well as in ALPS patients, dominant interference by mutant receptors blocked apoptotic but not nonapoptotic pathways, such as NF-nB, ERK, and c-Jun NH 2 -terminal kinase (JNK; ref. 29).We therefore examined these pathways in our mutant cell line and we confirmed these results because FasL triggered comparable level of ERK phosphorylation (see Fig. 4C) and JNK (data not shown) in both the parent and the Fas-mutated cell line.However, 7C11 was unable to induce any of these pathways (Fig. 4C for ERK; data not shown for JNK).Therefore, apoptotic resistance to anti-Fas agonistic mAb could not be attributed to the exacerbation of antiapoptotic pathways, as on the contrary, the whole Fas signal transduction machinery was completely blocked in response to agonistic mAb.
The effect of Fas death domain mutation is reversed by knocking down c-FLIP.From the previous data, we inferred that the step where the blockade of the Fas pathway occurred was located downstream the formation of Fas microaggregates and upstream the caspase-8 activation designating the DISC as the molecular target for this inhibition.A good candidate acting as a checkpoint molecule at the DISC level and responsible of the cell death signal inhibition is c-FLIP.To validate this hypothesis, we knocked down c-FLIP in the Fas Q257K-expressing cell line, using a small hairpin RNA (shRNA) with a sequence designed to target both long and short forms of c-FLIP (c-FLIP L and c-FLIP S ; ref. 30) and we analyzed whether the Fas-mediated signaling pathway was restored.
7C11-resistant Jurkat cells were transiently (Fig. 5A and B) or stably (Fig. 5C) transfected with empty or c-FLIP shRNA-expressing vectors and cells were tested for c-FLIP expression and sensitivity to sFasL and 7C11-induced cell death.The shRNA sequence used to target c-FLIP S/L is located into the coding sequence of the mRNA, which is shared by the long and short c-FLIP variants (30), meaning that this shRNA should down-regulate as efficiently the long c-FLIP than the short c-FLIP.However, we can notice that shRNA knocked down more efficiently the long form than the short form of FLIP in the transiently transfected cells.In contrast, c-FLIP S was efficiently down-regulated in long-term cultured stable clones (Fig. 5C, inset).The more pronounced effect on the long splice variant than the short splice variant at short-term is surprising because it has been reported that the c-FLIP S displays a shorter half-life than c-FLIP L (31).Although we do not have an indisputable answer to explain this phenomenon, it is possible that c-FLIP S mRNA is less accessible to shFLIP RNA than the longer form for reasons, such as the presence of more or different translational or splicing factors on the c-FLIP S than on the c-FLIP L mRNA.Nonetheless, this inhibition of c-FLIP expression was specific, as the expression of caspase-8, which shares sequence similarity with c-FLIP L , remained unchanged in the shRNA-transfected cells (Fig. 5A).The sensitivity of the Fas Q257K Jurkat cell line to sFasL was fully restored in c-FLIP knocked down cells because cell death level returned to the level of the control Jurkat cell line (Fig. 5A).More strikingly, these cells were strongly sensitized to 7C11-triggered death because up to 75% of the cells died (Fig. 5B).We isolated clones from this shFLIP-transfected bulk and we quantified the 7C11-triggered DNA fragmentation in these cells.Agonistic antibody treatment efficiently induced DNA fragmentation in all of these Fas-mutated cells (Fig. 5C).c-FLIP inhibits the DISC formation and the classic caspasedependent apoptotic pathway in Fas mutant-expressing cells.Next, we investigated whether c-FLIP down-modulation restored the formation of the DISC and was followed by the induction of a classic Fas apoptotic signal (caspase dependent) in the Fas Q257K Jurkat cell line.To define the molecular step altered by the c-FLIP Figure 4.Both apoptotic and nonapoptotic signaling pathways are inhibited on 7C11 binding.A, the Fas Q257K mutation did not impair the formation of the Fas microaggregates.Indicated cells were treated with the agonist mAb 7C11 and both monomeric Fas (white arrowheads ) and Fas-containing microaggregates (black arrowheads ) were quantified by an anti-Fas immunoblot.B, cells were incubated with 7C11 mAb (top ) or sFasL (bottom ) for the indicated times and then lysed.Ten micrograms of proteins were loaded in each lane, separated by SDS-PAGE, and a caspase-8 immunoblot was done.C, cells were treated with 1, 0.1, or 0.01 Ag/mL sFasL or mAb 7C11 for 15 min and then lysed.Fifty micrograms of protein were separated by SDS-PAGE, and phosphorylated ERK (P-ERK ) and total ERK (control for loading) were detected by immunoblot.Phorbol 12-myristate 13-acetate (10 ng/mL) and ionomycin (1 Ag/mL; P/I) treatment was used as positive control for ERK1/2 activation in T lymphocytes.
