Essential Role for Caspase-8 in Toll-like Receptors and NF κ B Signaling

In addition to its pro-apoptotic function in the death receptor pathway, roles for caspase-8 in mediating T-cell proliferation, maintaining lymphocyte homeostasis, and suppressing immunodeficiency have become evident. Humans with a germline pointmutationofCASPASE-8havemultipledefectsinTcells,B cells, and NK cells, most notably attenuated activation and immunodeficiency. By generating mice with B-cell-specific inactivation of caspase-8 ( bcasp8 (cid:2) / (cid:2) ), we show that caspase-8 is dispensable for B-cell development, but its loss in B cells results in attenuated antibody production upon in vivo viral infection. We also report an important role for caspase-8 in maintaining B-cell survival following stimulation of the Toll-like receptor and In response to caspase-8 is to a containing (cid:3)(cid:4) , resulted in delayed NF (cid:1) B nuclear translocation and impaired NF (cid:1) B transcriptional activity. Our study supports dual roles for caspase-8 in apoptotic and nonapoptotic functions and demon-stratesitsrequirementforTLRsignalingandintheregulationof NF (cid:1) B function.

tion by self-cleavage to form an activated caspase-8 tetramer complex in a process termed "proximity-induced activation" (5). Once activated, caspase-8 mediates apoptosis through a cell-specific type I or type II DR pathway. In type I cells, active caspase-8 at the DISC can directly process and activate caspase-3, leading to apoptosis. In type II cells, caspase-8 processes the BH3-only molecule Bid to form a truncated Bid molecule that translocates to the mitochondrial membrane and triggers cell death through the mitochondria (6).
In addition to their apoptotic roles, a new paradigm for caspases in non-apoptotic roles has become clear. In various studies, activation of caspases was detected in non-apoptoticactivated lymphocytes (7)(8)(9)(10). Additionally, lymphocyte activation and proliferation was defective in the presence of selective caspases inhibitors (11)(12)(13)(14). T-cell-specific deletion or overexpression of a dominant negative form of Fas-associated death domain protein (FADD), another DISC component, demonstrated its requirement for proper T-cell activation and IL2-dependent proliferation (15)(16)(17)(18). Together these findings implicate caspases and other DISC components in non-apoptotic functions.
Until recently, embryonic lethality prevented the generation and phenotypic characterization of adult mice missing caspase-8, necessitating tissue-specific inactivation (19 -21). We recently reported the conditional inactivation of caspase-8 in the T-cell lineage (tcasp8 Ϫ/Ϫ ) and identified important novel non-apoptotic roles for caspase-8 in maintaining peripheral T-cell homeostasis, in mediating expansion of T-cells upon in vitro activation, and in inducing an effective in vivo T-cell immune response against viral infections (20). Studies of human patients with a caspase-8 point mutation also support a non-apoptotic role for caspase-8. These patients are immunodeficient, show reduced activation of B, T, and NK cells in vitro, exhibit decreased levels of circulating antibodies and impaired immunoglobulin production in response to antigenic stimulation (22). Impaired activation of T, B, and NK cells from these patients was associated with defective NFB nuclear translocation (23). In response to TCR stimulation of T cells, caspase-8 contributes to the formation of the activation receptor-induced signalosome (ARIS) that contains CARMA1-BCL10-MALT1 and the IB kinase (IKK) complexes. Loss of caspase-8 impairs ARIS formation leading to defective NFB signaling in response to TCR stimulation. Toll-like receptors are important components of innate immunity and provide a first-line of defense against microbial infections. Binding of specific TLRs by their appropriate antigens activates B cells, induces their antigen-presenting ability, and stimulates the production of neutralizing antibodies (24,25).
A link between TLR pathways and DISC components has been identified in recent studies of caspase-8, FADD, Flip (FLICE-like inhibitory protein), and RIP-1 (receptor-interacting protein-1) (29 -32). Flip is recruited to the DISC and functions as an endogenous dominant-negative molecule by inhibiting FADD-caspase-8 interactions and thereby inhibits DR apoptosis (33). Deletion of FADD or FLIP in mouse embryonic fibroblats (MEFs) enhanced NFB activation in response to LPS stimulation (29,30), whereas RIP-1 deficiency resulted in decreased NFB activation or survival following TLR3 or TLR4 stimulation, respectively (31,32). The molecular mechanisms linking FADD, Flip, and RIP-1 with TLR pathways remain to be identified.
