Recently, it has been demonstrated that stimulated T cells bearing defects in caspase-8 fail to promote nuclear shuttling of NF-κB complexes. Such cells display strikingly similar proliferative and survival defects as T cells lacking Fas-associated death domain protein (FADD) function. We characterized NF-κB signaling in T cells bearing a dominant-negative FADD transgene (FADDdd). Whereas FADDdd T cells displayed proliferative defects following activation, these were not a consequence of aberrant NF-κB signaling, as measured by IKK/IκB phosphorylation and IκB degradation. There were no appreciable defects in nuclear translocation of p65/Rel using ImageStream, a flow-based imaging cytometer. Pretreatment with benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone, a potent caspase inhibitor, also failed to impede canonical NF-κB signaling. Secretion of IL-2 and up-regulation of various activation markers occurred normally. Thus, FADD does not play an essential role in NF-κB activation, suggesting an alternative route by which this adaptor promotes the clonal expansion of T cells.
To gain an understanding into the role of death-receptor signaling components during T cell development, transgenic mice were generated that express a dominantly interfering form of human Fas-associated death domain protein (FADD)3 (FADDdd) (1, 2, 3). The transgene is only expressed in T cells since it is driven by the p56lck proximal promoter that is specifically induced in thymocytes and, in some cases, mature T cells. FADDdd lacks the caspase recruitment DED domain but retains the carboxyl-terminal death domain. Thus, it is hypothesized that FADDdd retains the ability to bind death receptors but not caspase-8 molecules. As anticipated, T cells derived from this mutant strain of mice were refractory to DR-induced apoptosis. One would also anticipate that, due to the apoptotic insensitivity of these T cells, such mutant mice would be prone to the development of lymphoproliferative diseases. However, such diseases have not been found in FADDdd mice, nor in mice expressing the cowpox virus caspase-8 inhibitor CrmA (4, 5). Surprisingly, when stimulated with mitogens, FADD mutant T cells possess acute proliferative defects (1, 2, 3, 6, 7, 8).
T cell activation occurs when the TCR makes contact with MHC-peptide complexes and costimulatory receptors such as CD28 interact with their cognate ligands existent on the surface of APCs. An essential pathway instigated soon after TCR ligation involves the activation of NF-κB. Recently, it was reported that T cells from patients with caspase-8 deficiency (CED) failed to activate NF-κB following TCR stimulation (9). Given that NF-κB is required for T cell proliferation and survival and for the expression of IL-2 and its high-affinity receptor CD25, we sought to determine whether the proliferative and survival defects observed in FADDdd T cells might be due to an analogous deficiency in NF-κB signaling.
Materials and Methods
Mice, cell culture, ELISA, and Western blotting
1017-FADDdd transgenic mice (Tg(Lck-FADD)1Hed) previously described (2) were bred to homozygosity and maintained on the C57BL/6J background. Sex- and age-matched C57BL/6J mice were used as controls. Mice were bred and maintained in accordance with the animal use and care committee at the University of California, Irvine vivarium. Cell culture was performed as described previously (8). T cells were obtained by negative selection with magnetic beads using Abs to mouse MHC class II, CD45, and CD11b (Miltenyi Biotec). Cells were activated using anti-CD28 at 5 μg/ml and plate-bound anti-CD3 at 5 μg/ml. benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (zVAD-FMK) was used at 150 μM unless otherwise noted (Alexis); cells were pretreated for 60 min. at 37°C. For survival studies, T cells from three independent C57BL/6 mice were incubated in triplicate with IL-7 (10 ng/ml) and IL-15 (10 ng/ml) plus DMSO (0.2%) or benzyloxycarbonyl-Val-Ala-Asp (zVAD; 40 μM); medium was changed twice per day along with fresh zVAD. At the indicated times, cells were harvested volumetrically with a FACSCalibur (BD Biosciences) after staining the cells with Annexin VPE and 7-aminoactinomycin D (7AAD). ELISAs for IL-2 expression were conducted using the Ready-SET-Go kit, as per supplied protocol using three independent mice of each genotype (eBioscience). Western blotting using whole cell extracts was performed as described previously (8). The following Abs were obtained from Cell Signaling and used for immunoblotting: phospho-IKK, -IκB, -S6 kinase, -p44/42 MAPK, and total IκB and from Santa Cruz Biotechnology: p65 and Oct-1. Anti-mouse β-actin was used to verify equal protein loading (Abcam). Quantitation of the p65 Western blots was performed using NIH ImageJ software; arbitrary values represent the pixel intensity of each p65 band (minus background intensity) normalized against background corrected nuclear (Oct-1) or cytoplasmic (tubulin) band intensities for each lane.
