Although c-FLIP has been identified as an important player in the extrinsic (death receptor-induced) apoptosis pathway, its endogenous function in mature T lymphocytes remains undefined. c-FLIP may inhibit or promote T cell death as previous data demonstrate that the c-FLIPL isoform can promote or inhibit caspase 8 activation while the c-FLIPS isoform promotes or inhibits T cell death when overexpressed. Although the c-FLIPR isoform inhibits cell death in cell lines, its function in T cells remains unknown. To investigate the function of c-FLIP in mature T cells, we have generated several genetic mouse models with c-FLIP or its individual isoforms deleted in mature T cells. Surprisingly, we found that c-FLIP protects mature T cells not only from apoptosis induced by the death receptors Fas and TNFR but also from TCR-mediated and spontaneous apoptosis. Thus, c-FLIP plays an essential role in protecting mature T cells from a death signal induced through the TCR itself and is required for naive T cell survival. Our results demonstrate that c-FLIP functions beyond the extrinsic death pathway.
The c-FLIP molecule (also named CLARP, FLAME, I-FLICE, MRIT, usurpin, Casper, or CASH) plays important roles in death receptor-mediated signaling (1, 2, 3, 4, 5, 6, 7, 8). Three isoforms of c-FLIP derived from mRNA alternative splicing have been identified in human cells: 55kD c-FLIPL, 24kD c-FLIPR, and 26kD c-FLIPS (9, 10), whereas only c-FLIPL and c-FLIPR but not c-FLIPS are expressed in mouse cells (11). c-FLIPL contains two death effector domains at the N terminus and an inactive caspase-like domain at the C terminus. c-FLIPS and c-FLIPR only contain two death effector domains (9, 10, 11). All three isoforms inhibit death receptor-induced apoptosis, or the extrinsic death pathway, by interfering with pro-caspase 8 recruitment to the adaptor protein Fas-associated death domain when overexpressed (9, 10, 11, 12).
The roles of individual c-FLIP isoforms in mature T cells remain undefined as conditional deletion of c-FLIP in thymocytes prevents mature T cell development (13). Previous studies suggested that c-FLIPL may play a role in Th cell differentiation in transgenic (Tg)3 overexpression mouse models (14, 15, 16). c-FLIPL may promote or inhibit mature T cell death given that c-FLIPL can also promote pro-caspase 8 activation at low levels of expression (17, 18, 19, 20, 21). The roles of c-FLIPs (c-FLIPR in mouse) in mature T cells are controversial. For example, one report showed that overexpression of c-FLIPS/R decreases mature T cell survival in mouse (22), whereas another report using a similar system did not observe decreased T cell survival (23). Given these conflicting results, it is essential to assess the roles of endogenous c-FLIP in mature T cells using loss-of-function models. In this report, we have generated mouse models lacking c-FLIP or its individual isoforms in mature T cells. We show that the in vivo function of c-FLIP (c-FLIPL and c-FLIPR) is to protect mature T cells from death receptor-induced apoptosis. Importantly, we found that c-FLIP protects mature T cells from TCR-induced and spontaneous apoptosis. Thus, these results indicate that c-FLIP functions beyond the extrinsic death pathway.
Materials and Methods
c-FLIPf/f mice and c-FLIPR (previously termed c-FLIPS) bacterial artificial chromosome (BAC) Tg mice were generated in our laboratory as described (13, 24). To generate c-FLIPL BAC Tg mice, a mouse c-FLIP wild-type BAC DNA clone was modified using a system as described (25). In brief, mouse c-FLIPL cDNA, including its own stop codon and poly(A) signal, was fused to the start codon in exon 1 of the c-FLIP gene. A 60kb BAC fragment that is 15kb downstream of the nearest 5′ gene (Als2cr12) and 10kb upstream of the nearest 3′ gene (caspase 8) was used for injection. Three founders were obtained for each line and crossed to c-FLIPf/fER-Cre mice, and all mice displayed a similar phenotype. Animal usage was conducted according to protocols approved by the Duke University Institutional Animal Care and Use Committee.
