Clonal expansion of T cells is vital to adaptive immunity, yet this process must be tightly controlled to prevent autoimmune disease. The serine/threonine kinase death-associated protein kinase-related apoptosis-inducing kinase 2 (DRAK2) is a negative regulator of TCR signaling and sets the threshold for the activation of naive and memory T cells and selected thymocytes. Despite enhanced T cell activation, Drak2−/− mice are resistant to experimental autoimmune encephalomyelitis, an autoimmune demyelinating disease that resembles multiple sclerosis. However, the basis for this autoimmune resistance is currently unknown. In this study, we show that, in the absence of DRAK2 signaling, T cells require greater tonic signaling for maintenance during clonal expansion. Following stimulation, Drak2−/− T cells were more sensitive to an intrinsic form of apoptosis that was prevented by CD28 ligation, homeostatic cytokines, or enforced Bcl-xL expression. T cell-specific Bcl-xL expression also restored the susceptibility of Drak2−/− mice to experimental autoimmune encephalomyelitis and enhanced thymic positive selection. These findings demonstrate that DRAK2 is selectively important for T cell survival and highlight the potential that DRAK2 blockade may lead to permanent autoimmune T cell destruction via intrinsic apoptosis pathways.
Several parameters regulate a T cell’s decision to proliferate, including amount and avidity of TCR stimulation, presence of costimulation, and duration of T cell:APC interactions (1, 2, 3, 4). Peripheral T cell tolerance to self is, in part, mediated through low-avidity interactions between the TCR and self-Ag presented by nonactivated APCs that lack sufficient expression of costimulatory molecules, resulting in T cell anergy, clonal deletion, or T cell suppression (4, 5, 6, 7). When these signals are disrupted, the threshold for T cell activation is altered, often allowing inappropriate initiation of T cell responses to self-Ag and ultimately resulting in autoimmune disease. Consistent with this, exacerbated autoimmune susceptibility is a common phenotype observed in mice lacking negative regulatory molecules such as Cbl, GRAIL, MGAT5, and PTEN (8, 9, 10, 11, 12).
After an infection is cleared, most T cells generated during the clonal expansion phase are removed through either of two apoptotic pathways, termed activation-induced cell death (AICD)4 and activated cell autonomous death (ACAD) (13). AICD is thought to depend on an extrinsic form of apoptosis induced by the ligation of death receptors such as Fas on the surface of the T cell (14). Patients with mutations in Fas display increased survival of lymphocytes and develop a disease termed autoimmune lymphoproliferative syndrome (15). Also, mice with mutations in Fas (lpr) or Fas ligand (FasL) (gld) develop lymphadenopathies and an accumulation of autoreactive T cells in the periphery, indicating the importance of Fas-dependent AICD in lymphocyte homeostasis as well as peripheral tolerance (16, 17).
Although many of the molecular mechanisms behind AICD have been elucidated, much less is known about ACAD. ACAD is thought to depend on an intrinsic form of cell death similar to that induced by cytokine withdrawal. This form of death by neglect is regulated by the balance of pro (Bim, Bad, Bax, Bak)- and anti-apoptotic (Bcl-2, Bcl-xL, and inhibitor of apoptosis (IAP) family of proteins) factors, which themselves are subject to control by tonic signaling (18). Cytokine withdrawal results in reduced transcription of Bcl-2, Bcl-xL, and the cellular IAP family of caspase inhibitors (14, 15, 19). Reduced levels of these proteins relieve inhibition of Bim and Bad, allowing activation of Bax and Bak that can go on to initiate a mitochondrial form of apoptosis. T cells overexpressing Bcl-2 or Bcl-xL are resistant to this form of death, whereas Bim sensitizes activated T cells to apoptosis following viral clearance (20, 21, 22).
Factors that regulate T cell proliferation have also been shown to influence subsequent AICD and ACAD. Strong TCR stimulation induces genes not only involved in proliferation, but also in cell death such as Fas, FasL, and Bim (23, 24). Prolonged culture in high IL-2 concentrations leads to decreased expression of the caspase inhibitor c-FLIP, thus sensitizing these proliferating T cells to apoptosis (23). Recent evidence suggests that strong TCR stimulation not only initiates proliferation, but also encourages T cells to undergo AICD. For example, T cells from mice lacking an inhibitor of calcineurin, calcipressin-1 (Csp1), succumb to enhanced induction of FasL surface expression in response to TCR ligation, resulting in a reduced proliferative capacity due to exacerbated AICD (25). Although costimulatory signals clearly reduce the activation necessary for the commitment of naive T cells to clonal expansion, signals derived via costimulatory receptors are also crucial for activated T cell survival. CD28, 4-1BB, and OX40 promote activated T cell survival through sustained activation of Akt, which induces Bcl-xL and Bcl-2 expression (26, 27, 28, 29, 30, 31, 32, 33). Additionally, active Akt may also serve to limit apoptotic signaling by directly phosphorylating and inactivating Bad, as well as by interfering with the function of FOXO factors (34, 35). The latter may be crucial for dictating the apoptotic sensitivity of T cells, because a number of proapoptotic genes, including Bim, are transcriptionally regulated by FOXO factors (35, 36). It is clear that the coordination of survival and proliferative signaling, although highly complex, is essential for maintaining immune homeostasis and tolerance (37).
