Fas-associated death domain (FADD) is a negative regulator of T-cell receptor–mediated necroptosis

Abstract

Cell death is an important mechanism to limit uncontrolled T-cell expansion during immune responses. Given the role of death-receptor adapter protein Fas-associated death domain (FADD) in apoptosis, it is intriguing that T-cell receptor (TCR)–induced proliferation is blocked in FADD-defective T cells. Necroptosis is an alternate form of death that can be induced by death receptors and is linked to autophagy. It requires the death domain-containing kinase RIP1 and, in certain instances, RIP3. FADD and its apoptotic partner, Caspase-8, have also been implicated in necroptosis. To accurately assess the role of FADD in mature T-cell proliferation and death, we generated a conditional T-cell–specific FADD knockout mouse strain. The T cells of these mice develop normally, but lack FADD at the mature stage. FADD-deficient T cells respond poorly to TCR triggering, exhibit slow cell cycle entry, and fail to expand over time. We find that programmed necrosis occurs during the late stage of normal T-cell proliferation and that this process is greatly amplified in FADD-deficient T cells. Inhibition of necroptosis using an inhibitor of RIP1 kinase activity rescues the FADD knockout proliferative defect. However, TCR-induced necroptosis did not appear to require autophagy or involve RIP3. Consistent with their defective CD8 T-cell response, these mice succumb to Toxoplasma gondii infection more readily than wild-type mice. We conclude that FADD constitutes a mechanism to keep TCR-induced programmed necrotic signaling in check during early phases of T-cell clonal expansion.

The immune response to infection requires a delicate balance between cell expansion and contraction. During the expansion phase, programmed cell death is held in check by various mechanisms to allow lymphocytes to proliferate and to counter infection. These mechanisms include up-regulation of prosurvival factors and antiapoptotic proteins. As the immune response wanes, prodeath signals dominate to cause contraction of the responding lymphocytes (1–3). Although apoptosis has long been considered the premier form of programmed cell death, other forms of programmed death have more recently been identified (4). Despite its reputation as an unregulated, nonspecific event, necrosis can occur in a regulated and programmed fashion under certain circumstances. This process, termed programmed necrosis or necroptosis (5), is independent from caspases and can be triggered by the tumor necrosis factor (TNF) death receptors (e.g., Fas, TNF-R, TRAIL-R) when apoptosis is blocked (6). More recently, hyperautophagy has been suggested to cause necroptosis as well (7, 8). Necroptosis has also been suggested to contribute to the danger signals that alert the immune system to certain pathogenic invasions (4).

The death domain-containing kinase RIP1 is a multifunctional serine/threonine kinase known for its functions downstream of the death receptors to induce prosurvival signals such as NF-κB and MAPKs (9). RIP1 is also critical for necroptosis as Necrostatin-1, an inhibitor of RIP1 kinase activity, blocks this process (9). More recently, another member of the RIP1 family, RIP3, has also been shown to be crucial for TNF-induced programmed necrosis (10–12). In this context, RIP3 is found in a complex with RIP1, Caspase-8, and Fas-associated death domain protein (FADD) in cells stimulated with TNF in the presence of caspase inhibitors. In addition, RIP3^−/− mice suffer from impaired necrosis and reduced innate immunity when infected with vaccinia virus and die prematurely (12).

FADD is a 28-kDa adaptor protein that is a critical component of the death receptors’ apoptotic signaling pathway (13, 14). It initiates the formation of a death-inducing signaling complex and serves as a docking site for Caspase-8 (15). FADD-deficient cells exhibit impaired death receptor-mediated apoptosis. In addition, FADD-, Caspase-8-, or c-FLIP–deficient T cells fail to proliferate when stimulated through the T-cell receptor (TCR) complex (13, 16–18). Whereas cell cycle-specific phosphorylation of FADD is partially responsible for proper T-cell proliferation (19–21), the proliferative defect of Caspase-8–deficient T cells or T cells expressing a truncated FADD transgene (FADDdd) is associated with excess necroptosis that can be rescued with Necrostatin-1 (8, 22). Necroptosis in these cells, however, is due to hyperautophagy. As FADDdd transgenic mice suffer from defective T-cell development (23), mature T cells in these mice are likely to be abnormal. Abnormalities may also exist for FADD-deficient mature T cells from FADD^−/−→RAG-1^−/− chimeric mice (24), where T-cell development is severely impacted. To overcome this issue, we have generated a conditional FADD knockout mouse where T-cell development is normal. Herein we show that FADD does have a role in early cell cycle progression, but the absence of FADD reveals its dominant role to be as a negative regulator of necroptosis during T-cell expansion. However, hyperautophagy does not signal the observed necroptosis in these FADD knockout cells, nor does it involve RIP3. We conclude that FADD constitutes a mechanism to keep a presently unidentified pathway of RIP1-dependent necroptosis in check during early phases of the T-cell immune response.