PCR to assess the presence of the c-FLIPR or c-FLIPL BAC transgene was performed with 35 cycles of 94°C for 30 s, 52°C for 30 s, and 72°C for 120 s. The primers are: forward, 5′-GAGGTTGAGGGACTTGGCATG-3′; reverse, 5′-TCAGCAGGACCCTATAATCAG-3′.
In vitro deletion of c-FLIP
Pooled single-cell suspensions from spleen and lymph nodes (LNs)were lysed of RBCs by ACK buffer. Cells were cultured at 4 × 106/ml in complete RPMI 1640 medium in the presence of 0.2 μM 4OH-tamoxifen (Sigma-Aldrich) for 3 days. Live cells were purified by Lympholyte centrifugation (Cedarlane) according to the manufactory instruction. Live cells were recultured at 2 × 106/ml in the presence of different stimuli or subjected to Western and RT-PCR analysis.
Detection of active caspase 8
One day or 3 days after in vitro deletion of c-FLIP, live cells were purified and recultured in the presence or absence of 10 μg/ml anti-CD3 for 5 h. Fluorescence-labeled cell permeable caspase 8 inhibitor (FAM-LETD-FMK from Immunochemistry Technologies) was added to the culture during the last hour of incubation according to the manufactory instruction. Cells were surface stained and analyzed by flow cytometry.
Flow cytometric analysis
Cells were incubated with an FcR blocker (2.4G2; eBioscience), stained on ice for 30 min with FITC-, PE/Cy5-, or allophycocyanin-labeled mAb, and washed with PBS containing 2% FCS. A total of 1 to 5 × 104 events were collected on a FACScan flow cytometer (BD Biosciences) and analyzed using CellQuest (BD Biosciences) or FlowJo software. All fluorescence-labeled Abs, including anti-CD4 and -CD8 were obtained from eBioscience, Biolegend, or BD Biosciences. Apoptotic cells were defined by Annexin V and 7-AAD staining using an Annexin V-PE kit (BD Biosciences).
Abs used for Western blots were anti-c-FLIP (clone Dave-2; Alexis Biochemical) and anti-γ-tubulin (Santa Cruz Biotechnology).
Semiquantitative RT-PCR analysis
After in vitro deletion of c-FLIP, total RNA was extracted from live cells and subjected to RT-PCR analysis. The primers are: forward primer for both c-FLIPL and c-FLIPR: 5′GCTGCTGTGGTTCTGAACATG 3′; reverse primer for c-FLIPL: 5′CTTTGACTGTCACGGTATTCCAC 3′ and reverse primer for c-FLIPR: 5′GGTACTCCATACACTGGCTCC 3′. To detect c-FLIPL, the PCR was performed with 35 cycles of 94°C for 30 s, 64°C for 30 s, and 72°C for 60 s. To detect c-FLIPR, the reaction was performed with 35 cycles of 94°C for 30 s, 68°C for 30 s, and 72°C for 60 s.
Impaired survival of c-FLIP-deficient mature T lymphocytes after death receptor engagement
Our previous data have demonstrated that conditional deletion of c-FLIP (both c-FLIPL and c-FLIPR) in thymocytes results in severe impairment in the development of mature T lymphocytes in the periphery, excluding further analysis of c-FLIP function in these cells (13). To directly assess c-FLIP function in mature T cells, we crossed c-FLIPf/f mice with ER-Cre mice (26) to generate c-FLIPf/fER-Cre mice for the inducible deletion of c-FLIP. c-FLIPf/fER-Cre mice develop normally and have lymphoid compartments comparable to littermate controls (data not shown). As ER-Cre is expressed in all tissues, tamoxifen treatment of c-FLIPf/fER-Cre mice in vivo leads to lethality within 3 days, while all c-FLIPf/+ER-Cre littermates remain healthy (n = 5), suggesting that c-FLIP plays an essential role in host survival in other tissues. We used in vitro deletion of c-FLIP to determine the role of c-FLIP in mature T cell survival. Total spleen and LN cells were pooled and cultured in the presence of 0.2 μM 4OH-tamoxifen for 3 days. Live cells were purified and recultured overnight with different stimuli. Under this culture condition, 4OH-tamoxifen exhibited no obvious toxicity to control T cells (data not shown) and induced efficient c-FLIP deletion in c-FLIPf/fER-Cre T cells (Fig. 1,A). Consistent with previous results that c-FLIP functions as inhibitor of death receptor-induced apoptosis (10, 27), c-FLIP−/− T cells exhibit markedly increased apoptosis upon TNF-α and anti-Fas treatment compared with control T cells (Fig. 1 B and 3A), indicating that c-FLIP protects T cells from death receptor-induced death. After 3 days in vitro culture, 30–40% naive T lymphocytes remained alive. To exclude the possibility that an abnormal T cell population was enriched after in vitro culture, we also performed the in vitro culture in the presence of IL-7. As expected, minimal T cell death was observed in the presence of IL-7. c-FLIP-deficient T cells from culture with or without IL-7 exhibited similar survival defect (data not shown).