FACS labeling, microscopy, and ImageStream
FACS analyses were performed as described previously (8). Purified T cells were obtained from three independent mice from each genotype, activated as indicated, and labeled with 10 μM BrdU for the duration of the experiment using the BD Biosciences BrdU labeling system and 7AAD. Cells were fixed for 10 min at room temperature in 4% paraformaldehyde/PBS at 75 × 106 cells/ml. For microscopy, T cells were permeabilized with fixation buffer containing anti-p65 and incubated at 25°C for 20 min, followed by secondary staining with anti-Rb-FITC. Cells were fixed in 1% paraformaldehyde and resuspended in VectaShield plus 4′,6′-diamidino-2-phenylindole (DAPI). For ImageStream 100 (Amnis), cells were isolated, cultured, and stained as above. Before analysis, 7AAD (40 μM) was added to each sample. Image files of 10,000 cells for each sample were acquired. Following data collection, the spatial relationship between the NF-κB and nuclear images was measured using the similarity feature in the IDEAS software package (Amnis). The similarity score (a monotonic function of Pearson’s correlation coefficient between the pixel values of two image pairs) provides a measure of the degree of nuclear localization of p65 by computing the pixel intensity correlation between the anti-p65-FITC and 7AAD images. The scale ranges from negative infinity to positive infinity. Cells with low similarity scores exhibit no correlation of the images (hence, a cytoplasmic distribution of NF-κB), whereas cells with high scores exhibit a positive correlation of the images (hence, a nuclear distribution; see Fig. 4 B).
Results and Discussion
Since FADDdd T cells have defects in both proliferation and survival during cell cycle progression, we wanted to determine the stage during cell cycle progression in which FADDdd T cells manifest these defects. A comparison of continuous BrdU incorporation vs 7AAD staining was undertaken to determine the cycling parameters of FADDdd T cells with representative pseudocolored dot plots shown (Fig. 1,A). Upon stimulation with anti-CD3 and anti-CD28, a large proportion of wild-type cells incorporated BrdU (26%), although at this time, few wild-type cells had completed DNA synthesis (8%; Fig. 1,B). In contrast, very few FADDdd T cells had incorporated BrdU at 24 h (11%) and an even smaller fraction had completed synthesis (3%). At 48 h, most wild-type T cells had incorporated BrdU and many of these cells had completed the cell cycle, returning to G1. An increased fraction of FADDdd T cells incorporated BrdU at this later time, although there were many more cells remaining in G1 than for wild-type cultures. A large proportion of FADDdd T cells possessed subdiploid DNA content, consistent with decreased survival of these cells following activation (8). This failure to induce efficient cell cycle progression was not due to defective IL-2 expression, as assessed by ELISA (Fig. 2,A). These results are similar to those reported for FADD−/− T cells (6). As determined by flow cytometric analyses, we have found that FADDdd T cells induced normal surface expression of CD25, the high-affinity IL-2 receptor, and CD69 upon activation (Fig. 2 B). Similarly normal modulation of CD44 and CD62 ligand was observed (data not shown). Thus, the defects of FADDdd T cells cannot be attributed to aberrant IL-2 production, nor is there any obvious deficiency in signaling pathways proximal to the TCR. Such results are in contrast to those reported for T cells with known defects in NF-κB signaling, including T cells deficient in IKKβ, protein kinase Cθ, Bcl10, CARMA1, and MALT1/paracaspase (10).