Results

Generation and Phenotypic Analysis of tFADD^−/− Mice.

Conditional T-cell–specific FADD knockout (tFADD^−/−) mice (on the C57BL/6 background) were generated as described in Materials and Methods (Fig. S1). They are born at the expected Mendelian ratios and appear healthy with no gross abnormalities. PCR analysis shows that deletion occurs in the double-negative (DN) (CD4⁻CD8⁻) thymocyte population and is complete in the double-positive (DP) (CD4⁺CD8⁺) compartment with no detectable flox allele (Fig. 1A). Although a residual signal of the flox allele can be seen in CD8⁺ mature T cells, Western blot analysis shows that FADD protein is absent in both CD4 and CD8 T-cell subsets (Fig. 1B). tFADD^−/− thymi are of equal size and cellularity to wild-type littermate controls (Fig. 1C). The CD4/CD8 profiles, as well as analysis of the DN subsets, of FADD-deficient thymi are also comparable to those of littermate controls (Fig. 1D). Flow analysis of various selection markers showed normal positive selection in tFADD^−/− mice. As expected, FADD-deficient thymocytes are resistant to Fas-induced apoptosis (Fig. S2A). Older tFADD^−/− mice do not exhibit any signs of autoimmunity, suggesting that deletion of autoreactive thymocytes during development is functional.

Fig. 1.

tFADD^−/− mice exhibit normal T-cell development but have altered peripheral lymphoid organ compartments. (A) Thymocyte and peripheral T-cell populations from control (ctrl) and tFADD^−/− (ko) littermates were sorted and analyzed by PCR for the presence of the flox allele (flox) and subsequent deletion of the FADD exon 1 (del). (B) CD4 and CD8 T cells were sorted and either not stimulated or stimulated for 24 h with ConA/Phorbol myristate acetate (PMA). Whole-cell lysates were analyzed for expression of FADD; a nonspecific (NS) band demonstrates loading equality. (C) The average thymic cellularity of control and tFADD^−/− mice with ages ranging from 5 to 12 wk. Each dot represents one mouse. (D) CD4/CD8 profiles of control and tFADD^−/− thymi. (E) Lymphocytes from control and tFADD^−/− mice were analyzed for various cell types. (Upper) Lymph nodes were analyzed for B cells (B220) and T cells (TCR). (Lower) Splenocytes were analyzed for B cells (B220) and red blood cells (TER-119).

In contrast to their thymic cellularity, tFADD^−/− mice consistently possess enlarged lymph nodes and spleens (Fig. S2B; P < 0.05). Whereas the percentage of T cells in the peripheral organs of tFADD^−/− mice is reduced, the absolute cell number of CD4⁺ and CD8⁺ T cells is equal to that of littermate controls, suggesting T-cell maturation and migration to the periphery are unaffected (Fig. 1E). This is further supported by a lack of evidence for homeostatically proliferating cells (CD44^hi, CD62L^hi) in tFADD^−/− mice (Fig. S2C). Further analysis showed that tFADD^−/− mice exhibit a drastic increase of B cells in the lymph nodes and spleen (Fig. 1E). Whereas control and tFADD^−/− spleens typically maintain similar B-cell percentage profiles, the absolute number of B cells is higher in tFADD^−/− mice that exhibit splenomegaly. Interestingly, there is also an increase of TER-119⁺ red blood cells in the spleen but not in the bone marrow of tFADD^−/− mice (Fig. 1E). Thus, the enlarged lymph nodes and spleen observed in tFADD^−/−mice are due to increased B cells and red blood cells, respectively.