Impaired survival of c-FLIP-deficient T cells after TCR engagement
As TCR signaling induces caspase 8 activation (28, 29) and c-FLIP can either promote or inhibit caspase 8 activation (10, 17, 18, 19), and c-FLIP-deficient T cells have been found to have increased apoptosis following TCR stimulation in a Rag chimeric mutant model (30), we tested what roles c-FLIP might play after TCR stimulation of mature T cells in an ER-cre inducible system. We stimulated mature T cells that were deleted of c-FLIP in vitro with anti-CD3. Remarkably, TCR stimulation induced the death of 60–90% of CD4+ and CD8+ c-FLIP-deficient T cells (Fig. 1, B and C). Since it is possible that naive T cells could be more sensitive to tamoxifen-induced toxicity than activated cells, resulting in a cell population enriched for activated cells that would undergo activation-induced cell death upon TCR stimulation, we examined wild-type and c-FLIPf/fER-cre T cells cultured for 3 days with and without tamoxifen. We saw no differences in the activation profile (CD62L and CD44 expression) in cells treated with or without tamoxifen, indicating that the results we see are due to a defect in T cell survival following initial TCR stimulation in c-FLIP-deficient cells (data not shown). Interestingly, c-FLIP−/− T cells also exhibit a slightly but significantly elevated apoptotic rate without any stimuli (Fig. 1, B and C). It should be noted that in the absence of anti-CD3 stimulation, T cells were still in the resting stage after in vitro culture as determined by CD25 and CD69 staining (data not shown). These results suggest that c-FLIP regulates the survival of both resting and activated T cells. We then tested whether caspase 8 activity is enhanced in c-FLIP−/− T cells. One day after 4OH-tamoxifen induced deletion, c-FLIPf/fER-Cre T cells exhibited comparable caspase 8 activity as wild-type T cells due to the incomplete deletion of c-FLIP at this time point (Fig. 1, A and D). However, when c-FLIP deletion was completed at day 3, c-FLIP−/− T cells exhibited a 33% increase (17.3 vs 23.1%) without TCR stimulation and a 78% increase (20 vs 35.6%) with TCR stimulation in the percentage of cells containing active caspase 8 during a 5-h culture (Fig. 1,D). Furthermore, addition of benzyloxycarbonyl-Val-Ala-Asp (zVAD) restored the survival defects in c-FLIP−/− T cells with or without TCR stimulation (Fig. 2, A and B), suggesting that these cells die by apoptosis. These results suggest that c-FLIP plays a critical role in the survival of resting and activated T cells likely through the inhibition of caspase 8 activation.