Given the normal expression of activation markers and production of IL-2, but the dramatic decrease in cell cycle progression and survival of activated FADDdd T cells, we wished to specifically evaluate NF-κB signaling parameters in these cells. It has been previously reported that FADDdd T cells possess normal NF-κB signaling, but these experiments were conducted with long incubation times (11). Phosphorylated IKK was detected by 15 min in FADDdd T cells, with maximal phosphorylation observed 30 min after activation (Fig. 3 A); there were no apparent differences between wild-type and FADDdd T cells. Since the IKK complex phosphorylates IκB, thus leading to its degradation, we evaluated the phosphorylation status and degradation of IκB. Upon stimulation, there was an increase in the amount of phosphorylated IκB in both wild-type and FADDdd T cells. The level of phosphorylation decreased by 60 min, which correlated with increased degradation of IκB, since total levels of IκB decreased as stimulation progressed. Thus, IκB activation and degradation occurred normally in stimulated FADDdd T cells. Additionally, there were no differences in the phosphorylation status of ERK and S6 kinase, two kinases crucial for cell growth and survival (12), following TCR stimulation.
Caspase inhibitors such as zVAD-FMK impede T cell proliferation after TCR cross-linking, in some cases blocking IL-2 production, suggesting that activated caspases cleave a key component necessary for T cell activation (13, 14, 15). Thus, we determined whether or not blocking caspase activity via pretreatment with zVAD-FMK would inhibit IKK and IκB phosphorylation. Wild-type T cells were pretreated with a saturating level of zVAD-FMK (150 μM) before activation with anti-CD3 and anti-CD28 (Fig. 3,B). zVAD-FMK treatment had no effect on IKK or IκB phosphorylation, suggesting that caspase activity was not required for activation of the IKK complex. It has been reported by Su et al. (9) that, following caspase-8 blockade using zVAD-FMK, IKK phosphorylation was not detected in response to TCR ligation. However, in that study, cells were pretreated with zVAD-FMK twice a day for at least 4 days. We wished to determine what effect such prolonged treatment with zVAD-FMK would have on primary mouse T cells. As a control, cells not treated with zVAD-FMK received an equal volume of DMSO along with IL-7 and IL-15. Activated cells that were subjected to zVAD-FMK treatment for 4 days had no appreciable level of phosphorylated IKK or IκB, whereas control-treated cells demonstrated normal signal-dependent phosphorylation of IKK and IκB (Fig. 3,C). Surprisingly, the levels of endogenous IκB from zVAD-FMK-treated T cells were lower than in the control population, although equivalent levels of protein were loaded. It is possible that long-term zVAD-FMK treatment may have been toxic, since the recovery of cells treated with zVAD-FMK after 4 days was lower than for vehicle control cells (Fig. 3 D).
We next assayed for nuclear translocation of p65 via subcellular fractionation and Western blotting. Upon activation, there was appreciable translocation of p65 into the nucleus, and the extent of translocation between wild-type and FADDdd T cells was similar (Fig. 3,E). p65 localization was also evaluated using epifluorescent microscopy. Following fixation and permeabilization, cells were stained with an Ab against p65 and counterstained with DAPI. Untreated wild-type, FADDdd, and wild-type zVAD-FMK-treated cells all exhibited similar p65 cytoplasmic staining (Fig. 4,A). Following activation, the cells displayed a mostly nuclear staining pattern, demonstrating normal p65 translocation following TCR ligation. In accord with our previous fractionation data (Fig. 3 E), there was no difference in p65 translocation in activated FADDdd T cells compared with wild-type T cells. Similarly, zVAD-FMK-pretreated T cells induced p65 translocation as efficiently as control-treated T cells.
Although epifluorescent microscopy is an appropriate tool for visualizing the localization of p65 following T cell activation, we were concerned that only a small number of cells could be evaluated using this methodology. Furthermore, it is not possible to appropriately quantify the amount of translocation occurring within a mixed population of cells. To overcome these problems, activated T cells were subjected to analysis with the ImageStream quantitative imaging flow cytometer. With this instrument, six digital images (darkfield, brightfield, and four fluorescence colors) of each cell are simultaneously acquired in flow. The images acquired are analyzed using algorithm-based measurements of image-based features to facilitate the objective quantitation of p65 nuclear translocation on a cell-by-cell basis. Following data collection of cells stained with anti-p65 plus 7AAD, the spatial relationship between the NF-κB and nuclear images was measured using the similarity feature (B. E. Hall, T. C. George, and D. M. Coder, manuscript in preparation). Cells with low similarity scores exhibited a cytoplasmic distribution of NF-κB while cells with high scores exhibited a nuclear distribution (Fig. 4,B). By classifying cells in this manner, it is possible to assess the proportion of cells with a primarily nuclear staining pattern in a large population. These data were plotted as histograms representing the relative translocation levels (Fig. 4,C) and net translocation was determined for each cell type (Fig. 4 D). Although these net translocation values may be an underestimate of the proportion of cells with at least some nuclear p65, there were no apparent differences between these populations. Thus, p65 translocation did not depend on FADD, nor did it require caspase activity immediately following TCR cross-linking.