Death-associated protein kinase (DAPK)-related apoptosis-inducing kinase 2 (DRAK2) is a serine/threonine kinase distantly related to the DAPK family, a group of serine/threonine kinases thought to potentiate apoptosis (38). However, a direct role for DRAK2 in modulating apoptosis is a matter of controversy; ectopic expression of DRAK2 has been shown to induce apoptosis, but Drak2−/− mice display no discernable apoptotic deficits (38, 39, 40, 41). Although expressed at low levels in various tissues (42, 43), DRAK2 expression is highly enriched in lymphoid organs (40). DRAK2 negatively regulates TCR-mediated calcium mobilization and IL-2 production, and restricts clonal expansion following weak TCR stimulation. Consistent with these findings, Drak2−/− T cells initiate a proliferative response following suboptimal costimulation (40). However, unlike mice lacking other negative regulators of T cell activation, Drak2−/− mice do not develop overt signs of spontaneous autoimmune disease and display resistance to experimental autoimmune encephalomyelitis (EAE) following myelin oligodendrocyte (MOG) peptide immunization (40, 44). Given the paradoxical finding that mice lacking DRAK2 are refractory to autoimmune diseases, but possess hyperactive T cells, we hypothesized that DRAK2 may serve to maintain the survival of proliferating T cells. In this study, we provide evidence that heightened T cell apoptotic sensitivity leads to autoimmune resistance in Drak2−/− mice, suggesting that DRAK2 is an essential mediator of immunological tolerance.
Materials and Methods
Mice, Abs, and reagents
Drak2−/− mice bred onto the C57BL/6J background were generated, as previously described (40). Bim−/− mice were obtained from The Jackson Laboratory. Lckpr-Bcl-xL transgenic (Tg) mice were obtained from the laboratory of C. Thompson, University of Pennsylvania, Philadelphia, PA (22). Abs against CD4 (allophycocyanin, biotin), CD8 (biotin), B220, MHCII, CD11b, CD3, CD28, CD25, and IL-2 (clones JES6-1A12 and JES6-5H4), and streptavidin-PerCP were from eBioscience. Annexin V (PE, allophycocyanin) and Abs against CD4 (PE), CD8 (FITC), and CD69 were from Caltag Laboratories. Abs against CD8 (PerCP), Vβ7 (biotin), and Vβ8 (PE, FITC, biotin) were from BD Pharmingen. Staphylococcal enterotoxin B (SEB) was from Toxin Technology. The anti-DRAK2 mAb was purchased from Cell Signaling Technology.
Splenic T cells were purified by negative selection using biotinylated Abs against B220, MHC-II, and CD11b, followed by separation using streptavidin-conjugated magnetic beads (Miltenyi Biotec). Typical purity was greater than 95%. Wild-type splenic APCs were purified by depletion of T cells using biotinylated Abs against CD3, CD4, and CD8, followed by separation using streptavidin-conjugated magnetic beads.
Thymidine incorporation, cell recovery, and IL-2 production
Purified T cells were cultured in round-bottom wells with purified, mitomycin C-treated wild-type APCs and various concentrations of SEB. Proliferation was measured in triplicate by [3H]thymidine incorporation during the last 18 h of culture. Cell recovery was measured by timed FACS collection of cells stained with biotinylated anti-Vβ7 and anti-Vβ8/streptavidin-PerCP. IL-2 levels of supernatants were measured in triplicate by ELISA.