Both c-FLIPL and c-FLIPR isoforms protect T cells from death
Although two isoforms of c-FLIP, c-FLIPL and c-FLIPR, have been identified in mouse T cells, their function remains undetermined. It was shown that c-FLIPL inhibits caspase 8 activation at high expression levels, while it promotes caspase 8 activation at low expression levels (17, 18, 19). An enhanced caspase 8 activity was observed in c-FLIPL Tg T lymphocytes (20, 21). Although c-FLIPR is considered a caspase 8 inhibitor (9, 10, 11), its role has not been examined in mouse T lymphocytes. Furthermore, a recent report showed that overexpression of human c-FLIPS, which is very similar human c-FLIPR at both sequence and function level, promotes T cell death (22) and this result is contradictory to a previous publication (23). Thus, it is essential to examine the role of endogenous c-FLIPL and c-FLIPR in mature T cell survival. To achieve this, we have generated conditional knockout mice lacking either isoform by crossing c-FLIPf/fER-cre to BAC Tg mice expressing either mouse c-FLIPL or mouse c-FLIPR cDNA. As these cDNAs were inserted into the c-FLIP locus in BAC DNA, the expression of c-FLIPL or c-FLIPS in these mice is controlled by the endogenous c-FLIP promoter and regulatory elements. T lymphocytes from c-FLIPf/fER-cre.c-FLIPL BAC Tg mice that are treated with 4OH-tamoxifen express c-FLIPL but not c-FLIPR and therefore referred to as c-FLIPR−/− T cells (Fig. 2,C). Similarly, T lymphocytes from c-FLIPf/fER-cre.c-FLIPR BAC Tg mice that are treated with 4OH-tamoxifen express c-FLIPR but not c-FLIPL and therefore referred to as c-FLIPL−/− T cells (Fig. 2,C). Expression of either c-FLIPR or c-FLIPL in T lymphocytes was sufficient to rescue the survival defects in c-FLIP−/− T cells in the presence or absence of TCR stimulation (Fig. 2, A and B). These data suggest that both c-FLIPR and c-FLIPL isoforms are anti-apoptotic caspase 8 inhibitors in mature T cells at endogenous expression levels.
The role of death receptors in TCR-induced apoptosis of c-FLIP−/− T cells
Previous data suggests that TCR-mediated apoptosis is mediated through Fas/FasL or TNF-α/TNFRs in CD4+ and CD8+ T cells (31, 32, 33, 34, 35, 36, 37, 38, 39). The massive death of c-FLIP-deficient T cells upon TCR stimulation may be due to Fas- and/or TNFR-induced death signal. To determine the contribution of Fas and TNF-α during the apoptosis of c-FLIP−/− T cells upon TCR stimulation, c-FLIPf/fER-Cre mice were crossed onto lpr (Fas) mutant, TNF-α-deficient, and lpr/lpr/TNF-α double-deficient backgrounds. Fas mutation partially rescued the TCR-induced apoptosis while it completely rescued anti-Fas-induced apoptosis in c-FLIP−/− T cells (Fig. 3, A and C). In contrast, deletion of TNF-α had no obvious effect on the high apoptosis of c-FLIP−/− T cells (Fig. 3, B and C). Furthermore, c-FLIP-deficient T cells from an lpr/lpr/TNF-α double-deficient background had apoptotic rates similar to those from an lpr single mutation background (Fig. 3, B and C). Thus, these results suggest that Fas-mediated death signaling only contribute partially to TCR-induced apoptosis in c-FLIP-deficient T cells, while TNF-mediated death signaling has no obvious role in this process.
c-FLIP is critical for maintaining T cell homeostasis in vivo
To directly test whether c-FLIP is required for mature T cell survival in vivo, T cells from c-FLIPf/fER-Cre and c-FLIPf/+ER-cre mice were labeled with high or low amounts of CFSE, mixed at a 1:1 ratio, and transferred into unmanipulated wild-type hosts. Similar results were seen when host LN and spleen were analyzed, and the panel shown is from host LN. Five days after in vivo tamoxifen administration, the number of c-FLIP−/− T cells was decreased to 10–40% of that of transferred control T cells (Fig. 4). Thus, c-FLIP is required for T cell homeostasis in vivo.
Although c-FLIP has been identified as a critical player in death receptor-induced signaling for more than a decade, the function of endogenous c-FLIP in mature T cells remains unclear. Conflicting results suggest that c-FLIPL might inhibit or promote mature T cell death (17, 18, 19, 22, 23). We have generated several genetic models to investigate the function of c-FLIP in mature T cells. As expected, we found that c-FLIP protects T cells from death receptor-induced apoptosis via Fas and TNFR. However, we also found that c-FLIP protects mature T cells from TCR-induced as well as spontaneous cell death. Furthermore, both c-FLIPL and c-FLIPR isoforms function as anti-apoptotic proteins in mature T cells.
