DRAK2 is unique in that it is one of the few proteins that is specifically expressed in lymphoid tissue. Based on its similarity to DAP-like family members and direct experimentation (Sanjo et al., 1998
), we predicted that DRAK2 would play a role in T cell apoptosis. However, we demonstrate here that the absence of Drak2
does not affect apoptosis of T cells in a number of different assays. This includes apoptosis of thymocytes in vitro and two models of in vivo negative selection. In addition, there was no defect in death of mature T cells after in vitro stimulation or after an in vivo immune response. Furthermore, overexpression of Drak2
in Jurkat T cells did not result in an increase of apoptosis (data not shown). Therefore, we conclude that DRAK2 does not play a critical role in T cell apoptosis.
An analysis of Drak2−/− mice revealed that it encodes a negative regulator of T cell activation, and this is supported by several distinct phenotypic differences in T cell development and activation. First, positive selection of CD4+ T cells was enhanced as illustrated by the increase in the number of CD4+ thymocytes and an increase in the levels of CD5 and CD69 expression. There was an increase in memory-like T cells that was antigen specific. The requirement for costimulation was dramatically reduced although Drak2−/− T cells retained sensitivity to costimulation (). In fact, the effects of costimulation, including both proliferation and survival, were identical in Drak2−/− and wt T cells, with the exception that Drak2−/− T cells responded to lower doses of anti-CD28 antibodies (). Control experiments showed this hypersensitivity was not due to differences in APC or due to the fact that there were more CD62Llo T cells in Drak2−/− mice. Together, these data reveal that DRAK2 functions to negatively regulate signals involved in T cell activation.
The most surprising characteristic observed in Drak2−/−
mice was the resistance to EAE despite the T cell hypersensitivity. In other models of negative T cell regulation that have been tested for autoimmunity, the mice were either equally susceptible or more susceptible to disease when compared with wt mice. In fact, deficiencies in Cbl-b (Bachmaier et al., 2000
; Chiang et al., 2000
), Itch (Fang et al., 2002
), Sts-1 and Sts-2 (Carpino et al., 2004
), Mgat5 (Demetriou et al., 2001
), SHP-1 (Kozlowski et al., 1993
; Shultz et al., 1993
; Tsui et al., 1993
), Foxj1 (Lin et al., 2004
), CTLA-4 (Tivol et al., 1995
; Waterhouse et al., 1995
), PD-1(Nishimura et al., 1999
), SHIP (Helgason et al., 1998
), and Pten (Suzuki et al., 2001
) all resulted in an increased susceptibility to autoimmunity. Thus far, DRAK2 appears to be the only negative regulatory molecule that nonetheless conveys resistance to autoimmunity.
Due to the complexity of EAE, there are many opportunities for a regulatory molecule to affect disease. It is possible that there is a defect in CD4+
T cell expansion and survival in response to MOG immunization; however, this does not seem likely given the results of experiments measuring the T cell response in culture or in response to an infectious agent in vivo. In addition, at day 20 after immunization with MOG, there were similar numbers of antigen-reactive cells in the spleen and draining lymph nodes of Drak2−/−
and wt mice (data not shown). Rather, the loss of disease appears to originate from a loss of CNS infiltration. The number of MOG-specific cells in the spinal cord and brain of the Drak2−/−
mice was reduced compared to wt and Cbl-b−/−
mice, and this number correlated strongly with disease. We are currently investigating the possible reasons for the difference in migration to the central nervous system (CNS), and obvious possibilities include a disregulation of adhesion molecules, chemokines, or chemokine receptors. Consistent with this possibility, there were more CD62Llo
cells in the spleen of Drak2−/−
mice compared to wt mice (Supplemental Figure S4
Another way in which the absence of Drak2
could affect autoimmunity is by altering the cytokine expression of T cells. The amount and timing of cytokine expression is critical in EAE. An increase in Th2 cytokines can impart resistance to EAE and, paradoxically, IFN-γ has also been suggested to play an inhibitory role (Pedotti et al., 2003
; Willenborg et al., 1996
). In addition, T cells deficient in MAP kinase phosphatase 5 (Mkp5
) have elevated levels of activated JNK and produce more cytokines than wt T cells; however, Mkp5
-deficient mice are resistant to EAE (Zhang et al., 2004
). Clearly, the role of cytokines in EAE is not completely understood, but it is conceivable that an increase in either IFN-γ or IL-4 by Drak2−/−
T cells could result in decreased EAE ().
-deficiency could affect EAE as a consequence of altered B cell regulation. Previous experiments showed that B cells can play a protective role in EAE through the production of IL-10 (Cross et al., 2001
; Dittel et al., 2000
; Wolf et al., 1996
). Because DRAK2 is highly expressed in B cells, it is possible that Drak2−/−
B cells are also hypersensitive and produce more disease-inhibiting IL-10. Consistent with this, Drak2−/−
B cells were hyperproliferative to suboptimal stimuli in vitro (data not shown).