For Bcl-xL expression, purified T cells activated 6 days with SEB were surface stained with biotinylated anti-Vβ8/streptavidin-PerCP, then permeabilized with a Cytofix/Cytoperm kit, according to the manufacturer’s directions (BD Biosciences). Cells were then stained with either Bcl-xL (Cell Signaling Technology) or isotype control Abs, and then analyzed by FACS. For determining expression of IFN-γ and IL-17, T cells were harvested from the spleens of mice 14 days postimmunization with MOG35–55 emulsified in CFA, and magnetically sorted CD4+ T cells were stimulated with PMA plus ionomycin for 6 h in the presence of GolgiPlug (BD Biosciences). Cells were stained with anti-CD4 allophycocyanin, fixed, and permeabilized, as above, and then stained with anti-IFN-γ PE or anti-IL-17 PE. For intracellular DRAK2 levels, thymocytes from wild-type and Drak2−/− mice were harvested and stained with anti-CD4 allophycocyanin and anti-CD8 PE, followed by permeabilization, as above. Anti-DRAK2 was added for 30 min, followed by washing and staining with FITC-conjugated anti-rabbit Abs.
Quantitative real-time RT-PCR
Purified splenic T cells from either Drak2-deficient mice or wild-type littermates at a density of 106/ml were activated with plate-bound anti-CD3ε (1 μg/ml for optimal or 0.1 μg/ml for suboptimal stimulation) in the presence or absence of 1 μg/ml soluble anti-CD28 for either 24 or 48 h, and subsequently expanded in the presence of IL-2 (100 U/ml). Following isolation of total RNA with TRIzol solution, cDNA was generated from 1 μg of total RNA for each time point and condition using the Superscript First-Stand Synthesis System with oligo(dT) primers (Invitrogen). Samples were analyzed in triplicate by quantitative real-time RT-PCR with an iCycler using the iQ SYBR Green Supermix and specific primers for 40 amplification cycles. After normalization of all data to β-actin, fold changes were calculated by dividing the values for stimulated by unstimulated samples.
Murine stem cell virus/internal ribosome entry site/GFP (Mig) vectors encoding Bcl-XL were gifts from M. Croft (La Jolla Institute for Allergy and Immunology, La Jolla, CA). The ψ-eco packaging vector was a gift from O. Witte (University of California, Los Angeles, CA). Retrovirus was collected from the supernatants of calcium phosphate-transfected 293T cells and titered by GFP expression of infected 3T3 cells. T cells were infected by centrifugation for 90 min at 1800 rpm in the presence of retrovirus and 4 μg/ml polybrene at 48 and 72 h, following activation with SEB. Eighteen hours later, retroviral supernatant was replaced with fresh medium and cells were left to culture for a total of 8 days. Cells were then collected and stained with biotinylated anti-Vβ7 and anti-Vβ8/streptavidin-PerCP for FACS analysis.
EAE and histology
Mice were challenged with MOG35–55, as previously described (40). Mice were immunized with 125 μg of MOG35–55 (prepared in the laboratory of C. Glabe, University of California, Irvine, CA) emulsified in CFA containing H37Ra Mycobacterium tuberculosis (Fisher Scientific) in each hind flank on days 0 and 7. Mice also received i.p. injections of 200 ng of Bordetella pertussis toxin (List Biological Laboratories) in sterile PBS on days 0, 2, and 7. Mice were then monitored for signs of disease on the indicated days. Clinical disease following EAE induction was assessed using a previously described scale (40), as follows: 0.5, altered gait and/or hunched appearance; 1, limp tail; 2, partial hind limb paralysis; 3, complete hind limb paralysis; 4, complete hind limb paralysis and partial fore limb paralysis; and 5, death. Spinal cords and brains were removed at 25 day postimmunization. Spinal cord sections were fixed by immersion in 10% normal buffered formalin for 24 h for paraffin embedding. The severity of demyelination was determined by luxol fast blue staining of spinal cords and analyzed by light microscopy (45). Frozen brain sections were stained with Alexa-488-conjugated anti-CD3 (clone 2C11), followed by light microscopy.
Bim Western blotting
Wild-type and Drak2−/− T cells were cultured with mitomycin C-treated wild-type APCs pulsed with 2 μg/ml SEB. On days 0, 3, and 6, T cells were harvested, cleared of nonviable cells using Ficoll, and then magnetically sorted for Vβ7/8 expression using magnetic beads. Vβ7/8-purified T cells were lysed in complete lysis buffer (150 mM NaCl, 50 mM sodium fluoride, 10 mM β-glycerophosphate, 20 mM HEPES (pH 7.4), 1% Triton X-100 with 1 mM sodium vanadate, 1 mM PMSF, 1 mg/ml aprotinin, and 1 mg/ml leupeptin), and lysates were resolved on a 12% SDS-PAGE gel. The proteins were transferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore), prewashed in TBST, then blocked in TBST supplemented with 5% BSA and incubated with anti-BIM/BOD-purified polyclonal Ab (Imgenex) at a 1/1000 dilution. Following an overnight incubation with the primary Ab, the blot was washed three times in TBST, and incubated with peroxidase-labeled anti-rabbit secondary Ab (Vector Laboratories) for a 1-h incubation at room temperature. To visualize Western, the blot was washed three times in TBST, then incubated in ECL reagent (Amersham Biosciences) and exposed to x-ray film.