The likely target for c-FLIP in the survival of resting and TCR-stimulated mature T cells is caspase 8. Caspase 8 is actively involved in T cell homeostasis. TCR signaling rapidly up-regulates caspase 8 activity (28, 29). Considering the fact that weak TCR/MHC interaction is required for naive T cell survival, basal caspase 8 activity may also be present in resting T lymphocytes. Indeed, c-FLIP−/− T cells exhibit a marked increase of caspase 8 activity with or without TCR stimulation. Furthermore, because c-FLIP−/− T cells exhibit enhanced spontaneous and TCR-induced apoptosis, and zVAD completely rescues both defects, c-FLIP is the essential caspase 8 inhibitor in T lymphocytes. In the absence of c-FLIP, caspase 8 activity may elevate to a harmful level even without TCR stimulation. Thus, the major function of c-FLIP in resting naive T cells is likely to prevent the constitutively low levels of active caspase 8 from killing the cells.
Our data demonstrate that c-FLIPL and c-FLIPR, when expressed at endogenous levels, inhibit apoptosis. The role of c-FLIPL in regulating caspase 8 activity and cell death has been controversial. c-FLIPL was shown to promote caspase 8 activation at low expression levels and inhibit caspase 8 activation at high expression levels (17, 18, 19, 20, 21). Although c-FLIPR is generally considered a caspase 8 inhibitor (9, 10, 11), recent reports showed that overexpression of c-FLIPS, which is very similar to c-FLIPR, in T cells generated conflicting results regarding T cell death (22, 23). These controversial results are likely derived from the various overexpression systems used. By expressing the c-FLIPL or c-FLIPR isoform under its own regulatory elements, our data clearly show that both c-FLIPL and c-FLIPR are anti-apoptotic in resting and TCR-stimulated mature T lymphocytes.
Our data demonstrate that the Fas/FasL interaction, but not TNF-α, contributes only partially to the apoptosis of c-FLIP-deficient CD4+ and CD8+ T cells upon TCR engagement. Previous studies have demonstrated that preactivated T cells and T cell hybridoma cell lines undergo rapid apoptosis upon TCR restimulation through a Fas/FasL- or TNF-α/TNFR-dependent mechanism (31, 32, 33, 34, 35, 36, 37, 38, 39). However, our genetic data has clearly established that Fas does not cause all the cell death in TCR-stimulated c-FLIP-deficient T cells. Our results suggest that TCR engagement may activate caspase 8 independent of death receptors and Fas/FasL may function as an amplifier, but not an initiator, of caspase 8 activation in T lymphocytes. Although lpr mutation does not completely eliminate Fas signaling (40), it is unlikely that the mutant Fas still delivers a critical death signal during TCR-induced cell death in our system based on three lines of evidences. First, lpr mutation completely restores the survival defects upon the anti-Fas stimulation in c-FLIP−/− T cells (Fig. 3 A). Second, extensive previous investigations have shown that lpr mutation rescues the preactivated T cells from TCR-induced apoptosis (33, 35, 36, 38, 39). Third, in the report showing that lpr mutant is partially functional, the authors clearly demonstrated that lpr mutant completely loses the capacity to induce cell death while it can activate NF-κB signaling only in heterozygous (lpr/+) mice (40). In our experiments, only lpr/lpr mice were used. Therefore, lpr mutant can be considered as a complete loss-of-function of the death inducing capacity of Fas. Since death receptor mutations partially restore and zVAD completely restores the survival defects in c-FLIP−/− T cells upon TCR stimulation, this suggests that TCR signaling induces caspase activation through a Fas- (and TNFR-) independent mechanism. Thus, c-FLIP is critically needed to protect T cells from caspase activity during resting as well as activation phases.
We thank T. Matt Holl and Cheryl Bock for help in generating the c-FLIPL BAC Tg mice.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grants CA92123 and AI54683.
Abbreviations used in this paper: Tg, transgenic; BAC, bacterial artificial chromosome; LN, lymph node; zVAD, benzyloxycarbonyl-Val-Ala-Asp.