knockdown, we immunoprecipitated the DISC in the Jurkat cell line.Fas was immunoprecipitated to comparable amounts from the parent Jurkat as well as the Q257K cell line and the two clones expressing c-FLIP shRNA (Fig. 6A).FADD was efficiently coimmunoprecipitated from the parent Jurkat cell line but was barely detectable in the DISC of the 7C11-stimulated Jurkat-R and Jurkat-R shFLIP cells (Fig. 6A, arrowheads).Scan analysis of the blot revealed that 10 times less FADD was coimmunoprecipitated in the Jurkat-R DISC than in the parent Jurkat DISC.Similarly to FADD, c-FLIP S and c-FLIP L were less efficiently recruited to the Jurkat-R DISC than to the parent cell line DISC (Fig. 6A).As expected, c-FLIP down-expression led to a dramatic decrease of the quantity of FLIP present in the DISC of the ShRNA c-FLIP Jurkat-R clones compared with of the Jurkat-R.It is noteworthy that the decreased quantity of c-FLIP present in the DISC of the ShRNA c-FLIP Jurkat cells was correlated to a dramatic increase of caspase-8 recruitment, which became detectable (Fig. 6A).However, the cleavage of the small subunit of the procaspase-8 and the appearance of the fragment p43/41 were not detected in the DISC of shFLIP clones after 15 min and could explain why caspase-8 cleavage or DNA fragmentation (see Supplementary Fig. S2) is delayed in the c-FLIP knocked down clones (Fig. 6B compared with Fig. 4B).These results show that (a) FADD binding is not modulated by the c-FLIP expression and (b) the available amount of FADD is not a limiting factor in the DISC formation because recruitment of faint amount of FADD still allows binding and activation of a large quantity of caspase-8 (Fig. 6A).Therefore, in the context of a Fas heterozygous mutation, we show that c-FLIP down-regulation can restore cell death and allow the elimination of the mutated cell.
Next, we investigated whether the restoration of the DISC formation when c-FLIP expression was down-regulated led to the activation of caspases.Although the kinetics of caspase-8 activation was slightly delayed (by 45 min) when compared with the parent Jurkat cell line (compare Fig. 6B with Fig. 4B), the cleavage of caspase-8, which did not occur in the Fas Q257K cell line, was restored in the stable c-FLIP knocked down clones tested (Fig. 6B).The activation of the effector caspase-3 was also restored in c-FLIP knocked down cells when compared with the Fas Q257K cell line (Fig. 6B, bottom).Moreover, a pancaspase inhibitor (z-VAD), a caspase-8 inhibitor (IETD), or a caspase-3 inhibitor (LEHD) completely abrogated the 7C11-induced apoptosis (Supplementary Fig. S1).Therefore, restoration of apoptosis in a Fas-mutated cell line through the down-regulation of c-FLIP expression involved a classic caspase-mediated apoptotic signal.

Discussion
Using a Fas mutant-expressing acute leukemia T-cell line, we show that conversely to what could be expected, a Fas mutation by itself did not result in a complete blockade of the apoptotic signal.Indeed, we bring evidence that the permissivity of the apoptotic signal relies on the amount of c-FLIP expression.Therefore, c-FLIP behaves as a very potent checkpoint molecule in the earliest molecular step of the Fas-mediated signal.
Several published works have led to the conclusion that mutations into the Fas death domain can interfere in a dominantnegative manner with Fas signaling because homotrimers of WT Fas molecules are essential for transmitting a signal (see ref. 32 for review).Recently, we questioned this dogma when we found that cells expressing a heterozygous death domain-mutated Fas could still transduce nonapoptotic signals (29).In addition, in patients suffering from ALPS type Ia, heterotrimeric Fas complexes are still capable of recruiting FADD, but the threshold of activation is not reached.Indeed, in these cells, the binding of FADD to Fas was reduced by 75% compared with cells from healthy donors (33).It is noteworthy that in all family studies done thus far, one of the two parents from a diseased child also carries the Fas mutation but does not suffer from ALPS.Furthermore, in mice, the Lpr mutation causes exacerbated lymphoproliferation and autoimmunity in the MRL genetic background but the same mutation has a mild effect in the C57BL/6 mice strain.The data gathered in the present study suggest that the threshold required to trigger cell death on Fas stimulation may be increased in the ALPS type Iasuffering patient and the MRL mice, for a reason not solely linked to the Fas mutation.Therefore, according to our findings, it could be interesting to investigate the expression of c-FLIP in these pathologic conditions.