Enhanced death of Drak2−/− T cells following superantigen stimulation is blocked by tonic signaling through Bcl-xL
To investigate the potential that DRAK2 may contribute to survival signaling, purified Drak2−/− or wild-type T cells were stimulated in vitro with the superantigen SEB in the presence of wild-type, mitomycin C-treated APCs, and then analyzed for proliferation and SEB-reactive (Vβ7/8 TCR+) T cell recovery (46). Drak2−/− T cells had both decreased proliferation and live cell recovery in response to SEB (Fig. 1, A and B). This was not due to intrinsic defects of the SEB-reactive pool in Drak2−/− mice because the proportion of Vβ8+ T cells, as well as the CD4:CD8 ratio within this pool, was similar to wild-type mice (supplemental Fig. 1, A and B).5 Also, SEB-activated Drak2−/− T cells up-regulated activation markers and produced IL-2 to a similar extent as wild-type T cells, excluding decreased activation or distinct TCR repertoire pools in Drak2−/− mice as a cause of reduced SEB-mediated proliferation (supplemental Fig. 1, C and D).5 To determine whether this decreased proliferation and recovery were due to enhanced apoptosis, purified T cells were CFSE labeled and subsequently analyzed by FACS after 8 days of SEB culture. Recovery of live proliferating Drak2−/− T cells was significantly reduced following SEB stimulation (Fig. 1,C, left panel). A greater proportion of dividing Drak2−/− T cells was annexin Vhigh, demonstrating that the majority of these T cells responded to SEB by proliferating and subsequently undergoing apoptosis (Fig. 1,C, right panel). Drak2−/− T cell death following superantigen stimulation was Fas independent, because treatment with Fas:Fc failed to restore clonal expansion (Fig. 1 D) or survival of these cells (data not shown).
That FasL blockade failed to rescue abortive proliferation in this context was consistent with previously published data indicating that superantigen-mediated T cell apoptosis occurs via a death receptor-independent pathway. Rather, this death is mediated by control over the relative expression of prosurvival and proapoptotic Bcl-2 family proteins (14, 18, 47, 48). This form of death can be prevented by immunological adjuvants and costimulatory survival signals that promote sustained Bcl-2 and Bcl-xL expression (27, 31, 32, 49, 50, 51, 52, 53). To determine whether costimulation might restore Drak2−/− T cell survival, purified T cells were stimulated with SEB in the presence of agonistic anti-CD28 to mimic enhanced costimulatory conditions. Exogenous CD28 ligation restored Drak2−/− T cell proliferation and live cell recovery in response to SEB (Fig. 2,A), and this rescue was most profound for high-dose superantigen stimulation (Fig. 2,B). CD28 not only promotes survival of activated T cells, but amplifies TCR signaling and subsequent T cell activation (54). Indeed, anti-CD28 enhanced SEB-mediated IL-2 production by both wild-type and Drak2−/− T cells (supplemental Fig. 2),5 although exogenous IL-2 did not promote Drak2−/− T cell survival (data not shown). However, anti-CD28 reduced the proportion of dividing Drak2−/− T cells that were annexin Vhigh (Fig. 2, C and D). Surface expression of CD28 was indistinguishable between wild-type and Drak2−/− T cells following SEB activation for 8 days (supplemental Fig. 3),5 demonstrating that the increased costimulatory requirement in the mutant T cells is not due to diminished CD28 expression. Addition of the homeostatic cytokines IL-7 and IL-15 also restored Drak2−/− T cell survival (Fig. 2 E), suggesting that Drak2−/− T cells have a greater dependence on survival factors to prevent their apoptotic demise following SEB activation. Addition of blocking Abs to IL-7 and IL-15 failed to block anti-CD28-mediated rescue of SEB-stimulated Drak2−/− T cells (data not shown), suggesting that such tonic signaling may occur independently via cytokine or costimulatory receptors.