In this study, we have investigated the roles of caspase-8 in development, apoptosis, and proliferation of B lymphocytes. As expected, caspase-8-deficient B cells were resistant to CD95mediated killing; however unlike T lymphocytes (20), we found that caspase-8 is not required for development and homeostasis of B lymphocytes. Our study identifies an impaired response of caspase-8-deficient B cells to TLR4 and links this defect to a delayed nuclear translocation of NFB, and defective transcription of its target genes in the absence of caspase-8. We also report that loss of caspase-8 leads to reduced phosphorylation of p65 at serine 536, which is known to regulate p65 nuclear translocation (34). Finally, in response to TLR4 stimulation, caspase-8 was found to be transiently recruited to a complex that contains IKK␣␤ thereby placing this caspase in a signaling complex important for cell survival and immune responses.

EXPERIMENTAL PROCEDURES
Mice-Casp8-conditional mutant mice (20) were intercrossed with CD19Cre transgenic mice (35) to obtain bcasp8 mice. All mice studied were in a mixed 129/J ϫ C57BL/6 genetic background and were genotyped by PCR (primer sequences and PCR conditions available upon request). Controls used in the different experiments correspond to either wild-type, casp8 fl/wt CD19-Cre, or CD19-Cre mice. The former two genotypes produced similar results compared with the wild-type controls. Mice used for each experiment were littermates. All experiments were performed in compliance within the Ontario Cancer Institute Animal Care Committee Guidelines.
B-cell Purification and Thymidine Incorporation-For all the experiments, B cells were purified by negative selection or selected by activated cell sorting (FACS). For negative selection, splenocytes were subjected to red blood cell lysis, depleted of macrophages using anti-rat IgG magnetic beads (Dynabeads, Dynal) following anti-CD11b and anti-Gr-1 (eBioscience) labeling. T cells were eliminated using anti-Thy 1. . Lipoteichoic acid (LTA) and ultrapure LPS (invivo-Gen) were used at 10 g/ml and 5 g/ml, respectively.
Cell Cycle and CFSE Analysis-For cell cycle analysis, purified B cells were fixed in 70% ethanol and stained with 5 g/ml of propidium iodide as previously described (20). Purified splenic B cells (10 6 ) were collected in serum-free medium and incubated with CFSE (Molecular Probes-5 M) at 37°C for 10 min. Following three washes in Iscove's medium, 10% fetal calf serum, the CFSE-labeled cells were stimulated for 24, 48, and 72 h with LPS (10 g/ml). For each time point, cells were stained for survival using 7-amino-actinomycin D (7-ADD), and the levels of CFSE incorporation determined by FACS analysis.
Immunofluorescence-Cells (10 5 ) were used for cytospin centrifugation. Cytospin specimens of B cells were fixed with 4% PFA for 30 min at room temperature. Cells were incubated with 1:50 dilution of anti-p65 Ab (C-20, Santa Cruz Biotechnology) for 45 min at room temperature. Labeling was revealed using anti-rabbit-conjugated to fluorescein isothiocyanate (Jackson Immunoresearch). Cells were stained with Hoescht for 5 min and then mounted with Fluoromount (Southern Biotec). Images were collected on a Leica TCS-NT/SP confocal microscope (Leica Microsystems) using a ϫ63 oil immersion objective NA 1.32, zoom X.

RESULTS
To study the function of caspase-8 in B lymphocytes, mice carrying the caspase-8 mutation targeted to the B-cell lineage (casp8 fl/fl ;CD19Cre, also called bcasp8 Ϫ/Ϫ ) were generated by crossing mice homozygous for the caspase-8-floxed allele (casp8 fl/fl ) (20) with mice expressing the Cre recombinase under the control of the B-cell-specific CD19 promoter (CD19Cre) (35). Southern blot analysis indicated the highly efficient deletion of caspase-8 gene in purified peripheral B cells from bcasp8 Ϫ/Ϫ mice (Fig. 1A). Western blot analysis of serial dilutions of protein extracts prepared from purified B cells confirmed the loss of caspase-8 protein in bcasp8 Ϫ/Ϫ cells (Fig. 1B).