FADD-Deficient T Cells Have Impaired Early Cell Cycle Transition.

To understand the nonapoptotic function of FADD in mature T cells, we analyzed the ability of FADD-deficient T cells to proliferate in response to TCR ligation. Similar to published work using FADD^−/−→RAG1^−/− chimera mice (13), stimulation with anti-CD3/CD28 antibodies revealed a defect in the ability of FADD-deficient T cells to accumulate in culture over time (Fig. 2A). The effect was most pronounced in the CD8 T-cell subset, as compared with that of the CD4 T-cell subset. Defects were also found when we stained CD8⁺ FADD-deficient T cells with K_i-67, which detects non-G₀ cells, and when we examined cell division using carboxyfluorescein succinimidyl ester (CFSE) (Fig. 2B and Fig. S3A). In contrast to T cells from FADD^−/−→RAG1^−/− chimera mice, however, CD4⁺ FADD-deficient cells from tFADD^−/− mice do proliferate to some extent, showing only a slight delay in K_i-67 expression and cell division. A pulse-chase experiment using concurrent propidium iodide (PI) staining and BrdU labeling also showed cell cycle entries of FADD-deficient cells with only a delay in progression through the cell cycle (Fig. S3B).

Fig. 2.

FADD-deficient T cells do not proliferate well. (A) Purified peripheral T cells were activated by plate-bound anti-CD3 and anti-CD28 antibodies and total cells were counted daily via trypan blue exclusion to obtain the growth curves. Solid lines, T cells from three separate tFADD^−/− mice; dotted lines, wild-type (wt) T cells from littermate controls. Data are representative of multiple independent experiments (>10) with similar results. (B) Purified control (shaded) and FADD-deficient (solid lines) T cells were labeled with CFSE, activated with anti-CD3/CD28 antibodies for the indicated times, and stained for CD4 and CD8. Data are representative of multiple experiments (>3) with similar results.

In T cells from FADD^−/−→RAG1^−/− chimera mice, loss of FADD affects the machinery at all phases of the cell cycle (24). Despite normal development, phosphorylation mimetic mutant FADD(S191D) T cells similarly exhibit early cell cycle defects that specifically affect the cell cycle master regulator FoxM1 and the Rb protein (21). However, T cells from tFADD^−/− mice exhibit normal cell cycle protein profiles. Thus, atypical expression of many cell cycle proteins in T cells from FADD^−/−→RAG1^−/− chimeras is likely a consequence of abnormal T-cell development in the absence of FADD.

FADD-Deficient T Cells Die by Necroptosis During Proliferation.

Although FADD-deficient T cells from tFADD^−/− mice are capable of some cell division, they ultimately do not accumulate in culture. FADD-null T cells do exhibit a slower cell cycle kinetic, but this 2-h lag is unlikely to account for the overt proliferation defect. We thus decided to examine the rate of cell death in these T-cell cultures. The forward/side scatter (FS/SS) profiles of the tFADD^−/− T-cell cultures suggested an increase in necrotic death (necrotic cells are FS^lo/SS^hi; Fig. 3A). Costaining of Annexin V and PI also pointed to necrotic death of FADD-deficient T cells following activation. Specifically, early apoptotic T cells (Annexin V⁺,PI⁻) could not be detected in FADD-deficient cultures, whereas phosphatidyl serine exposure only occurred concurrently with compromise of the plasma membrane (Annexin V⁺,PI⁺) (Fig. S4A). Measurement of the mitochondrial membrane potential further suggests that FADD-deficient T cells are dying by necrosis. Gating on live cells within the culture, we found relatively little loss of 3,3′-dihexyloxacarbocyanine iodide (DiOC6) in either control or FADD-deficient T cells (Fig. S4B). However, DiOC₆ loss is increased in the tFADD^−/− culture when the entire culture is analyzed (Fig. S4B), suggesting the mitochondrial membrane is compromised following breakdown of the plasma membrane, a hallmark of necrosis.

Fig. 3.