CD28-mediated costimulation and IL-7/IL-15 all promote T cell survival through the up-regulation of anti-apoptotic proteins such as Bcl-2 and Bcl-xL (31, 37, 55). Using quantitative real-time PCR, we observed that anti-CD3-stimulated Drak2−/− T cells displayed diminished levels of Bcl-xL following 6-day culture in IL-2 when compared with wild-type T cells, although levels of expression of Bcl-xL after 24 h were indistinguishable (Fig. 3,A). Similarly, we found that Drak2−/− T cells stimulated for 6 days with SEB-pulsed APCs failed to maintain Bcl-xL expression, and this was restored to wild-type levels by addition of anti-CD28 (Fig. 3,B). Consistent with diminished Bcl-xL expression, we observed a greater proportion of DiOC6low/hydroethediumhigh Drak2−/− T cells following mitogenic stimulation (Fig. 3, C and D), a probable consequence of mitochondrial disruption known to occur during intrinsic apoptosis (56). To determine whether expression of Bcl-xL was sufficient to promote Drak2−/− T cell survival, SEB-activated T cells were infected with a Mig retrovirus encoding the infection marker GFP with or without Bcl-xL, and then collected on day 8 for FACS analysis. Because cell division is required for infection by murine stem cell virus-based retroviruses, the relative percentage of live-gated GFP+ cells recovered was used as an indicator of proliferating T cell survival. Retrovirus-enforced expression of Bcl-xL rescued SEB-responsive Drak2−/− T cell recovery (Fig. 3,E; supplemental Fig. 4),5 consistent with the hypothesis that Drak2−/− T cells are hypersensitive to intrinsic apoptosis following superantigen stimulation. As an alternative means to validate this hypothesis, Drak2−/− mice were bred with a Tg mouse line in which Bcl-xL is constitutively expressed in T cells (Lckpr-Bcl-xL Tg), and also with mice deficient in the proapoptotic Bcl-2 family member Bim (22, 57). Previous studies using these mouse lines have demonstrated the rescue of T cells from activated cell death (20, 31, 47). Importantly, Bim has been shown to be essential for SEB-induced T cell death (47). Although Tg Bcl-xL expression enhanced the recovery of wild-type T cells, there was even greater recovery of Drak2−/− × Bcl-xL Tg T cells following SEB stimulation (Fig. 3,F), supporting our findings with retrovirus-enforced Bcl-xL expression. A Bim deficiency improved the recovery of wild-type and Drak2−/− T cells to a similar extent, consistent with the role of this BH3-only proapoptotic Bcl-2 homologue in promoting T cell apoptosis (13). However, Bim expression was indistinguishable between wild-type and Drak2−/− T cells following SEB stimulation (Fig. 3 G). This finding suggests that Bim contributes to the apoptotic hypersensitivity of Drak2−/− T cells, but that its expression is not likely a target of DRAK2 activity. Taken together, these results demonstrate that Drak2−/− T cells are apoptotically hypersensitive and require heightened tonic signaling to maintain their survival.
Blockade of apoptosis by T cell-intrinsic Bcl-xL expression restores EAE sensitivity to Drak2−/− mice
Drak2−/− mice are resistant to EAE, and display a profound lack of T cells recruited to the CNS 21 days following MOG-peptide immunization (40). Our results with SEB suggested that enhanced T cell-instrinsic apoptosis might account for the resistance of Drak2−/− mice to this autoimmune disease. To test this hypothesis, EAE was induced in wild-type, Drak2−/−, Bcl-xL Tg, and Drak2−/− × Bcl-xL Tg mice by immunization with MOG35–55 peptides plus CFA (40, 44). Consistent with previous results, Drak2−/− mice had diminished clinical scores following MOG peptide immunization, with two independent experiments shown (Fig. 4, A and B). Although Tg Bcl-xL expression did not significantly alter disease severity in an otherwise wild-type background, this transgene completely restored the susceptibility of Drak2−/− mice to EAE. Because the Bcl-xL transgene is T cell specific, these results not only implicate apoptotic hypersensitivity as a mechanism of resistance of Drak2−/− mice to EAE, but also demonstrate that this resistance is T cell intrinsic. Consistent with this, we observed restoration of CD3+ T cells in the CNSs of MOG35–55-immunized Drak2−/− mice and enhanced demyelination upon expression of the Bcl-xL transgene (Fig. 4, C and D). We also observed a defect in the maintenance of MOG35–55-reactive CD4+ T cells as assessed by ex vivo culture of splenic T cells with Ag-pulsed APCs following 7 days of MOG immunization (Fig. 5,A). Immunization of Drak2−/− mice failed to give rise to increased ratios of IFN-γ- or IL-17-expressing CD4+ T cells (Fig. 5, B and C), although Drak2−/− mice have unimpaired differentiation of either Th1 (40, 44) or Th17 (Fig. 5,D) subsets following culture under biasing conditions. These findings suggest that the maintenance of MOG-reactive Th1 and Th17 cells, subsets known to be required for EAE (58), is dependent upon DRAK2. Consistent with this, we observed that the Bcl-xL transgene restored an increased fraction of MOG-reactive IFN-γ-expressing Th1 clones following 14 days of in vivo MOG peptide challenge (Fig. 5,E). Although we did not observe significant restoration of MOG-reactive Th17 cells by the Bcl-xL transgene in Drak2−/− spleens (Fig. 5 F), it may be that such cells were instead recruited to the CNS to promote EAE, or that rescued anti-MOG Th1 cells may have supplanted the need for Th17 cells (59). Taken together, our data provide evidence that DRAK2 plays an essential role in maintaining the survival capacity of MOG-reactive T cells in vivo during EAE.