Finally, the effect of Drak2 on autoimmunity may be through regulatory T cells. Given that Drak2−/− conventional T cells are hypersensitive to antigenic stimulation, it is possible that Drak2−/− regulatory T cells are also hypersensitive to stimulation, and the response of the regulatory T cells could dominate over that of the conventional CD4+ T cells that induce disease. However, in vitro, Drak2−/− regulatory T cells did not suppress T cell activation more effectively than wt T cells (data not shown). In addition, we did not detect an increase in the number of CD4+CD25+CD69− or CD4+GITR+T cells in the CNS or draining lymph nodes at day 20 after MOG immunization in Drak2−/− mice compared to wt mice (data not shown). However, this does not eliminate the possibility that there are more regulatory T cells present early in the response, and this is sufficient to inhibit disease.
In conclusion, the absence of Drak2 does not appear to negatively impact the health or fecundity of mice. Yet, a Drak2 deletion renders mice resistant to autoimmune disease though fully capable of mounting a normal immune response to an infectious virus. It thus presents an evolutionary enigma–what was the selective pressure to evolve a regulatory molecule that inhibits the effectiveness of an immune response and increases the propensity for autoimmunity? Whatever the answer, we note that DRAK2 constitutes a cell type-specific target that could be inhibited to treat autoimmune disease without the risks normally associated with immune inhibition.
Drak2−/− mice were generated by homologous recombination in 129SVJ embryonic stem cells. The genomic clone corresponding to Drak2 was obtained by screening a 129SVJ mouse genomic lambda phage library (Stratagene, La Jolla, CA) by using full-length Drak2 cDNA as a probe. Most Drak2−/− mice analyzed were backcrossed to C57Bl/6 at least six generations, unless indicated in the figure legend. In all cases, littermates were used as controls. Cbl-b−/− mice were a generous gift from Hua Gu and were backcrossed 12 generations to C57BL/6.
Purification of Lymphoid Populations
DN, CD8+, CD4+, and postselection DP thymic subsets were purified from C57BL/6 thymuses by sorting with a FACSVantage (Becton Dickinson, San Jose, CA) after magnetic depletion with MACS beads (Miltenyi Biotech, Auburn, CA). Preselection DP thymocytes were obtained from a MHC I/II−/− (β2M−/−;I-Ab−/−) thymus and sorted for expression of CD4 and CD8 with the FACSVantage. Peripheral T cells were purified from the lymph nodes by negative selection with biotin-CD4 or CD8 and B220, CD11b, DX5, and MHC class II (I-Ab) antibodies (eBioscience, San Diego, CA), followed by separation with streptavidin-conjugated magnetic beads (Miltenyi Biotech, Auburn, CA). B cells were purified from the spleen by positive selection with B220-conjugated magnetic beads (Miltenyi Biotech, Auburn, CA). NK cells and macrophages were obtained from splenocytes that were first depleted of T and B cells followed by sorting based on CD11b and NK1.1 expression. Macrophages were CD11b+ NK1.1− and NK cells were NK1.1+ with about half also expressing CD11b. Dendritic cells were obtained from collagenase-digested spleens that were then spun on a BSA gradient. CD19+ and Thy1+ cells were removed by magnetic sorting, and the remaining cells were sorted for expression of CD11c with a FACSVantage. All subsets were greater than 94% pure.
Flow Cytometric Analyses
Single cells suspensions of thymus, lymph node, and spleen were stained with FITC-, PE-, TC-, PerCP-, and APC-conjugated antibodies against CD4, CD8, CD3, CD5, CD25, CD69, CD44, TCRβ, CD62L, PD-1, ICOS, 41BB, CD27, CD127 (IL-7R) Ly6C, CTLA4, and OX40 (Pharmingen, San Diego, CA; Beckman Coulter, Fullerton, CA; and eBioscience, San Diego, CA). Cells were analyzed by flow cytometry on a FACSCalibur (Becton Dickinson, San Jose, CA). Analysis was performed with FlowJo software (TreeStar, Inc., Ashland, OR).
Purified T cells or whole spleen were labeled with 5,6-carboxyfluorescein diacetate succinimidyl ester (CFSE, Molecular Probes, Eugene OR) at 2.5 μM in PBS for 10 min at 37°C, washed with RPMI containing 10% FCS, and incubated with plate bound anti-CD3 or peptide as described below.
Anti-CD3 was plate bound by first binding goat anti-hamster (Vector Laboratories, Burlingame, CA) followed by the appropriate concentration of anti-CD3 ascites (145-2C11). Anti-CD28 ascites (N37) were added at varying concentrations. After 24–72 hr, T cells were harvested and stained with Annexin V (Pharmingen, San Diego, CA).