We have evaluated early NFκB activation parameters in T cells derived from FADDdd transgenic mice and have found no apparent defect in IKK activation or IκB degradation. Furthermore, we have determined that there was normal p65 nuclear translocation by western blotting, epifluorescent microscopy, and ImageStream multispectral imaging flow cytometry. Although T cells from patients carrying mutations in caspase-8 failed to activate NF-κB following TCR activation (9), our results demonstrate that FADD itself likely does not participate in such signaling. Although it is conceivable that FADDdd T cells might maintain some basal level of caspase-8 activity due to incomplete inhibition, it remains that these T cells are highly refractory to proliferation following TCR stimulation (Fig. 1). An alternative explanation may be that there are striking differences in the signaling routes undertaken by activated murine and human T cells. Such a possibility awaits further elucidation of the means by which caspase-8 might promote the activity of the IKK complex.
Recently, Su et al. (9) demonstrated that caspase-8 interacts directly with the CARMA1/Bcl10/MALT1 complex (9). Using immunoprecipitation, they found that FADD and caspase-8 both bound to the CARMA1/Bcl10/MALT1 complex, but that FADD remained in this complex only transiently. FADD was not detected in the complex at the time of IKK phosphorylation, suggesting that FADD is dispensable for IKK activation, while caspase-8 is required. Thus, it remains uncertain whether or not FADD and caspase-8 act in concert or independently to promote T cell proliferation. Although the simplest explanation is that FADD and caspase-8 constitute analogous signaling platforms following Ag receptor and death receptor stimulation, equally likely is that they serve independent functions. Thus, it is tenable that caspase-8 may be activated in an entirely FADD-independent fashion for the propagation of a NF-κB signal following Ag receptor ligation. However, our results using zVAD-FMK call into question the putative requirement for caspase-8 enzymatic activity in mediating this process in murine T cells. A 1-h pretreatment using an inhibitor concentration in excess of that sufficient to block caspase cleavage and apoptosis did not prevent the activation of the NF-κB pathway, although a 4-day pretreatment apparently did just this (Fig. 3,C). We attribute the defective NF-κB response in long-term-treated cultures to the aforementioned zVAD-FMK toxicity (Fig. 3 D). Alternatively, it has been demonstrated that caspase inhibitors such as zVAD-FMK not only arrest apoptosis, but promote autophagic death by blocking basal caspase-8 activity (16). It is possible that the 4-day pretreatment with zVAD-FMK caused these cells to undergo autophagy, potentially interfering with multiple signaling cascades, including those that converge onto NF-κB.
It is clear that many of the proteins involved in apoptotic pathways have emerging roles in survival and proliferation. Although it appears that caspase-8 and FLIPL may promote NF-κB signaling in some contexts (17, 18, 19), FADD is likely involved in an alternative cell cycle checkpoint pathway that coordinates lymphocyte cell division and survival (20, 21) and is not a priori involved in promoting survival via NF-κB activation. Although at first these functions might seem idiosyncratic, such a checkpoint could provide for high-fidelity control over lymphocyte homeostasis.
We thank Drs. Naomi Morrissette, Stephen Hedrick, and Andreas Strasser for thoughtful comments regarding this manuscript.
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 the National Institutes of Health Grant AI050606 (to C.M.W.) and by Training Grants T32GM007311 and T32CA09054 (to A.F.A. and J.C.S., respectively).
Abbreviations used in this paper: FADD, Fas-associated death domain protein; zVAD-FMK, benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone; zVAD, benzyloxycarbonyl-Val-Ala-Asp; IKK, IκB kinase; 7AAD, 7-aminoactinomycin D.