Enhanced thymocyte-positive selection in Drak2−/− mice revealed by a Bcl-xL transgene
Given that DRAK2 serves to control the activation and survival of peripheral T cells, we turned our attention to thymocyte selection to determine whether an analogous duality in its function may be present during thymopoiesis. In previous studies, we observed a role for DRAK2 in setting the threshold for TCR-induced Ca2+ mobilization in double-positive thymocytes during selection (60). Consistent with this, intracellular staining with anti-DRAK2 mAbs revealed predominant expression in thymic subsets undergoing, or having completed selection (Fig. 6,A; supplemental Fig. 5).5 Namely, we observed high levels of DRAK2 in single-positive and transitional T cells. Elevated DRAK2 expression was also observed in CD4low/CD8low thymocytes, a population thought to represent thymocytes undergoing selection (61). Although Drak2−/− mice displayed modestly enhanced thymocyte-positive selection when bred with AND and OT-II Tg mice, no obvious differences in positive selection were observed in non-TCR Tg backgrounds, and negative selection was not apparently impacted by this deficit (40). Because TCR signal strength is a fundamental determinant for the outcome of double-positive thymocyte selection, we wished to establish whether the absence of an enlarged population of positively selected thymocytes in non-TCR Tg Drak2−/− mice might be due to the apoptotic demise of such cells. To assess this, we evaluated the thymic CD4 vs CD8 phenotypes of Drak2−/− mice with or without the Bcl-xL transgene. Although Bcl-xL Tg mice possessed a significant increase in the proportion of CD8 single-positive thymocytes, we observed a dramatic increase in the proportion of both CD4 and CD8 single-positive thymocytes derived from Drak2−/− × Bcl-xL mice in five separate experiments, with representative data shown for littermates of the indicated genotypes (Fig. 6,B). We failed to observe significant differences in thymus size or cellularity between littermates in these studies (data not shown). Although the Bcl-xL transgene led to enhanced proportions of both CD4+ and CD8+ peripheral T cells, the absence of DRAK2 did not appear to affect these peripheral lymphocyte numbers (Fig. 6 C). These data suggest that DRAK2 restricts thymocyte selection by negatively regulating TCR signaling, and that in its absence, selected T cells are hypersensitive to apoptosis. Given that no major differences in CD4+ and CD8+ peripheral T cell populations were observed beyond those provided by enforced Bcl-xL expression, our results also reveal that DRAK2 most likely does not control the maintenance of peripheral naive T cells.
Our studies indicate that whereas DRAK2 blocks T cell activation (40, 60, 62), this lymphoid-enriched serine-threonine kinase also plays an essential role in maintaining the survival of T cells activated under specific contexts. We have found that Drak2−/− T cells are more susceptible to intrinsic apoptosis, suggesting that DRAK2 activity promotes the survival of T cells during clonal expansion following antigenic stimulation. Consistent with these findings, we observed diminished T cell survival following antigenic stimulation with SEB and with MOG peptides. Because SEB and MOG reactivity were restored by a Bcl-xL transgene, our results are in accordance with the hypothesis that DRAK2 signaling impacts the balance in expression of prosurvival vs proapoptotic Bcl-2 family members during the course of a T cell response to these Ags.