TCR-stimulated FADD-deficient T cells die by necrosis and can be rescued by inhibition of RIP1 kinase activity but not by inhibition of hyperautophagy. (A) T cells from control and tFADD^−/− littermates were activated with anti-CD3/CD28 antibodies and analyzed by forward scatter (FS) versus side scatter (SS) for evidence of necrotic death. (B) The accumulation of T cells in culture after activation as in A was determined by daily cell counts using trypan blue exclusion. Cultures were also stained for CD4 and CD8 to determine the absolute cell numbers for each T-cell subset. Necrostatin-1 (Nec) was added at 30 μM where indicated. (C) (Upper) Activated T cells from control and tFADD^−/− mice were either not treated or treated with Necrostatin-1 and subsequently analyzed at the indicated time points for Ki67, CD4, and CD8 expression. (Lower) Same as the top panels except T cells were labeled with 0.5 μM CFSE preactivation. Shown is the plot for CD4 at day 3 postactivation and CD8 at day 4 postactivation. (D) Purified T cells were CFSE-labeled, activated as in A, and either not treated (NT) or treated with 0.5 mM 3-MA, and surface-stained for CD8 postharvest. Data are representative of multiple experiments (>3) with similar results.

Recently, a caspase-independent program of cell death that results in a necrotic phenotype has been identified and termed programmed necrosis (25, 26). In particular, necroptosis is a form of programmed necrosis that is dependent upon RIP1 kinase activity for execution. Necrostatin-1 is a specific inhibitor of RIP1 kinase activity (9) and has been used to block necroptosis in various contexts. Using this compound, we found that inhibition of RIP1 kinase activity by Necrostatin-1 fully rescued the inability of FADD-deficient T cells, both CD4 and CD8, to proliferate, as measured by multiple parameters (Fig. 3 B and C). Interestingly, treatment with Necrostatin-1 also consistently increased the accumulation of control T cells in culture versus untreated control T cells at later time points (Fig. 3B). Thus, a basal level of necroptosis occurs in activated T-cell cultures, suggesting that necroptosis may actually be a physiological event in the context of T-cell proliferation and the immune response. FADD(S191D) mice, which express a mutant form of FADD that mimics constitutive phosphorylation at serine 191, also exhibit proliferative defects but which can be specifically attributed to a blockage at the S-phase transition. Unlike tFADD^−/− T cells, Necrostatin-1 treatment is unable to fully rescue the impaired proliferation of FADD(S191D) T cells (Fig. S5A). This is likely due to the maintained capacity of FADD(S191D) T cells to undergo death by apoptosis in response to impaired S-phase progression. Thus, necroptosis appears to occur uninhibited in the absence of FADD during T-cell expansion and is the main cause of proliferative failure of FADD-deficient T cells.

Necroptosis of Proliferating FADD-Deficient T Cells Is Not Due to Hyperautophagic Signaling, RIP3 Signaling, or Failure of Caspase-8 to Cleave RIP1.

To understand the mechanism of tFADD^−/− T-cell necroptosis, we analyzed various pathways or molecules previously identified to be involved in necroptotic signaling. FADD has previously been linked to ATG5, a component of the autophagy pathway (8, 27, 28), and T cells expressing a truncated form of the FADD protein (FADDdd) exhibit hyperautophagy and subsequent necroptotic death (8). We analyzed the role of autophagy in the necroptotic phenotype of the tFADD^−/− T cells by using 3-methyladenine (3-MA), an inhibitor that blocks autophagosome formation (29). We found that this inhibitor was unable to rescue the death of activated tFADD^−/− T cells (Fig. 3D). In fact, 3-MA was toxic to primary T-cell cultures at the published concentrations (0.5–1 mM), as evidenced by the reduced number of wild-type T cells (Fig. 3D and Fig. S5B). If titrated to less toxic ranges (0.1 mM), 3-MA was still incapable of rescuing FADD-deficient T-cell proliferation (Fig. S5C). Addition of 3-MA post-TCR activation at 48 h also failed to rescue tFADD^−/− T-cell proliferation. In contrast, a similar concentration of 3-MA (0.5 mM) could rescue z-VAD-induced autophagic cell death of L929 cells (30) (Fig. S5D). These data suggest that if hyperautophagy does ensue in the process of FADD-deficient T-cell death, it is likely a consequence of the necroptotic program rather than the cause. Hyperautophagy as a consequence of necroptosis was previously observed by others (9), but is in contrast to a recent publication that suggests 3-MA can rescue the proliferation defect observed in T cells expressing FADDdd (8).