We propose a ''multistep threshold'' model, in which various factors control the Fas activation threshold.In this model, expression of a ''dominant-negative'' Fas mutant does not completely abrogate the apoptotic signal because a potent multimeric sFasL or the down-expression of c-FLIP was able both A, the parent and Jurkat-R cells were transiently transfected with control shRNA and with the c-FLIP shRNA-containing vector for the Jurkat-R.The indicated cells were incubated with sFasL for 18 h, and cell death was quantified using MTT assay.Inset, c-FLIP and caspase-8 immunoblots.B, the Jurkat-R cells were transfected with control or c-FLIP shRNA-containing vector, together with the pEGFP plasmid at a 3:1 molecular ratio.The cells were incubated with the indicated inducer, and death of the green cells was determined by measuring the sub-G 1 population using a propidium iodide DNA staining.C, Jurkat-R cells were transfected with empty or c-FLIP shRNA-containing vector, and two independent stable shRNAexpressing clones were isolated.These stable clones and the control cell line were lysed.Inset, immunoblot of anti-FLIP and anti-caspase-8.
to overcome the increased apoptotic threshold and to trigger cell death.Then, the inhibition of the apoptotic signal due to the expression of a Fas mutant can be reversed in acting upstream or downstream the Fas receptor.Recently, a checkpoint molecule has been characterized in the signaling pathway of another member of the TNFR family (34).Indeed, these authors described that a competition between the NF-nB-regulated molecule c-FLIP and caspase-8 favors nonapoptotic or apoptotic signaling, respectively.The triggering of the NF-nB signal leads to c-FLIP overexpression and to cell survival, whereas, on the other hand, the inhibition of NF-nB decreases c-FLIP expression and promotes cell death.Similarly to TNFR1, Fas engagement also triggers nonapoptotic pathways.However, we could not observe a competition between these pathways because numerous cells, among which is the Jurkat cell line, which trigger a robust NF-nB, still undergo apoptosis on Fas engagement (30).Therefore, for Fas signaling, c-FLIP still exerts a checkpoint ancestral function as in the TNFR pathway, but this role has evolved toward the decision to transmit or to block the whole signal.
In type II cells, the DISC is not efficiently formed, but nevertheless, the apoptotic signal triggered on Fas engagement by FasL is transmitted (35) and these cells are resistant to cell death mediated by anti-Fas divalent agonistic antibodies (36).Recently, it has been reported that Fas is concentrated into membrane subdomains called lipid rafts or microdomains in type I cells and excluded from them in type II cells, such as the Jurkat cell line (36), and we reported that the forced relocalization of Fas inside these lipid rafts in type II cells improved the cell death signaling (37).It is conceivable that in type I cells or in type II cells where Fas has been retargeted into the lipid rafts, these proteolipidic structures compete with c-FLIP for the access to the DISC by a still unknown mechanism and thereby allow the transduction of an efficient Fas-mediated cell death signal.
To our knowledge, we provide here the first demonstration that inhibition of Fas-mediated cell death due to the expression of a nonfunctional intracellular mutant can be overcome by acting on the expression of a downstream protagonist of the Fas apoptotic signal.In general, the dysregulation of the Fas apoptotic pathway has been proposed as a mechanism of oncogenesis and has also been implicated in the relapse after chemotherapeutic treatment (16).Moreover, point mutations of the Fas receptor have been found in 11% of non-Hodgkin's lymphoma, in ALPS, and in solid tumors (11).Here, we showed that the blockade of Fas apoptosis resulting from such mutations can be reversed by down-expressing c-FLIP; therefore, we suggest that developing tools to knock down c-FLIP in human could restore the Fas sensitivity and favors the elimination of various Fas-resistant transformed cells.A, for each condition, 10 8 of parent Jurkat, Jurkat-R, and stable c-FLIP shRNA-expressing Jurkat-R cells were incubated for 0 or 15 min with 7C11 (1 Ag/mL) and lysed.Then, Fas was immunoprecipitated as described in Materials and Methods; the immunoprecipitated complex and total cell extracts were separated by 10% SDS-PAGE and immunoblots were carried out with the indicated antibodies.Asterisks, nonspecific staining (APO-1-3 IgG3).For the anti-FADD immunoblot, 2 Â 10 8 cells were lysed and the DISC was immunoprecipitated.The ECL film was overexposed to detect the faint amount of FADD associated to Fas in the 7C11-resistant Jurkat cells.B, cells (10 6 ) were incubated with 7C11 (1 Ag/mL) for the indicated times and then lysed.One hundred micrograms of proteins were separated by SDS-PAGE.Caspase-8 and caspase-3 activation was analyzed by immunoblot method.