Although a DRAK2 deficiency leads to defective survival of MOG-reactive cells in EAE, such mice have no significant defects in immunity to acute viral infection. Drak2−/− mice display overtly normal antiviral T cell responses to lymphocytic choriomeningitis virus (LCMV) and murine hepatitis virus (MHV), and generate antiviral memory T cells (40, 44, 63). Although the bases for the differences in apoptotic hypersensitivity are currently unclear, we hypothesize that necessity for DRAK2 signaling in peripheral T cells may be dependent upon the inflammatory context or nature of the Ag. Unlike an antiviral response in which pathogen-associated molecular patterns elicit an array of costimulatory survival molecules and inflammatory cytokines, EAE results from an autoimmune response against self-Ag induced in the absence of sustained pathogenic danger signals (64, 65). It is therefore possible that Drak2−/− T cells do not receive sufficient costimulatory- or cytokine-mediated survival signals within the CNS during EAE development to maintain their survival. Consistent with this, Tg Bcl-xL expression restored susceptibility of Drak2−/− mice to EAE. Additionally, Drak2−/− mice develop neuroinflammatory disease indistinguishable from Drak2+/+ mice following neurotropic MHV infection, a model in which chronic viral infection promotes sustained and/or enhanced costimulatory ligand and inflammatory cytokine expression within the CNS (44, 66, 67, 68). Although there may be a variety of alternative explanations for the differential requirements for DRAK2 signaling in antiviral vs autoimmune settings, we have not observed increased apoptotic sensitivity of Drak2−/− T cells responding to MHV or LCMV. Thus, it is likely that the apoptotic hypersensitivity of Drak2−/− T cells is limited by tonic factors present in an antiviral context.
Previous work has established that DRAK2 restricts TCR-dependent signaling in developing thymocytes, peripheral T cells, and antiviral memory CD8+ T cells (40, 60, 63). In this study, we have demonstrated that DRAK2 also serves to maintain the survival of activated T cells and selected thymocytes. Expression of a Bcl-xL transgene in Drak2−/− mice led to restoration of MOG-induced EAE, and enhanced the recovery of positively selected thymocytes. Given the enhanced recovery of positively selected thymocytes in Drak2−/− × Bcl-xL mice, it is possible that the restoration of EAE sensitivity observed might be due to enlarged populations of peripheral T cells in these mice. However, this explanation is unlikely because the Bcl-xL transgene only modestly affected the recovery of naive peripheral CD4+ T cells, the subset that responds to MOG35–55. Although a DRAK2 deficiency during thymopoiesis may very well alter the repertoire of the peripheral T cell pool, we suggest that DRAK2 also plays an important role in peripheral T cell survival. This hypothesis is strengthened by the finding that retrovirus-enforced expression of Bcl-xL rescued peripheral T cell responses to SEB. Thus, the restoration of EAE sensitivity by the Bcl-xL transgene most likely reflects enhanced survival of peripheral effector T cell populations. Recently, it has been demonstrated that transgene-enforced overexpression of DRAK2 led to enhanced T cell apoptosis and a resultant decrease in memory T cell pools (43). Although seemingly contradictory to the results presented in this study, we note that these mice were produced using a human β-actin promoter, a promoter likely to give rise to unregulated and ectopic expression of the DRAK2 transgene in a variety of cell types. It is likely that DRAK2 expression levels must be carefully regulated during T cell development and in naive T cells, because this kinase plays significant roles in setting TCR signaling thresholds (40, 60). As well, overexpression in these cells may trivially lead to apoptosis, as has been observed in other cell types that do not normally express DRAK (39, 41, 69). However, the basis for enhanced T cell apoptosis in such Tg mice remains to be described.
Although the direct targets of DRAK2’s enzymatic activity are currently unknown, DRAK2 most likely participates at some level in Ca2+ signaling in primary lymphocytes. Drak2−/− T cells and thymocytes mobilize Ca2+ to a much greater degree than wild-type T cells in response to weak TCR stimulation. Recently, we have observed that Ag receptor stimulation of T and B cells led to DRAK2 autophosphorylation of Ser12 (62). The Ag receptor-mediated induction of DRAK2 enzymatic activity is highly dependent upon Ca2+ mobilization, because calcium chelators blocked Ser12 autophosphorylation. Moreover, depletion of intracellular Ca2+ is sufficient to activate DRAK2 because treatment of lymphocytes with the sarco/endoplasmic reticulum Ca2+-ATPase pump inhibitor thapsigargin alone induces DRAK2 Ser12 autophosphorylation. These studies suggest that DRAK2 serves in a negative feedback loop to restrict levels of Ca2+ in lymphocytes responding to antigenic stimulation. Although DRAK2 lacks a calmodulin-binding motif, that Ca2+ modulates the activity of DRAK2 is not surprising given that other DAPK family members are also subject to its regulation.