Caspase-8 has been shown to cleave and inactivate RIP1 during apoptosis, but this mechanism is also unlikely to be responsible for keeping RIP1-mediated necroptosis in check in actively proliferating T cells (31). We did not detect an increase in the incidence of cleaved RIP1 in activated control versus tFADD^−/− T-cell cultures at early time points (Fig. 4A). Therefore, a failure of RIP1 cleavage by Caspase-8 cannot explain the necroptotic phenotype of tFADD^−/− T cells.

Fig. 4.

Necroptosis of FADD-deficient T cells does not induce RIP1 cleavage or a RIP1-RIP3 complex. (A) Whole-cell extracts from activated control and tFADD^−/− T cells at various time points were blotted for RIP1 expression. RIP1c denotes cleaved RIP1. (B) Whole-cell extracts were prepared as in A, subjected to RIP3 immunoprecipitations (IP), and subsequently blotted for RIP1 or RIP3. Coimmunoprecipitation of phosphorylated RIP1 (pRIP1) with RIP3 in L929 cells undergoing necroptosis [+ tumor necrosis factor (TNFα) and z-VAD] serves as a positive control. The asterisk denotes a nonspecific band that we were unable to distinguish as either a nonspecific protein or an unmodified RIP1 that was pulled down nonspecifically with control IgG as previously reported by others (44). (Lower) Whole-cell extracts (WCE) were immunoblotted for RIP1 or RIP3 expression. Data are representative of several experiments with similar results.

Recently, RIP3, another RIP kinase family member, has been identified as a critical partner of RIP1 during TNF-R1-induced necroptosis, and this association induces phosphorylation of both proteins (10, 32). Although T cells die by necroptosis through RIP1, RIP3 does not appear to be involved in this context. First, whole-cell extracts of L929 cells undergoing necroptosis show phosphorylation of RIP1, which is not present in untreated L929 or T-cell extracts (Fig. 4B Lower). Second, coimmunoprecipitation experiments fail to show a phospho-RIP1 associating with RIP3 in activated tFADD^−/− T cells at various time points (Fig. 4B). In contrast, a slower-migrating phosphorylated RIP1 band was coimmunoprecipitated with anti-RIP3 antibodies in L929 cells undergoing necroptosis, as shown previously (10, 11). Thus, TCR-induced necroptosis appears to be RIP3-independent.

tFADD^−/− Mice Are More Susceptible to Toxoplasma gondii Infection.

To address the consequence of the lack of FADD and uncontrolled T-cell necroptosis in an immune response setting, we performed Toxoplasma gondii infection studies. T. gondii is an intracellular parasite that causes a robust CD8 T-cell–dependent immune response (33), and CD8-deficient mice (on the C57BL/6 background) die within 20–25 d postinfection (34). We found that the absence of FADD renders mice more susceptible to chronic T. gondii infection. Whereas infection of wild-type C57BL/6 mice with a relatively low dose of parasite led to the death of 50% of mice at 50 d postinfection (Fig. 5A Top), 100% of tFADD^−/− mice died by 31 d postinfection. At the same time, tFADD^−/− mice also exhibit more weight loss than the control group (Fig. 5A Bottom). We subsequently analyzed the serum from mice at days 8 and 14 postinfection and performed cytokine profile studies using the cytokine bead array assay. Interestingly, the tFADD^−/− serum levels of TNF, IL-6, IL-10, MCP-1, and IL-12 were normal (Fig. S6). In contrast, γ-IFN was significantly elevated in the infected tFADD^−/− mice at day 14 postinfection (Fig. 5B). This may be due to increased γ-IFN production by NK cells and antigen-presenting cells in response to the necrotizing T cells and uncontrolled parasite replication during these later stages of the immune response (35–37). Thus, these data suggest that programmed necrosis in activated FADD-deficient T cells might trigger increased levels of γ-IFN, to promote the immune response, as seen in other systems (4, 11), but these cytokines are not sufficient to protect the mice from infection as there is a lack of functional CD8⁺ T cells. Thus, an enhancement of the immune response by necroptosis is context-dependent.