T cells, purified as described above, were stimulated in Th1 or Th2 conditions as described previously (Rengarajan et al., 2002
). After 4 days, the cells were restimulated for 5 hr in the presence of Brefeldin A, fixed, and permeablized with Cytofix/Cytoperm Plus staining kit (Pharmingen, San Diego, CA). The treated cells were then stained with PE-anti-IL-2, PE-anti-IFN-g, or PE-anti-IL4 (eBioscience, San Diego, CA).
LCMV virus was a generous gift from Raymond Welsh. Mice were infected with 2 × 105 PFU of LCMV via intraperitoneal injection. T cells were isolated from the spleen at various days postinfection and stimulated in vitro for 5 hr in the presence of gp33 or gp61 peptides, IL-2, and Brefeldin A, then stained for intracellular IFN-γ with the cytofix/cytoperm plus kit to detect LCMV-specific T cells. In addition, blood was collected via cardiac puncture in the presence of 15 μl of 0.5 M EDTA. The blood was spun down and plasma was frozen at −80°C and used to determine levels of IFN-γ by ELISA.
T cells were purified as described above and stimulated with anti-CD3 and anti-CD28. After the various time points, the cells were harvested into ice-cold PBS to stop the reaction. The cells were lysed in a way to separate the cytoplasmic fraction from the nuclear fraction as described previously (Frasca et al., 2002
DRAK2 mRNA Expression
Purified T cells were stimulated with plate-bound anti-CD3 for 24 hr, harvested, washed with PBS, and frozen at −20°C. RNA was isolated from frozen cell pellets with RNeasy mini kits (Qiagen, Valencia, CA) and labeled and hybridized as described (Wodicka et al., 1997
) to mouse GeneChip arrays (Affymetrix, Santa Clara, CA). Duplicate chips were used for each condition. Data was analyzed with Resolver Gene Expression Data Analysis software (Rosetta Inpharmatics, Seattle, WA).
Calcium flux was measured by labeling cells with 2 μM Fura red, 1 μM Fluo-4, and 0.2% Pluronic (Molecular Probes, Eugene, OR) for 30 min at room temperature in serum-free media. Cells were washed twice and rested for 20 min in the dark. For stimulation, biotinylated anti-CD3 (eBioscience, San Diego, CA) was added to labeled cells on ice for 15 min, washed, then resuspended with prewarmed streptavidin, and analyzed by flow cytometry. Calcium mobilization was determined by measuring the ratio of Fluo-4:Fura red with FlowJo software (Treestar, Inc., Ashland, OR).
TCR Signaling Analysis
The APC were peritoneal macrophages that were harvested by peritoneal lavage 5 days after injection with 1 ml thioglycollate. The macrophages were labeled with CD11b-conjugated magnetic beads (Miltenyi Biotech, Auburn, CA) and positively selected on a magnetic column. The magnetically labeled APC were then incubated with 5 μM OVAp or 50 μM G4 peptides for 3 hr at 37°C. Purified CD8+ T cells were added to the APC, spun at high speed for 5 sec, and incubated for an additional 5 min at 37°C. Ice-cold PBS was added to stop the stimulation, and tubes were incubated on ice and centrifuged at 4°C. Each sample was then run over the magnet again to separate the APC and T cells. The T cell fraction was washed in PBS and lysed with the cytoplasmic lysis buffer described above. Equal amounts of total protein were loaded on a 10% SDS gel for Western blot analysis with antiphosphorylated ERK (Rabbit polyclonal), phosphorylated p38 (28B10), phosphorylated IκB-α (Ser 32, rabbit polyclonal) all purchased from Cell signaling, (Beverly, MA), ZAP-70 (Transduction Labs, Lexington, KY), and phosphorylated c-Jun (KM-1, Santa Cruz Biotechnology, Inc., Santa Cruz, CA).
Mice were immunized with 125 μg of MOG35-55 (MEVGWYRSPFSR VVHLYRNGK, Genemed Synthesis, San Francisco, CA) emulsified in complete Freund’s adjuvant containing 0.4 mg of H37Ra mycobacterium tuberculosis (Fisher Scientific, Tustin, CA) in each hind flank. The mice also received 200 ng of Bordetella pertussis toxin (List Biological, Campbell, CA) intraperitoneally immediately after immunization and again on day 2. On day 7, the mice were boosted with another 125 μg of MOG35-55 emulsified in complete Freund’s adjuvant, followed by one injection of 200 ng Bordetella pertussis toxin. Each day, the mice were scored for disease using the following scale: 0, no signs of disease; 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 forelimb paralysis. Mouse cages were coded and individual mice were scored without reference to genotype. Mice were euthanized when they reached a score of 4.