Analysis by flow cytometry of cell subpopulations in bone marrow (BM), spleen, and lymph nodes (LN) did not show any differences between bcasp8 Ϫ/Ϫ mice and control littermates ( Fig. 1C and data not shown). The proportion of T and B cells in LN and spleens and total lymphocyte numbers in thymus, spleen, LN, and BM were similar in control and bcasp8 Ϫ/Ϫ mice ( Fig. 1D and data not shown). Therefore, unlike its function in the T-cell lineage, caspase-8 is not essential for the maintenance of B-lymphocyte homeostasis.
To confirm the functional deletion of caspase-8, we assessed the sensitivity of casp8 Ϫ/Ϫ B cells to CD95-induced apoptosis. Purified peripheral B cells were activated in vitro for 3 days with anti-IgM/IL4, and then treated with Fas ligand, in the presence or absence of cycloheximide, an inhibitor of protein biosynthesis known to potentiate Fas killing (37). Annexin-V/PI staining demonstrated that while B cells from control mice were sensitive to CD95-induced killing, bcasp8 Ϫ/Ϫ cells were resistant (Fig. 1E). These results confirm the essential role for caspase-8 in DR-mediated cell death.
Vesicular stomatitis virus (VSV) infection in mice induces a primary infection phase, which consists of a T-cell-independent response characterized by the production of neutralizing IgM antibodies by B cells. The second phase of the infection,  which is T-cell-dependent, leads to the production of IgG immunoglobulin (36,38). We investigated the role of caspase-8 in the in vivo B-cell response to VSV infection. Control and bcasp8 Ϫ/Ϫ mice were subjected to intravenous immunization with 2 ϫ 10 6 PFU of VSV, and the production of VSV-neutralizing IgM and IgG antibodies was measured 4, 8, and 12 days later (39). While the levels of IgM antibodies directed against VSV were not affected in bcasp8 Ϫ/Ϫ mice, the production of VSV-neutralizing IgG was reduced in bcasp8 Ϫ/Ϫ mice (Fig. 2). This finding is consistent with the increased susceptibility of human patients deficient for caspase-8 to viral infections (22). These findings suggest that caspase-8 deficiency in B cells leads to impaired in vivo responses to VSV infection.
B cells are essential components of the innate immune response against microbial infection (24). TLR signaling is important for the activation of B cells and other components of the immune response by microbes (40). Given the connection between DISC components and TLR signaling (29 -32), we investigated the effect of caspase-8 deficiency on TLR signaling. This was accomplished by treating purified B-cell populations with antigens specific for different TLRs. B-cell expansion was measured by [ 3 H]thymidine incorporation following treatment with increasing doses of different Escherichia coli-derived LPS, ultrapure LPS, CpG, LTA, or poly (I:C) to identify TLR4, TLR9, TLR2, and TLR3 responses, respectively. Whereas the stimulation by CpG consistently produced similar B-cell expansion in the control and bcasp8 Ϫ/Ϫ backgrounds, responses of casp8 Ϫ/Ϫ B cells to poly(I:C), LPS ( Fig. 3A and supplemental Fig. S1), and LTA (supplemental Fig. S2) were reduced compared with controls, demonstrating that caspase-8 does indeed play a role in TLR2, TLR3, and TLR4 signaling. Similarly, CFSE labeling indicated defective expansion of casp8 Ϫ/Ϫ B cells in response to LPS and poly(I:C) stimulation ( Fig. 3B and data not shown). These defects could not be attributed to reduce TLR3 and TLR4 protein levels in bcaps8 Ϫ/Ϫ cells as their cell surface expression levels comparable to controls (data not shown). No major defect in proliferation of bcasp8 Ϫ/Ϫ B cells was observed in response to anti-IgM, anti-IgM, and IL4, or anti-IgM and anti-CD40 compared with wild-type controls (data not shown).