Fig. 5.

tFADD^−/− mice exhibit decreased survival in response to T. gondii infection. (A) tFADD^−/− mice and their littermate controls were infected intraperitoneally with T. gondii. Survival data (top panel) and relative body weight loss (bottom panel) were compiled over a 50-d period. (B) Serum γ-IFN levels of control or tFADD^−/− Toxoplasma-infected mice at day 14 postinfection. Cytokine levels were determined by the cytokine bead array assay and using the corresponding standards.

Discussion

FADD has long been known to be an important mediator of lymphocyte proliferation in addition to its classical role in death receptor-mediated apoptosis. However, the mechanism by which FADD exerts its proliferative function has been elusive. Work done by our lab, as well as others, has pointed to a link between FADD and cell cycle regulation (19–21, 24, 38, 39). Using a FADD knockout model that allows for proper T-cell development, we now show that although FADD does have a role in early cell cycle transition, the dominant function of FADD is inhibition of programmed necrosis during T-cell proliferation.

We have previously shown that FADD phosphorylation occurs upon entry into the cell cycle and that altered phosphorylation of FADD leads to cell cycle defects (20, 21, 39). The phenotype of the tFADD^−/− T cells corroborates this role for FADD in early cell cycle transitions. Activated FADD-deficient T-cell cultures have a larger percentage of cells that have failed to enter the cell cycle and undergo only one cell division. For comparison, T cells from Caspase-8 knockout mice also have proliferative defects that can be rescued with Necrostatin-1 (8, 22), but they do not possess the cell cycle phenotype of the FADD-deficient T cells; Caspase-8 knockout T cells divide at the same rate as control T cells (22). This suggests a cell cycle-related role for FADD that is independent of Caspase-8. Interestingly, the defect of FADD-deficient cell cycle entry can be overcome by itself to a certain extent, as demonstrated by cell cycle entry of FADD-deficient T cells, although their cycling is slower than wild-type T cells. However, accumulation is not restored, and this is particularly evident for CD8 tFADD^−/− T cells. In contrast, FADD(S194D) T cells still succumb to apoptotic cell death in response to strong S-phase impairment and are insensitive to Necrostatin-1 rescue.

Necroptosis is a specific form of programmed necrosis that is executed by RIP1 kinase activity and has been shown to be physiologically relevant in certain human pathologies (5). Although the role of necroptosis in the immune system is less clear, it has been suggested to act as a danger signal to the immune system in certain contexts (4). Furthermore, necroptosis may constitute a checks-and-balances type of mechanism for instances in which apoptosis is ineffective for various reasons. We show here that FADD-deficient T cells fail to proliferate primarily due to an increased rate of programmed necrosis and that inhibition of RIP1 kinase activity by Necrostatin-1 is able to fully rescue the proliferative defect of these T cells. This suggests that FADD is normally a negative regulator of RIP1-dependent necroptosis during the T-cell expansion phase. This is in contrast to Jurkat T cells where, in the absence of Caspase-8, stimulation of the death receptors, specifically Fas, TNF-R1, and TRAIL-R, leads to necroptosis (6) and FADD is specifically required for the Fas- and TRAIL-R-mediated process (6).

Caspase-8 and FADD have been proposed to be involved in a complex that inhibits hyperautophagy during T-cell proliferation, thereby keeping necroptosis in check (8). T cells from transgenic mice expressing a truncated form of FADD (FADDdd) exhibit proliferative defects that can be reversed by inhibition of autophagy. In contrast, autophagy inhibitors cannot rescue tFADD^−/− T-cell proliferation. We speculate that the differences in phenotypes between FADDdd and tFADD^−/− T cells can be attributed to abnormal T-cell development in FADDdd mice (23) which produces an unusual population of mature T cells. Furthermore, the FADDdd protein is not truly a dominant-negative protein, as previously thought. The level of transgenic FADDdd protein is at least 100-fold that of endogenous FADD protein expression (40); expression of FADDdd at physiological levels failed to suppress Fas-induced apoptosis (40). It is not clear whether the defective proliferation of Caspase-8–deficient T cells can be rescued by an autophagic inhibitor or not. However, there are several phenotypic differences between tFADD^−/− and T-cell–specific Caspase-8 knockout (tcaspase-8^−/−) mice. Whereas T-cell development is also normal in tcaspase-8^−/− mice (22, 41), old tcaspase-8^−/− mice suffer from lymphoproliferative and lymphoinfiltrative immune disorder, with many of them dying within 1 y (41). The phenotypic discrepancies between the Caspase-8 and FADD models suggest that Caspase-8 and FADD may not act in the same proliferative/anti-death pathways.