Figure 1 .
Figure 1.Selection of cells resistant to antibody but not FasL-mediated cell death.A to D, cells were incubated for 18 h with the indicated inducers, and quantification of cell death was measured using the MTT assay.A, Jurkat cells were incubated with agonistic anti-Fas antibody 7C11 or sFasL.Dashed line, maximum of cell death reached on stimulation with the agonistic antibody.B, top, quantification of Fas expression was done by flow cytometry analysis; bottom, immunoblot analysis of the indicated components of the Fas-mediated cell death pathway.C, the selected Jurkat resistant (Jurkat-R ) and parental cell lines were incubated with 7C11 or soluble TRAIL (sTRAIL ).D, comparison of the nucleotide and peptide sequences of WT Fas and Fas isolated from the Jurkat resistant cell line.Data are representative of at least three independent experiments.

Figure 3 .
Figure 3. Fas Q257K expression in a WT Jurkat dominantly interferes with agonistic mAb but not sFasL-induced cell death.Cells were incubated for 18 h with the indicated inducers, and cell death was measured using the MTT assay.A, flow cytometric expression of the WT and mutated Fas in the Fas-deficient murine T-cell line WR19L, on stable transfection (top ).Effect of the indicated apoptotic inducers on the parent and stably transfected WR19L cells.B, the parent Jurkat cell line was transfected with Fas Q257K cDNA and various stable clones were selected, characterized for Fas expression and for sensitivity toward FasL and 7C11.Points, mean of at least three independent experiments; bars, SD.

Figure 5 .
Figure 5. c-FLIP knockdown restores cell death in Fas mutant-expressing cell line.A, the parent and Jurkat-R cells were transiently transfected with control shRNA and with the c-FLIP shRNA-containing vector for the Jurkat-R.The indicated cells were incubated with sFasL for 18 h, and cell death was quantified using MTT assay.Inset, c-FLIP and caspase-8 immunoblots.B, the Jurkat-R cells were transfected with control or c-FLIP shRNA-containing vector, together with the pEGFP plasmid at a 3:1 molecular ratio.The cells were incubated with the indicated inducer, and death of the green cells was determined by measuring the sub-G 1 population using a propidium iodide DNA staining.C, Jurkat-R cells were transfected with empty or c-FLIP shRNA-containing vector, and two independent stable shRNAexpressing clones were isolated.These stable clones and the control cell line were lysed.Inset, immunoblot of anti-FLIP and anti-caspase-8.

Figure 6 .
Figure6.The knocking down of c-FLIP restores a classic Fas-mediated apoptotic signal in cells harboring a mutated Fas.A, for each condition, 10 8 of parent Jurkat, Jurkat-R, and stable c-FLIP shRNA-expressing Jurkat-R cells were incubated for 0 or 15 min with 7C11 (1 Ag/mL) and lysed.Then, Fas was immunoprecipitated as described in Materials and Methods; the immunoprecipitated complex and total cell extracts were separated by 10% SDS-PAGE and immunoblots were carried out with the indicated antibodies.Asterisks, nonspecific staining (APO-1-3 IgG3).For the anti-FADD immunoblot, 2 Â 10 8 cells were lysed and the DISC was immunoprecipitated.The ECL film was overexposed to detect the faint amount of FADD associated to Fas in the 7C11-resistant Jurkat cells.B, cells (10 6 ) were incubated with 7C11 (1 Ag/mL) for the indicated times and then lysed.One hundred micrograms of proteins were separated by SDS-PAGE.Caspase-8 and caspase-3 activation was analyzed by immunoblot method.