Cell cycle and cell death were analyzed in response to TLR4 stimulation to characterize the lack of expansion of LPS-stimulated casp8 Ϫ/Ϫ B cells. Based on the PI-stained cell cycle profiles and the quantification of the sub-G 1 population, the defective expansion of casp8 Ϫ/Ϫ B cells in response to TLR4 stimulation was attributed to increased cell death (Fig. 4A). Annexin-V/PI staining experiments were also performed and indicated that the increased cell death of TLR4 stimulated casp8 Ϫ/Ϫ B cells was attributed to apoptosis (Fig. 4B). Apoptotic cell death was further confirmed by electron microscopy (Fig. 4, C-F) and increased caspase-3 processing in LPS-stimulated casp8 Ϫ/Ϫ B cells compared with controls (data not shown). Together, these data suggest that caspase-8 is required for the activation of B cells and maintaining their survival in response to TLR3 and TLR4 stimulation.
The IKK complex, composed of IKK␣, ␤, and ␥ or NEMO plays an important role in mediating NFB nuclear translocation in response to TLR (46). We therefore tested whether caspase-8 associates to this complex upon TLR4 activation.   MARCH 9, 2007 • VOLUME 282 • NUMBER 10

JOURNAL OF BIOLOGICAL CHEMISTRY 7419
Immunoprecipitation was performed using anti-IKK␣␤ and/or anti-caspase-8 antibodies on protein extracts derived from untreated or LPS-stimulated casp8 Ϫ/Ϫ or control B cells. These studies indicated that caspase-8 co-immunoprecipitates with IKK␣␤ following 30 min of LPS stimulation of control B cells (Fig. 5A). The specificity of the IKK␣␤ immunoprecipitations was shown by Western blot using anti-IKK␣␤ antibodies. The recruitment of caspase-8 to a complex containing IKK␣␤ in response to LPS stimulation was confirmed in 3T3-immortalized fibroblasts. As shown in Fig. 5B, IKK␣␤ association to caspase-8 in 3T3 was detected as early as 4 min following TLR4 activation and appeared to be optimal at 7-min post-LPS stimulation. Thus, our data identify that TLR4 stimulation drives caspase-8 transient recruitment to IKK␣␤, an important regulator of NFB functions.
NFB is an important downstream target of IKK␣␤ and its activation, in response to TLR4 stimulation, leads to transcriptional activation of a subset of target genes. Thus, the effect of caspase-8 mutation on TLR4-induced NFB transcriptional activation was assessed using real time PCR to quantify the transcription levels of the NFB target genes including IL6, TNF-␣, IFN-␤, and IP-10. cDNA, representative of total cellular mRNA, was prepared from cell-sorted control and casp8 Ϫ/Ϫ B cells activated with LPS for 0, 2, 4, and 7 h. As expected, increased expression of these genes was observed in B cells after treatment with LPS; however, the level of transcriptional induc-

Caspase-8 Role in TLR and NFB Signaling
7420 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 282 • NUMBER 10 • MARCH 9, 2007 tion of these genes over control was significantly lower in casp8 Ϫ/Ϫ B cells (Fig. 5C). Therefore, caspase-8 is important for NFB transcriptional activation in response to TLR4-stimulated B cells.
To investigate the mechanisms leading to the defective NFB-mediated responses of TLR4-stimulated casp8 Ϫ/Ϫ B cells, we focused on known signaling events downstream of LPS/TLR4 activation, including Akt, MAPK-p38, and IkB␣ phosphorylation. LPS stimulation of control and casp8 Ϫ/Ϫ B cells for 30 min, 1 h, and 2 h showed comparable levels of expression and phosphorylation of Akt and MAPK-p38 ( Fig. 6A and data not shown). Phosphorylation of IB␣ results in its proteasome-mediated degradation followed by nuclear translocation and transcriptional activation of NFB (44). Expression levels and phosphorylation of IkB␣ were similar in casp8 Ϫ/Ϫ and control B cells at all time points after LPS stimulation (Fig. 6A).