Recent work demonstrates that in certain cell types, RIP3 works in a phosphorylation-dependent manner with RIP1 to induce necroptosis signaled through TNF-R1 (10–12). However, in FADD-deficient T cells undergoing necroptosis, RIP3 does not appear to associate with RIP1. Thus, T-cell necroptosis is dependent upon RIP1 but not RIP3. Furthermore, we are so far unable to block necroptosis in tFADD^−/− T cells by neutralizing Fas, TNF-R1, and TRAIL-R signaling alone or in combination. Our data suggest TCR-induced necroptosis to be death receptor- and autophagy-independent, thus pointing to the existence of an alternate RIP1-dependent necroptotic pathway downstream of TCR signaling. This necroptotic signal may normally act at the end of the immune response to aid in T-cell contraction, but loss of FADD leads to uncontrolled, early induction of necrotic death.

Wild-type T cells, like FADD-deficient T cells, are positively affected by Necrostatin-1 treatment. Necrostatin-1 has also been shown to increase the life span of activated primary macrophages (26). Thus, necroptosis may be a general mechanism for mediating immune system homeostasis and may play a role in abating the late phase of the T-cell response. In this regard, it is possible that an undetectable increase in the steady-state level of necroptosis in tFADD^−/− mice may enhance proinflammatory signaling and contribute to the increased red blood cell and B-cell populations we observe in these mice.

Recently, a genome-wide siRNA screen to identify regulators of necroptosis found many key immune system proteins to be involved, further suggesting that necroptosis is indeed a relevant cell-death mechanism within the immune system (26). Although our tFADD^−/− data show how uncontrolled necroptosis is specifically detrimental to T-cell expansion and function, especially in the case of T. gondii infection, we can speculate the need for necroptosis of other cell types during times of infection in which caspases have been specifically targeted and inactivated. It is also likely that in certain pathological contexts, necroptosis is preferred over apoptosis, as necrotic death increases inflammation and immune cell infiltrates to the affected area. However, our FADD conditional knockout model clearly demonstrates the importance of keeping necroptosis in check within certain cell types, especially T cells, during specific phases of the immune response in which cell death would be counter to a productive immune response.

Materials and Methods

Generation of tFADD^−/− Mice.

LoxP sites were placed around FADD exon 1 (Fig. S1A). Mice generated were crossed to CD4-Cre (42) to generate CD4-Cre/FADD^fl/fl (tFADD^−/−) mice.

T-Cell Analysis.

Purified mature T cells were activated with plate-bound anti-CD3 and anti-CD28. Necrostatin-1 (30 μM) or 3-MA were added at the start of the culture and supplemented when cultures were expanded. 3-MA was also added at 48 h postactivation, instead of at culture initiation. Cells were counted in triplicate per condition each day of culture by trypan blue and analyzed by flow cytometry. For more detailed information regarding T-cell culture conditions, see Materials and Methods in SI Text.

T. gondii Infections.

Parasites of the Prugnaud strain of T. gondii were grown in fibroblasts and harvested by standard protocol (43). Mice were infected intraperitoneally with 400 Pru parasites.

Cytokine Bead Array Assay.

Blood was collected from mice at the indicated days postinfection via tail-vein bleed. The cytokine bead array assay was performed according the manufacturer's protocol and samples were read on an LSR II flow cytometer (BD Biosciences). The data are representative of two independent experiments, with each experiment having 300 replicates of each cytokine reading.

See SI Text for other procedures and additional details on materials and methods.