Ca2+ has long been known to play significant roles in lymphocyte activation and survival (70). The recent discoveries of Stim1 and Orai1, the components of store-operated channels that allow heightened and sustained cytoplasmic Ca2+ following depletion of endoplasmic reticulum stores (71, 72, 73, 74), have provided further impetus for investigating the exquisite control lymphocytes must exert over this fundamental second messenger. Importantly, Ca2+ controls the balance between life and death for many cell types, and its dysregulation under pathological states is often deleterious (75). In the context of T cells, Ca2+ signaling controls both the activation and survival status of responding clones. One example of this is the factor Csp1, a negative regulator of calcineurin. Csp1 sets the threshold for NF-AT activity in T cells via its restriction of calcineurin, and its deletion not only leads to enhanced T cell activation, but heightened expression of FasL and AICD hypersensitivity (25). The myelin basic protein splice variant golli also controls the activation and survival status of T cells in vivo by restraining store depletion-induced Ca2+ signaling in primary T cells (76). Similar to Drak2−/− T cells, golli−/− T cells are hyperproliferative to TCR cross-linking, produce enhanced levels of IL-2 upon suboptimal stimulation, and are refractory to MOG35–55-induced EAE (77). Although the basis for the enigmatic autoimmune resistance of golli−/− mice remains to be determined, it may be that T cells lacking golli are apoptotically hypersensitive due to defective Ca2+ homeostasis in an analogous fashion.
In our work, we have established that DRAK2 plays dual roles in T cells, as follows: first, to restrict signal transduction downstream of the TCR, and second, to maintain the survival of activated T cells. Because DRAK2 is an enzyme and may have multiple substrates in lymphocytes, it is quite possible that these are independent enzymatic functions. For example, naive T cell activation may be restricted by one signaling cascade regulated by DRAK2, whereas blockade of apoptosis via the regulation of prosurvival factors like Bcl-xL may be orchestrated by a discrete pathway independently affected by DRAK2 activity. Alternatively, DRAK2 may affect these distinct features of T cell biology through the same signaling cascade. Studies are currently underway to carefully establish the targets of DRAK2 activity to address this important question. Nevertheless, the studies presented in this work suggest that DRAK2 may be an important target for certain classes of autoimmune disease. Although DRAK2 activity is necessary for the development of EAE, DRAK2 most likely is involved in the development of other autoimmune diseases. This provokes an important question: why would a protein such as DRAK2 significantly impact autoimmune responses, but only modestly participate in antimicrobial immunity? It is likely that DRAK2 plays other roles in controlling immunity. For example, DRAK2 function is essential for T cell help during germinal center reactions, and Drak2−/− mice fail to produce high-affinity Abs; rescue of germinal center Drak2−/− T cells with a Bcl-xL transgene led to a restoration of germinal centers and high-affinity Abs (78). Furthermore, Drak2−/− mice display defects in the maintenance of encephalitogenic West Nile virus-reactive T cells, leading to resistance of these mice to lethal encephalitis (79). Despite this, Drak2−/− mice possess normal immune responsiveness to acute viral infection with LCMV and MHV (40, 44). Thus, reagents to specifically block DRAK2 function may very likely have significant clinical utility. Most current therapies against autoimmune disease result in generalized immunosuppression. Given the results presented in this study, inhibition of DRAK2 may specifically prevent autoimmunity by targeting self-reactive T cells for clonal deletion without dramatically affecting the integrity of a normal immune response against acute microbial infection.
We thank Professors David A. Fruman, Thomas E. Lane, and Aimee Edinger, and members of the laboratory of C.M.W. for insightful comments regarding this manuscript. We thank Drs. Charles Glabe and Saskia Milton for MOG35–55 peptide synthesis and purification. We also thank Drs. Stephen Hedrick, Michael Croft, Garry Nolan, and Pippa Marrack for reagents and protocols.
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 grants from the National Institutes of Health (RO1-A63419 and RO1-AI50506 to C.M.W., and T32-AI60573 to S.J.R.), and by fellowships from the Arthritis National Research Foundation (to C.M.W. and M.G.).
Abbreviations used in this paper: AICD, activation-induced cell death; ACAD, activated cell autonomous death; Csp1, calcipressin-1; DAPK, death-associated protein kinase; DRAK2, DAPK-related apoptosis-inducing kinase 2; EAE, experimental autoimmune encephalomyelitis; FasL, Fas ligand; LCMV, lymphocytic choriomeningitis virus; MHV, mouse hepatitis virus; MOG, myelin oligodendrocyte; SEB, staphylococcal enterotoxin B; Tg, transgenic.
The online version of this article contains supplemental material.