To further investigate the mechanisms for the defective TLR4-induced NFB transcriptional activation in casp8 Ϫ/Ϫ B cells, nuclear translocation of the NFB-p65 subunit was assessed by immunofluorescence using control and casp8 Ϫ/Ϫ B cells stimulated with LPS for 0, 30, 45, 60, and 120 min. Similar to controls, NFB-p65 was almost exclusively cytoplasmic in untreated casp8 Ϫ/Ϫ B cells (Fig. 6, B and C). However, LPS-induced NFB-p65 nuclear translocation was delayed in casp8 Ϫ/Ϫ B cells at 30, 45, 60, and 120 min post-LPS treatment relative to controls (Fig. 6, B and C). Similarly, Western blot analysis of cytoplasmic and nuclear fractions from bcasp8 Ϫ/Ϫ B cells and their controls confirmed a delayed LPS-induced nuclear translocation of NFB-p65 in the absence of caspase-8 (Fig. 6D). Similar delayed NFB-p65 nuclear translocation was observed in bcasp8 Ϫ/Ϫ B cells in response to TLR3 stimulation (supplemental Fig. S3).
Phosphorylation of NFB-p65 at serine 536 is known to play a major role in its nuclear translocation and transactivation (34), and therefore we assessed the effect of loss of caspase-8 on this NFB-p65 phosphorylation. Interestingly, the level of phosphorylated NFB-p65 at serine 536 was reduced in both cytoplasmic and nuclear fractions from LPS-stimulated bcasp8 Ϫ/Ϫ B cells compared with controls (Fig. 6D). Therefore defective phosphorylation of NFB-p65 at serine 536 could potentially contribute to the NFB-impaired shuttling and function in LPS-stimulated casp8 Ϫ/Ϫ B cells.

DISCUSSION
CASPASE-8 deficiency has been recently associated with human diseases. These diseases include immunodeficiency, cancer, and metastasis (22). Characterization of CASPASE-8 function in immunodeficient patients has demonstrated that in addition to its apoptotic functions, caspase-8 also carries important non-apoptotic functions.
In this study, we report important roles for caspase-8 in the regulation of NFB function in response to TLR4 stimulation. We have previously reported that caspase-8 is dispensable for thymocyte development but required for peripheral T-cell homeostasis and T-cell survival in response to activation stimuli (20). Deletion of caspase-8 in early lymphoid progenitors was also reported to cause early differentiation arrest indicating its important role in early hematopoietic development (21). Similarly, deletion of caspase-8 in all hematopoietic cells using Vav-iCre transgenic mice that express Cre in early hematopoietic progenitors including T-and B-cell progenitors, results in embryonic lethality. 4 Our present study demonstrates that CD19-Cre-mediated caspase-8 deletion in B lymphocytes does not affect B-cell development or homeostasis: B-lymphocyte subpopulations were similar in bcasp8 Ϫ/Ϫ mice compared with controls. We also identified defective in vivo production of neutralizing antibodies, one of the main functions of B cells, in the absence of caspase-8 in B cells. Because helper T cells in bcasp8 Ϫ/Ϫ mice express caspase-8 and are fully functional, the deficient immunoglobulin production following VSV infection of bcasp8 Ϫ/Ϫ mice is likely a result of defects in (i) T/B-cell interactions, (ii) immunoglobulin switch, or (iii) B-cell expansion. Our present study supports a key role for caspase-8 in the immune response of B cells and for the proper production of pathogen-neutralizing antibodies.
Toll-like receptors are important components of innate immunity and provide a line of defense against microbial infections. Stimulation of specific TLRs induces a distinct pattern of 4 B. Lemmers, data not shown.

Caspase-8 Role in TLR and NFB Signaling
gene expression, which leads to antigen-specific acquired immunity and activation of the innate immune response pathway. Caspase-8-interacting proteins such as FADD and c-FLIP, or their downstream responsive molecules such as RIP, have been implicated in the regulation of the signaling by some TLR such as TLR3 and TLR4 (29 -32). TRAIL-R, known to induce apoptosis and to form a death signaling complex with FADD, and caspase-8 was reported recently to negatively regulate TLR2, TLR3, and TLR4 signaling in macrophages and BM-derived dendritic cells (47). Our study demonstrates that caspase-8 plays an essential role in B-cell activation and expansion in response to TLR2, TLR3, TLR4, but not TLR9.
In contrast to a recent study by Beisner et al. (48), our data support a role for defective NFB signaling in the impaired response to TLR3 and TLR4 stimulation of B cells deficient for caspase-8. A possible explanation for this discrepancy is that we have analyzed the effect of caspase-8 loss on NFB-p65 nuclear translocation at earlier time points post-TLR3 and -TLR4 stim-ulation. However, Beisner et al. have analyzed this effect at much later time points when the delay of NFB-p65 nuclear translocation is not obvious anymore. Thus, similar to human B cells from caspase-8 mutant patients (23), loss of caspase-8 in mouse B cells impairs nuclear translocation of NFB-p65.
TLR4 stimulations activate two major signaling pathways, both dependent on NFB-mediated transactivation of genes encoding pro-inflammatory cytokines and survival molecules, among other immune mediators. One pathway is a MyD88-dependent and involves the sequential activation of IRAK-4, binding of TRAF6 to phospho-IRAK1, activation of AP-1 through the MAP kinase pathway, and NFB nuclear translocation (49). This pathway induces the early phase of NFB-mediated transactivation of genes including IL6 and TNF-␣. The second LPS-activated pathway is MyD88-independent, and activates the late phase of NFB-mediated gene expression, driving the production of IFN-␤ and the IFNinducible gene IP-10 (50).
Our real-time PCR analysis identified that in response to TLR4 stimulation, casp8 Ϫ/Ϫ B cells exhibit defective expression of early and late NFB-dependent pro-inflammatory cytokines and other immune mediators. We have also identified a temporal delay in the nuclear translocation of NFB-p65 in TLR4stimulated casp8 Ϫ/Ϫ B cells. Delayed NFB nuclear translocation has been shown to compromise NFB-mediated gene expression, and cell function upon activation (51). These observations highlight the importance of the delayed nuclear translocation of NFB in casp8 Ϫ/Ϫ B cells, in causing the defective production of inflammatory cytokines upon LPS treatment. Thus, we propose a model whereby caspase-8 is involved in the timely activation of NFB downstream signals in response to TLR4 activation. In the absence of caspase-8, NFB signaling in response to TLR4 activation is not blocked, but delayed. Nevertheless, this delay is sufficient to impair the transduction of NFB survival signals in TLR4-stimulated caspase-8-deficient B cells.
Interestingly, we have also identified that in response to TLR4 stimulation, caspase-8 transiently associates with the IKK␣␤, proteins important for NFB signaling (52). The IKK complex is responsible for the phosphorylation of various substrates including NFB-p65 in response to various stimuli and Hoescht (red) from control and bcasp8 Ϫ/Ϫ B cells stimulated with LPS for 0, 30, 45, 60, and 120 min. Data shown are representative of five independent experiments. C, percent of nuclear NFB as measured by anti-p65 nuclear staining as in B. D, Western blot analysis of the levels of NFB-p65 in cytoplasmic and nuclear fractions from LPS-stimulated purified B cells from bcasp8 Ϫ/Ϫ and control mice. B cells were stimulated with LPS as in A, and cell fractions prepared and used for Western blots with anti-p65, anti-p65 Ser 536 , anti-PARP, and anti-actin. including pro-inflammatory signals (52). Interestingly, we have identified that in response to LPS/TLR4 stimulation, caspase-8 transiently associates with the IKK complex. We also demonstrate that caspase-8 deficiency in TLR4-stimulated caspase-8deficient B cells leads to decreased phosphorylation of NFB-p65 at serine 536. This phosphorylation is mediated by the IKK complex, modulates NFB nuclear translocation, and therefore, its impairment could contribute to the defective NFB nuclear translocation and functions observed in TLR4-stimulated bcasp8 Ϫ/Ϫ cells. Our findings support a role for defective NFB signaling in the impaired responses to TLR4 activation in the absence of caspase-8 and provide a molecular mechanism for the immunodeficiency of caspase-8 mutant patients.